1 INFO-VAX	Thu, 08 Jun 2000	Volume 2000 : Issue 318       Contents:P Re:  =?iso-8859-1?q?=AC=FC=B0=D3=A4=BD=A5q=BCx=A4=7E_Looks?= LikeChiCom SPAM to ' Re: Actual vs Theoretical Network Speed # Re: Breathe New Life Into Your VAX! 
 Re: C bashing / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures / Re: C bashing (was Re: VMS File Caching Futures . Re: Charon VAX on PC emulator - Asynch DECNET? Command Line question  Re: Command Line question  Re: Command Line question  Re: Command Line question ( Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger, Re: Detecting memory leak in C with debugger1 Digital TCP/IP Services and Alternet Telnet Ports & Files-11 ODS-2 Readability for FreeBSD$ Re: Hobbyist vms 7.2 alpha questions$ Re: Hobbyist vms 7.2 alpha questions$ Re: Hobbyist vms 7.2 alpha questions; Re: is anyone able to get a DE500-BA working under 6.2 1h3? ; Re: is anyone able to get a DE500-BA working under 6.2 1h3? * Microsoft loses, www.compaq.com disappears* Microsoft loses, www.compaq.com disappears. Re: Microsoft loses, www.compaq.com disappears. Re: Microsoft loses, www.compaq.com disappears Re: MMOV 2.2 and VMS 7.1) Multithreading and event flags limitation  Re: OpenVMS AND THE INTERNETA Re: OpenVMS commentaries (was Re: Gartner commentary on Wildfire) " OpenVMS V7.3 SDK Info; Freeware V5: Re: Part numbers for Qbus SCSI + ESDI controllers, anyone?: Re: Part numbers for Qbus SCSI + ESDI controllers, anyone?' SMTP_RECV_MAIL from TCPIP$smtp_receiver , Re: Syntax-coloring editors (was: C bashing) system slow  Re: system slow  tcp/ip printing  tcp/ip printing 8 Re: What is the simplest way to network two VMS systems? Re: X25 and psi$k_reset 7 xxx.EDT to convert various forms of special characters?   F ----------------------------------------------------------------------  % Date: Wed, 07 Jun 2000 13:49:30 -0700 ! From: Shane.F.Smith@Healthnet.com Y Subject: Re:  =?iso-8859-1?q?=AC=FC=B0=D3=A4=BD=A5q=BCx=A4=7E_Looks?= LikeChiCom SPAM to  C Message-ID: <OF1B0B0F9A.0C27127F-ON882568F7.00726586@HEALTHNET.COM>    Jim Agnew wrote: >Bradford Hamilton wrote:  >> >> Mike Rechtman wrote:  >> > >> > Terry C. Shannon wrote: >> > >F >> > > There are few things I hate as much as SPAM fron Bill Clinton's3 >> > > "strategic partner," a.k.a. Communist China. 	 >> <snip>  >>H >> > >      "=AC=FC=B0=D3=A4=BD=A5q" <ddddccccbbbb@pchome.com.tw> wrote=  in message 2 >>                                              ^^? >> Sorry to pick a nit here, but isnt this message from Taiwan?  >> >> :-)
 >ah so... ;-)   C Or if you're talking to the person who generated it, "Ah sole....."    Shane=   ------------------------------   Date: 7 JUN 2000 20:04:11 GMT 4 From: karcher@thuria.waisman.wisc.edu (Carl Karcher)0 Subject: Re: Actual vs Theoretical Network Speed5 Message-ID: <7JUN00.20041193@thuria.waisman.wisc.edu>   K In a previous article, "Gord Coulman" <nospam_gcoulman@ccinet.ab.ca> wrote: E ->Can anyone recommend a good utility to measure the actual bandwidth K ->available between two machines?  A Win9x utility is what I need (only one  ->OpenVMS machine).   A You can get TTCP.C for VMS by searching for "TTCP" in the OPENVMS G DSNLINK database. TTCP for Win 9x and NT is available in a java version I from http://www.ccci.com/tools/ttcp/download.htm. There's also a compiled G version for Window/NT from http://ben.cat.pdx.edu/netlab/products/ttcp/  (I've never tried this one).   ------------------------------  % Date: Wed, 07 Jun 2000 19:57:07 +0100 + From: Neil Pike <100577.553@compuserve.com> , Subject: Re: Breathe New Life Into Your VAX!1 Message-ID: <VA.00000dac.1f90004b@compuserve.com>   @  I believe there are ODBC drivers that will allow you to do this   ?  >     Here is a question for all; how do you do the following:  > J >         Connect to an RBS file on a VAX system using SQL 7 and NT 4 SP5? > M >     I've been assigned to figure out a way to "mirror" the old VAX data and K > "new" SQL Server data.  My knowledge in this area is very limited and any N > assisstance would be greatly appreciated!  Our ablilty to do this would saveF > myself and a few other hundreds of hours of writting/testing code toG > generate textfiles on both systems, transfer them via Reflections and 	 > update.   W Q.     Where can I get an OLE-DB/ODBC/JDBC driver to connect SQL Server to product xyz?   T A. The best overrall lists of ODBC/JDBC/OLEDB vendor's I know is run by Ken North :-  C    http://ourworld.compuserve.com/homepages/Ken_North/odbcvend.htm  C    http://ourworld.compuserve.com/homepages/Ken_North/jdbcvend.htm  C    http://ourworld.compuserve.com/homepages/Ken_North/oledbven.htm     ===    v1.01 2000.02.05% Applies to SQL Server versions  : All . FAQ Categories                  : Connectivity5 Related FAQ articles            : odbcunix, mac, jdbc % Related Microsoft Kb articles   : n/a % Other related information       : n/a + Authors                         : Neil Pike +  Neil Pike MVP/MCSE.  Protech Computing Ltd "  (Please reply only to newsgroups)  SQL FAQ (428 entries) see  \  forumsb.compuserve.com/gvforums/UK/default.asp?SRV=MSDevApps (sqlfaq.zip - L7 - SQL Public)  or www.ntfaq.com/sql.html  or www.sql-server.co.uk  or www.mssqlserver.com/faq    ------------------------------  % Date: Wed, 07 Jun 2000 17:11:54 -0400 * From: David A Froble <davef@tsoft-inc.com> Subject: Re: C bashing- Message-ID: <393EBA9A.BD0462FC@tsoft-inc.com>    Christopher Smith wrote:  L > It's difficult to figure out whether to keep trying to convince people, or7 > just give up and let them figure it out the hard way.   N People believe what they WANT to believe.  This is real easy when they can getM someone to confirm what thay want to believe, and billie-boy sure does that.  L Best solution is to give your opinion once.  Any more, and people will buildO resistance to seeing reality, cause you're shoving it in their face before they O are ready to see it on their own.  Eventually, they will see the light, and may N remember that you were right on that topic the next time you say something, orN they will be forever blind.  Two things to remember, 1) it doesn't matter, andP 2) it REALLY doesn't matter.  Keeps your blood pressure within tolorable limits.   Dave   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  $ Date: Wed, 7 Jun 2000 11:45:20 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures$ Message-ID: <bLw%4.73$Rp5.65@client>  J "Wayne Sewell" <wayne@tachysoft.xxx.044962.killspam.0138> wrote in message( news:mm0UqdQic5Db@tachxxsoftxxconsult... [snip]G > It had to be *ported*; it did not necessarily have to be *converted*.  VeryI > little of the vms source code written in macro had to be converted to a E > different language to support alpha, if any.   Some of it had to be  modifiedH > because of vaxisms in the code, but it did *not* have to be converted.K > Digital basically created a macro *compiler*, treating vax macro like any  other L > computer language, including even optimization.   In other words, a single vax L > instruction produces many alpha machine instructions.  Contrast this to anI > assembler, in which there is a one-to-one correspondence between source E > instruction and machine instruction.   From the viewpoint of a risc 
 processor,J > a cisc instruction set *is* a high level language.  In any case, much of the J > vms source code is *still* macro, built with the vax *assembler* and the alpha , > *compiler* from basically the same source. >  > G > You accused someone else in this thread of pulling information out of  their I > ass.  It appears you are doing the same.  You simply assumed all of the  macro H > code had to be converted to another language, and then you assumed theJ > conversion was to C.  If conversion had actually been required, it could have > just as easily been to Bliss.   F Quite right, I have been taken to task on this one, and found wanting.  A I do suspect (without proof) ;-) that Tru64 Unix is written in C. ) Would you describe this as a robust tool?   J And consider the C that is used to construct OpenVMS. It is 1/3 C or 1/4 CG or something like that, at any rate.  Is it (therefore) not robust?  Is L OpenVMS not to be trusted as a secure operating system?  Should we switch to IBM hardware and MVS?   I The point remains, bad software is written in every language.  It is true L that you can do truly stupid things in C.  But that is not a language fault.I It is a programmer fault.  Further, there are stupid programmers in every ' discipline of software (unfortunately).   J Almost all of the backlash against C is (I think) due to unreasonable fearL of UNIX by many legacy programmers who simply do not understand the language and fear change. --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_p I C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm    ------------------------------  $ Date: Wed, 7 Jun 2000 11:55:20 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures& Message-ID: <LUw%4.75$Rp5.1501@client>  @ "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageD news:Pine.LNX.4.05.10006071622410.13762-100000@Mufasa.pubserv.com...( > On Tue, 6 Jun 2000, Dann Corbit wrote: > I > > Produce some statistics or admit that you are completely full of crap  and E > > just spouting without a grain of knowlege.  The Y2K problems (for 	 instance) L > > were mostly COBOL.  The greatest amount of defects is COBOL, but that isE > > largely because it is also the greatest volume of code.  I have a  feeling K > > you are a blowhard.  Legacy programmers fear C programmers because they  are $ > > afraid of Unix (Tru64 included). > K > Could you define "legacy programmer," please?  You wouldn't be using that K > in the context of "anyone who doesn't happen to program in c," would you? I > I can tell you from experience that there are languages out there which : > meet and surpass the quality of c for most applications.  ? Someone maintaining code written long ago in whatever language. L Languages can be better for particular applications.  Fortran is pretty goodC at number crunching because it does not allow aliased pointers, for J instance.  You seem to think that there is such a thing as "the quality of, C" which is a profoundly ignorant statement.  ' The quality of C code is a function of: ! 1. Proper mapping of requirements  2. Proper design 3. Proper implementation 4. Proper testing  5. Proper documentation   L Just like any other languge.  Furthermore, it is a function of the talent ofL the programmer and the entire software development team, just like any other	 langauge.   J > That aside, why should "legacy programmers" (that's an ammusing term) be > afraid of unix?   L They should not be.  It is an unreasoning fear, but anyone who denies it has his head in the sand.   8 >(Has it occured to you that unix is an older technologyJ > than most things still in use today -- a few multics installations still- > around, and other things, but in general...  > L > Then, thirdly, please tell me why being afraid of unix would also make one > afraid of c programmers?  H It's an observation and a correct one.  People are afraid of Unix.  ThisL forum is absolutely struck with fear over it.  I can make a deja news searchI and prove the point, but someone who denies that doesn't care about facts L anyway since they are ignoring them to start with.  People also fear WindowsG NT.  I think both types of fear are silly and unreasonable and yet they L exist.  I'm not saying that *you* have such fear.  But many, many people do.  0 >(Before you answer, consider that c code can beL > built on anything from cp/m to MVS (VMS included, allthough it's not as ifH > the o/s were written in it... ;) these days)  I hope you don't mean toG > suggest that c has unbreakable ties to the unix o/s, since that would A > probably make it a worse choice to develop a new o/s in, right?   L Historically, a lot of C programming is done on Unix systems.  They all comeL with a C compiler for free [at least in general] which may be a contributingH factor.  If you have not observed this connection, then you are not very" observant.  I see it all the time.   > / > I may just be missing something, let me know.   ( Lots of things.  I hope the above helps.  L I'm not trying to bash OpenVMS.  It is one of my favorite operating systems.G I am not trying to bash COBOL or Fortran or BASIC or Bliss or Pascal or L PL/1.  Every tool has its place, and in the hands of a master it can produce a work of art.  I C bashing is done by people cringing in fear of something that they don't  understand.  --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_p I C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm    ------------------------------   Date: 7 Jun 2000 19:11:35 GMT # From: system@niuhep.physics.niu.edu 8 Subject: Re: C bashing (was Re: VMS File Caching Futures* Message-ID: <8hm6p7$ta$1@husk.cso.niu.edu>  / "Dann Corbit" <dcorbit@solutionsiq.com> writes: 8 >"Peter LANGSTOEGER" <eplan@kapsch.net> wrote in message  5 >The Y2K problems (for instance) were mostly COBOL.     J And would have been in whatever language was used in the 60-70s (-80s?) by$ cost conscious mainframe programmers   ------------------------------  $ Date: Wed, 7 Jun 2000 15:35:53 -0400' From: "Bill Todd" <billtodd@foo.mv.com> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures( Message-ID: <8hm80p$gh7$1@pyrite.mv.net>  D Larry Kilgallen <Kilgallen@eisner.decus.org.nospam> wrote in message% news:murWTdLG9YGg@eisner.decus.org... J > In article <009EB385.0A8C9D0B@SendSpamHere.ORG>, system@SendSpamHere.ORG' (Brian Schenkenberger, VAXman-) writes:  > I > > Clean your KLOK.  It really take very little effort to create quality I > > software.  It needs only a process and a desire to write quality s/w.  > G > No, it also requires time.  Some people are not given enough, perhaps D > because they are competing against those for whom quality is not a > significant priority.   E And that, IMHO (yes, once in a while it's humble), is the only really 7 significant observation made so far in this discussion.   L Just for the record, I like C, and I like to comment it.  Probably because IK feel the same way about assembler.  And come hell or high water, I take the3F design and coding time necessary to make sure that I don't have enoughE run-time bugs that having a pickier language would make a difference.   F That's a luxury not everyone has quite as much leeway (or chutzpah) toH enjoy, so other languages may be better for them - I wouldn't presume toI say, though for myself if I spent less time making sure I knew what I was:H doing, I'd be afraid that non-syntactical errors (that might well escapeD discovery at run time) might sneak in in greater numbers even if the language caught the others.e   - bill   ------------------------------  + Date: Wed, 07 Jun 2000 19:44:03 +0000 (   )n3 From: Christopher Smith <chriss@Mufasa.pubserv.com>u8 Subject: Re: C bashing (was Re: VMS File Caching FuturesJ Message-ID: <Pine.LNX.4.05.10006071904540.13762-100000@Mufasa.pubserv.com>  & On Wed, 7 Jun 2000, Dann Corbit wrote:  B > "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageM > > Could you define "legacy programmer," please?  You wouldn't be using thats  A > Someone maintaining code written long ago in whatever language.nN > Languages can be better for particular applications.  Fortran is pretty goodE > at number crunching because it does not allow aliased pointers, for L > instance.  You seem to think that there is such a thing as "the quality of. > C" which is a profoundly ignorant statement.  J Good enough.  I'd say that there is a "quality of C," which contributes toB the quality of the programs built with it.  Here's how I make that connection:S  6 There is a quality of programs, I'm sure you'll agree.  D I'm sure that you'd also agree that a compiler is a program like any1 other, and subject to varying degrees of quality.i  G Now, it follows that since a compiler can only follow specifications (IoH notice you stress proper design below -- think of this in the same way),A the quality of the compiler, and hence the quality of the code iswF dependant on the quality of the specification which will produce them.  J Note that I'm not saying that the quality of the C language is too low forJ one application or another, but that there is a quality factor involved in it, as with any other language.O  I AFAIK that's pretty sound reasoning, but let me know if you see a problemi with it.  H Let me say that I'm not going to suggest that the quality of any programI is completely dependant on the compiler -- that would be silly -- but the-I quality of the compiler, and the language's spec, play an important part.r  ) > The quality of C code is a function of:i# > 1. Proper mapping of requirementsa > 2. Proper design > 3. Proper implementation > 4. Proper testingV > 5. Proper documentatione  G No argument there.  People should also realize that any language in usei> aside from assembly (which is certainly error proned) has beenF developed using a lower-level (more error-proned...) language, so it'sH obviously not impossible to write airtight code in most languages, given& the patience and understanding for it.  I I don't think that this is what the argument is really about though.  ThepJ people arguing on the side of other languages simply believe that for someA (or any?) purpose, the language(s) they use is/are better than C.r  G Of course, C doesn't generate bad code by itself, it takes a programmer-E who doesn't write good code, but the argument is that it's simpler tom( write good code in some other languages.  N > Just like any other languge.  Furthermore, it is a function of the talent ofN > the programmer and the entire software development team, just like any other > langauge.a   err... yes, see above.. :)  N > They should not be.  It is an unreasoning fear, but anyone who denies it has > his head in the sand.u  F I've seen a dislike, and a resentment sometimes, but I'm not sure I've9 seen fear -- perhaps you interpret it differently than I.=  N > > Then, thirdly, please tell me why being afraid of unix would also make one > > afraid of c programmers?  J > It's an observation and a correct one.  People are afraid of Unix.  ThisN > forum is absolutely struck with fear over it.  I can make a deja news search  G Again, I've seen dislike and resentment, of course.  Honesly, though, I F think that it might be the other way around -- people who dislike unix will sometimes blame it on C.  e  I Certainly people who dislike one are more prone to dislike the other, butr> I'm not sure that it's a 100% correlation, or even a majority.  K > and prove the point, but someone who denies that doesn't care about factslN > anyway since they are ignoring them to start with.  People also fear WindowsI > NT.  I think both types of fear are silly and unreasonable and yet theyMN > exist.  I'm not saying that *you* have such fear.  But many, many people do.  H Oh, I certainly fear windows nt, in that I'm afraid somebody will try to% use it in a "smart" weapon one day ;).  G I think it's often not change that people are worried about.  For some,-H change might be the issue, but for myself (and several people I know whoJ do dislike unix and c -- I honestly like it for some things) it's an issue of quality.D  H It's unfair to say that language X has a larger number of buggy programsF written in it, and therefor is a bad language; this is in the same wayB it's unfair to say that people who dislike c also dislike unix. ;)   People still do this, though.N  N > Historically, a lot of C programming is done on Unix systems.  They all comeN > with a C compiler for free [at least in general] which may be a contributingJ > factor.  If you have not observed this connection, then you are not very$ > observant.  I see it all the time.  A You'd be suprised -- or maybe not -- a lot of the commercial unixo6 offerings charge lots of extra money for the compiler.  E It makes sense, though, since the entire system -- more or less -- isLI written in C, that a lot of new development on unix would also use C.  Ite@ doesn't, however, make sense that people would assume unix and C inseperable.  N > I'm not trying to bash OpenVMS.  It is one of my favorite operating systems.I > I am not trying to bash COBOL or Fortran or BASIC or Bliss or Pascal ordN > PL/1.  Every tool has its place, and in the hands of a master it can produce > a work of art.  D No argument there, though you might want to be careful with the termB "legacy" in this group ;) (Seriously, old doesn't necesarilly meanI not-as-good.  I've seen too much incredible technology that was developediI years ago, never caught on, and just recently has had a weak attempt at a)H re-implimentation with "newer" methods that completely missed the mark.)  K > C bashing is done by people cringing in fear of something that they don't 
 > understand.v  D I'm still not sure that a lot of this "bashing" is directed at the CH language, so much as it's directed (knowingly, or not) at the people whoH misuse it. In some cases, sure, people bash C for having null-terminatedC strings, or the like -- but all that really means is that for theiroF purposes another language could be better.  I, for one, wouldn't use aF null-terminated string to fetch a username or password ;) (not that itH couldn't be done in C, given the inclination, with a pascal string, or aI character array of a fixed size)  For that, another language might be thee better choice.  G Of course, like nearly anything else related to computers, it sometimeso? becomes a religeous issue, and then everything just goes crazy.n  E Personally I try to choose the best language available to me for each I given job, and not to fall into the trap of using language N because it's F the same language I've used the last six years.  Programming languagesJ are, in fact, so diverse, that I have no idea how anyone could help but do this.a   Regards,   Chrisb  O =============================================================================== @ "My two cents"			(http://rootworks.com/twocentsworth.cgi?128562)= Christopher Smith(chriss@pubserv.com)			Prgramer^W Programmer4 Prime Synergy of Champaign, IL.-% -------------------------------------7I "Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and H weighs 30 tons, computers in the future may have only 1,000 vacuum tubes; and weigh only 1.5 tons." -- Popular Mechanics, March 1949  O -------------------------------------------------------------------------------u   ------------------------------  $ Date: Wed, 7 Jun 2000 15:52:58 -0400' From: "Bill Todd" <billtodd@foo.mv.com>o8 Subject: Re: C bashing (was Re: VMS File Caching Futures( Message-ID: <8hm90u$k1s$1@pyrite.mv.net>  I 'Dangerous' is clearly in the eye of the beholder.  I don't consider yourgI examples 'dangerous' for my own use of C, because they're not the kind of-E errors I make.  People who do make that kind of error would likely benK well-advised to use a different language, or at least compiler options thatg( help catch them (if such options exist).  D And if I were an application developer, I might want a language thatL provided better support for some of the application concerns that just don'tK crop up in system-level development.  Or not:  there's something to be saidaI for sticking with a language and coding style you're comfortable with, as H long as it works well for you (and it's really hard to be sure that someH other choice would be better for whoever may inherit the code later:  ifH someone's going to screw up what I wrote, I'd at least like it not to be6 because they don't know the language it's written in).   - bill  . John Vottero <John@MVPSI.com> wrote in message4 news:C15945A9D9EFCF11BA8B08002BBF1CCC0CD71A@BERRY... > >f > > Dan Sugalski wrote:iA > > > Right, but being forced to care means that if you miss, your > > lose. Computer@ > > > languages are supposed to help the programmer avoid making > > mistakes. C H > > > requires the use of some constructs that are terribly error-prone. > >e > > requires ? ? ? ? ? ? > >c> > > Perhaps you mean "C code that can be ported to Unix" or "C > > code that wasw: > > written on Windows". But I disagree that "C" by itself$ > > "requires" dangerous constructs. > >h >p > How about: >  >   if (xyz = abc)	 >       {  >       dothis(); 	 >       }  >t > or maybe:  >t >    for (i=0; i<100; i++);r	 >       {u >       do that();	 >       }r >v > or maybe this: >  >    /* ) >    **  C comments keep going and going!s >    *\  <---oops!' >    printf("This will never print\n");  >s >    /*'" >    **  And, there are no errors! >    */yC >    printf("C has lots of problems, that's why there's C++ :)"\n);l >- >u >h   ------------------------------  " Date: Wed, 7 Jun 2000 19:33:19 GMT- From: "Richard D. Piccard" <piccard@ohio.edu> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures( Message-ID: <393EA374.843E8D95@ohio.edu>   Dann Corbit wrote: >    [snip]   > L > Almost all of the backlash against C is (I think) due to unreasonable fearN > of UNIX by many legacy programmers who simply do not understand the language > and fear change.  @ While I am sure that fear of unix is responsible for some of theF backlash against C, I find "almost all" a bit hard to swallow.  I, forE one, dislike every implementation of C that I have worked with (whichtG does not include any of the DEQ C compilers), because they have all hade@ artificial limitations on the size of static arrays that I couldF declare.  I was *FORCED* to use dynamic memory allocation and pointersF explicitly for large arrays, even though they fit into memory and wereE going to be used throughout program execution, rather than having thet< compiler do that housekeeping for me.  The ability to do theG housekeeping was in the compiler -- I could declare small static arrayssE -- but it was implemented in a way that only worked for small arrays,sH rather than being capable of use for any size array.  I think that couldD have been solved by a language definition that required a conformingC compiler to included full size static arrays, instead of permittingfB conforming compilers to specify a maximum size for static arrays.   B It also strikes me as particularly lame that the default parameterE passing mechanism (reference or value) depends on the variable type.  B That is a language definition problem, inherited from the originalH effort, and unfixable because of the need to maintain compatibility with existing code.    	 						RDP  -- eB ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  + Date: Wed, 07 Jun 2000 18:34:22 +0000 (   ).3 From: Christopher Smith <chriss@Mufasa.pubserv.com> 8 Subject: Re: C bashing (was Re: VMS File Caching FuturesJ Message-ID: <Pine.LNX.4.05.10006071827310.13762-100000@Mufasa.pubserv.com>  # On Wed, 7 Jun 2000, it was written:I  B > "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in message  H > > So, in short, no thanks, I'd rather design my own gui than have this* > > (poor) compiler try to modify my code.  F > > I have a pet peeve with visual c++, so sorry if I'm ranting.  Your3 > > original point is good, and I don't dispute it.s  N > I used the MS compiler as my sample because I don't have any experience withH > the GCC or VMS C compilers.  I had tried older compilers for differentN > systems, and they simply stunk.  As for the junk the MS Compiler Application  I Things change, I guess... I do happen to have (don't shoot me!) a machineeH at home that runs the original (more or less) unix system V r 3.0 stuff,J including the compiler.  I wouldn't say that it stinks, but it's certainly? not a walk in the park to write for it, so I see what you mean.e  G Having no experience with older dec c compilers, I can't say much aboute those.  M > Wizards put in, it does get in the way sometimes, but on the whole, it does F > a good job of handling the messy stuff in MS-Windows programming andI > includes debug build checks for invalid arguments in almost every call,aM > which definitely helps catch pointer errors that the compiler itself can't.o  F It's that messy-stuff-handling feature that's bitten me more than onceJ ;)... basically I've had experiences where the powers-that-be want the guiF redesigned several times over, and finally the thing forgets to removeE some of its code that's no longer valid due to changes.  The code, ofhG course gets thrown in at compile time, and causes a run-time exception.sH The source-level debugger won't find it because it's in a comment...etc, etc.  J With any luck, of course, that's been fixed by now, but somehow I'm not so sure of it.t   Regards,   Chris   O =============================================================================== @ "My two cents"			(http://rootworks.com/twocentsworth.cgi?128562)= Christopher Smith(chriss@pubserv.com)			Prgramer^W Programmerd Prime Synergy of Champaign, IL.e% -------------------------------------eI "Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes andtH weighs 30 tons, computers in the future may have only 1,000 vacuum tubes; and weigh only 1.5 tons." -- Popular Mechanics, March 1949 cO -------------------------------------------------------------------------------a   ------------------------------  % Date: Wed, 07 Jun 2000 16:10:47 -0400t- From: JF Mezei <jfmezei.spamnot@videotron.ca>.8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EAC46.ACEFAF39@videotron.ca>  & "Brian Schenkenberger, VAXman-" wrote:K > ficult and understanding the construct even more difficult.  In addition, J > you're very likely NOT going to find a comment explaining the 'expected'L > function of such code.  Many a C programmer seems to be from a school thatC > C is a self-documenting language obviating the need for comments.a  M I once had to deal with a pretty smart C programmer who had written all sorts=L of screen handling routines for DOS. Not only was there no documentation forL HIS routines, but he didn't have proper storage/names for the libraries thatJ contained his routines, and not all his libraries were compaible with eachJ other. But *HE* knew which ones were compiles recently and were up to date4 and, of course, he had never had problems with them.  M I looked like an idiot because I could not get his routines to work, while hen never had problems.Q  K This is not because it was written in C, it was because it was written by asM hacker who loved to do fancy stuff such as redefining printf etc etc etc. ButnM the guy did produce code fast by himself. While I didn't get good marks while.H working there, when they did hire a extra person to work with that smartL hacker, the management realised what had been going on, and the smart hackerN was let go. It had gotten to a point where he was purposefully making his codeH difficult to understand and use as a job security feature to prevent HIS6 work/program from being maintained/improved by others.  I Just because the new crop of programmers are learning on C or worse, C++,tL doesn't mean that the language is bad. It means that their education is bad.   ------------------------------  # Date: Wed, 07 Jun 2000 20:19:50 GMTf4 From: "Michael D. Ober" <mdo.@.wakeassoc.com.nospam>8 Subject: Re: C bashing (was Re: VMS File Caching FuturesD Message-ID: <G7y%4.4266$2X2.156422@newsread2.prod.itd.earthlink.net>  @ "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageD news:Pine.LNX.4.05.10006071904540.13762-100000@Mufasa.pubserv.com...( > On Wed, 7 Jun 2000, Dann Corbit wrote: >hD > > "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageJ > > > Could you define "legacy programmer," please?  You wouldn't be using that >v   <<<<< CLIPPED >>>>>>>>  G > > and prove the point, but someone who denies that doesn't care aboutn factsrH > > anyway since they are ignoring them to start with.  People also fear Windows K > > NT.  I think both types of fear are silly and unreasonable and yet theytL > > exist.  I'm not saying that *you* have such fear.  But many, many people do.r >pJ > Oh, I certainly fear windows nt, in that I'm afraid somebody will try to' > use it in a "smart" weapon one day ;)a  L The US Navy tried a year or so ago.  Their "smart" weapon was a warship thatG ended up drifting at sea (no pun intendend) for several hours while thepF software was fixed.  Note this wasn't strictly an NT bug, but rather aH poorly coded application running on the MS-SQL Server.  NT itself didn'tF crash, but the SQL server database was so badly corrupted it had to beI restored from backup.  An unintialized variable shut down the whole ship."I NT isn't a real time OS, so it isn't the right choice for a weapon.  I'vecJ seen VMS clusters can made into real-time systems by appropriate tweaking.   >.
 > Regards, >- > Chris- >- >-L ============================================================================ ===s@ > "My two cents" (http://rootworks.com/twocentsworth.cgi?128562)= > Christopher Smith(chriss@pubserv.com) Prgramer^W Programmer ! > Prime Synergy of Champaign, IL.n' > -------------------------------------1K > "Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and:J > weighs 30 tons, computers in the future may have only 1,000 vacuum tubes< > and weigh only 1.5 tons." -- Popular Mechanics, March 1949L > -------------------------------------------------------------------------- -----c >i	 Mike Obert   ------------------------------   Date: 7 Jun 2000 20:11:42 GMTe* From: helbig@astro.rug.nl (Phillip Helbig)8 Subject: Re: C bashing (was Re: VMS File Caching Futures. Message-ID: <8hma9u$6t9$3@info.service.rug.nl>  6 In article <8hm10u$p5u$1@mailint03.im.hou.compaq.com>,5 hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) writes: s  A >  largely for various aspects of maintainability.  MAIL was one iF >  such area, and has been ported over -- and as Larry comments, some D >  of the more indiosyncratic and obscure behaviour of the old MAIL A >  utility was not correctly emulated, hence various of the MAIL cB >  problem reports, and hence the inclusion of MAIL/OLD.  Some of   F As the originator of this thread, the reason I use MAIL/OLD is that I G often SPAWN out of EDT.  DEFINE KEY GOLD CON Z AS "XLATESPAWN Z^ZREF." iE gives a "command tables have invalid format" if callable EDT is used e! from MAIL as opposed to MAIL/OLD.e   ------------------------------  $ Date: Wed, 7 Jun 2000 13:41:03 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com>m8 Subject: Re: C bashing (was Re: VMS File Caching Futures' Message-ID: <Ery%4.100$Rp5.1953@client>s  8 "Richard D. Piccard" <piccard@ohio.edu> wrote in message" news:393EA374.843E8D95@ohio.edu... >e > Dann Corbit wrote: > >  >h > [snip] >  > >II > > Almost all of the backlash against C is (I think) due to unreasonabled fearG > > of UNIX by many legacy programmers who simply do not understand then language > > and fear change. >iB > While I am sure that fear of unix is responsible for some of theH > backlash against C, I find "almost all" a bit hard to swallow.  I, forG > one, dislike every implementation of C that I have worked with (whichcI > does not include any of the DEQ C compilers), because they have all hadeB > artificial limitations on the size of static arrays that I couldH > declare.  I was *FORCED* to use dynamic memory allocation and pointersH > explicitly for large arrays, even though they fit into memory and wereG > going to be used throughout program execution, rather than having thea> > compiler do that housekeeping for me.  The ability to do theI > housekeeping was in the compiler -- I could declare small static arrays/G > -- but it was implemented in a way that only worked for small arrays,iJ > rather than being capable of use for any size array.  I think that couldF > have been solved by a language definition that required a conformingE > compiler to included full size static arrays, instead of permittingpC > conforming compilers to specify a maximum size for static arrays.   G If you make it static, it is not such a problem.  Large auto arrays are   problems on some implementation.  D > It also strikes me as particularly lame that the default parameterF > passing mechanism (reference or value) depends on the variable type.D > That is a language definition problem, inherited from the originalJ > effort, and unfixable because of the need to maintain compatibility with > existing code.   ?Puzzle?  K In C, absolutely everything is passed by value.  There are no references inoL C.  Even when using a pointer, a copy of the pointer is passed to the called	 function.e --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_pmI C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm0   ------------------------------  $ Date: Wed, 7 Jun 2000 13:46:29 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com>c8 Subject: Re: C bashing (was Re: VMS File Caching Futures' Message-ID: <Lwy%4.103$Rp5.1817@client>D  @ "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageD news:Pine.LNX.4.05.10006071904540.13762-100000@Mufasa.pubserv.com... [mega-snippage]B    I agree on the following points:  F 1.  There is definitely some really bad C code, written by incompetent. nincompoops.  I have seen it with my own eyes.  H 2.  Use the right tool for the right job.  For some jobs, the C language6 *is* a square peg trying to cram through a round hole.  H 3.  I don't dislike legacy systems nor denigrate their use.  Legacy alsoD stands for "Thoroughly debugged and has withstood the test of time."  J I actually think that a large portion of the time, it is a serious mistake- to abandon a legacy system for something new.r --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_pwI C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htmi   ------------------------------  % Date: Wed, 07 Jun 2000 16:57:30 -0400u- From: JF Mezei <jfmezei.spamnot@videotron.ca> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EB736.1117E921@videotron.ca>   Dann Corbit wrote:M > In C, absolutely everything is passed by value.  There are no references intN > C.  Even when using a pointer, a copy of the pointer is passed to the called > function.a  M But isn't this how it really works when the high level language is translatedaL into machine code anyways ? When you push arguments onto the stack, isn't it7 in fact a copy of the argument that goes on the stack ?    ------------------------------  + Date: Wed, 07 Jun 2000 20:55:58 +0000 (   ):3 From: Christopher Smith <chriss@Mufasa.pubserv.com>s8 Subject: Re: C bashing (was Re: VMS File Caching FuturesJ Message-ID: <Pine.LNX.4.05.10006072050060.13762-100000@Mufasa.pubserv.com>  & On Wed, 7 Jun 2000, Dann Corbit wrote:  H > 1.  There is definitely some really bad C code, written by incompetent0 > nincompoops.  I have seen it with my own eyes.  J Well, I'm sure -- almost sure -- some of my old co-workers could give your/ incompetent aquaintences a run for their money.i  L > I actually think that a large portion of the time, it is a serious mistake/ > to abandon a legacy system for something new.e  I It's too bad more people don't think that way, instead of responding in aaC lemming-like fashion to whatever the newest trend is by immediatelyl" throwing everything else out... :)  E It's also too bad that more people can't agree that each language haseD strong-points, and leave the decision on which language to use to beA decided on the basis of individual need. (If all programmers weret& competent, that would be enough ... ;)   Regards,   Chrism  O ===============================================================================i@ "My two cents"			(http://rootworks.com/twocentsworth.cgi?128562)= Christopher Smith(chriss@pubserv.com)			Prgramer^W Programmer  Prime Synergy of Champaign, IL. % -------------------------------------eI "Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and?H weighs 30 tons, computers in the future may have only 1,000 vacuum tubes; and weigh only 1.5 tons." -- Popular Mechanics, March 1949 eO -------------------------------------------------------------------------------:   ------------------------------   Date: 7 Jun 2000 20:51:14 GMT * From: helbig@astro.rug.nl (Phillip Helbig)8 Subject: Re: C bashing (was Re: VMS File Caching Futures. Message-ID: <8hmck2$7so$1@info.service.rug.nl>  5 In article <393EAC46.ACEFAF39@videotron.ca>, JF Mezeit' <jfmezei.spamnot@videotron.ca> writes: /  L >This is not because it was written in C, it was because it was written by aN >hacker who loved to do fancy stuff such as redefining printf etc etc etc. ButN >the guy did produce code fast by himself. While I didn't get good marks whileI >working there, when they did hire a extra person to work with that smartgM >hacker, the management realised what had been going on, and the smart hacker O >was let go. It had gotten to a point where he was purposefully making his codeUI >difficult to understand and use as a job security feature to prevent HIS 7 >work/program from being maintained/improved by others.r  G There's a line in the UNIX HATER'S HANDBOOK which says "when your unix a: sysadmin mentions `security', he's talking about his job".   ------------------------------  " Date: Wed, 7 Jun 2000 22:32:05 GMT9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)s8 Subject: Re: C bashing (was Re: VMS File Caching Futures+ Message-ID: <q4kjgCMILrMa@eisner.decus.org>n  V In article <LUw%4.75$Rp5.1501@client>, "Dann Corbit" <dcorbit@solutionsiq.com> writes:B > "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageF > news:Pine.LNX.4.05.10006071622410.13762-100000@Mufasa.pubserv.com...  L >> Could you define "legacy programmer," please?  You wouldn't be using thatL >> in the context of "anyone who doesn't happen to program in c," would you?J >> I can tell you from experience that there are languages out there which; >> meet and surpass the quality of c for most applications.d > A > Someone maintaining code written long ago in whatever language.s  A Therefore, someone maintaining code which has enough quality thataB there has been no need to discard it.  That is, maintainable code.   ------------------------------  " Date: Wed, 7 Jun 2000 21:09:48 GMT- From: "Richard D. Piccard" <piccard@ohio.edu> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures( Message-ID: <393EBA10.530974B2@ohio.edu>   Dann Corbit wrote: > : > "Richard D. Piccard" <piccard@ohio.edu> wrote in message$ > news:393EA374.843E8D95@ohio.edu...   [snip]  F > > It also strikes me as particularly lame that the default parameterH > > passing mechanism (reference or value) depends on the variable type.F > > That is a language definition problem, inherited from the originalL > > effort, and unfixable because of the need to maintain compatibility with > > existing code. > 
 > ?Puzzle? > M > In C, absolutely everything is passed by value.  There are no references in>N > C.  Even when using a pointer, a copy of the pointer is passed to the called > function.   F Yes, but passing a pointer by value is effectively the same as passing the data by reference. p  	 						RDP  -- oB ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  % Date: Wed, 07 Jun 2000 17:22:00 -0400 * From: David A Froble <davef@tsoft-inc.com>8 Subject: Re: C bashing (was Re: VMS File Caching Futures- Message-ID: <393EBCF8.E96911AE@tsoft-inc.com>o   "Michael D. Ober" wrote: > N > As to why C programs have more flaws than BLISS - simple - there's more codeJ > in C than in BLISS.  It would be interesting to know how much of the new+ > code in VMS is actually ported from Unix.n >  > Mike Ober.  M No!  The reason C programs have more flaws than other languages is that the C N compiler will not catch simple mistakes nearly as well as many other compilersO do.  Programmers are human, and will make simple typing mistakes, transpose twohO characters, and such.  Not all are catchable, but every time a compiler catchessP a simple mistake, one of two things happen.  Either the programmer is spared theM effort of tracking down the mistake, or, the mistake makes it into productionl# where really bad things can happen.r  P My favorite bitch, and some of you will not agree, is doing a test, ie; ( if a =N b  vs  if a == b ) and in addition to performing the test, one variable is setN equal to the other.  Things such as this are real easy to do, and real easy toN miss when looking for the error.  (Disclaimer, not sure of the syntax, but youL get the idea.)  Not even going to get into null terminated strings and such.   Dave   -- =4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Wed, 07 Jun 2000 17:31:31 -0400e* From: David A Froble <davef@tsoft-inc.com>8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EBF33.4319353@tsoft-inc.com>   Dann Corbit wrote: > K > C bashing is done by people cringing in fear of something that they don't 
 > understand.u  G Or maybe it gets bashed more than others because it truly deserves it. j	 Possible?b  L Dave, who's not in fear of C, Unix, or NT, but reserves the right to despise them.    -- n4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Wed, 07 Jun 2000 17:39:20 -0400v* From: David A Froble <davef@tsoft-inc.com>8 Subject: Re: C bashing (was Re: VMS File Caching Futures- Message-ID: <393EC108.2627B9D5@tsoft-inc.com>.  L > > Oh, I certainly fear windows nt, in that I'm afraid somebody will try to) > > use it in a "smart" weapon one day ;)t  P I just hope it's aimed at me.  Thaat's the only sure way I can be assured NOT to be hit. :-)h   Dave   -- i4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  $ Date: Wed, 7 Jun 2000 16:41:59 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com>n8 Subject: Re: C bashing (was Re: VMS File Caching Futures' Message-ID: <e5B%4.139$Rp5.2449@client>e  7 "David A Froble" <davef@tsoft-inc.com> wrote in messagek' news:393EBCF8.E96911AE@tsoft-inc.com...  > "Michael D. Ober" wrote: > >aK > > As to why C programs have more flaws than BLISS - simple - there's more  codeL > > in C than in BLISS.  It would be interesting to know how much of the new- > > code in VMS is actually ported from Unix.  > >b > > Mike Ober. > I > No!  The reason C programs have more flaws than other languages is thatg the C F > compiler will not catch simple mistakes nearly as well as many other	 compilerseC > do.  Programmers are human, and will make simple typing mistakes,a
 transpose two>I > characters, and such.  Not all are catchable, but every time a compilery catchesXG > a simple mistake, one of two things happen.  Either the programmer is@
 spared theD > effort of tracking down the mistake, or, the mistake makes it into
 production% > where really bad things can happen.n >rI > My favorite bitch, and some of you will not agree, is doing a test, ie;n  if a = L > b  vs  if a == b ) and in addition to performing the test, one variable is setvH > equal to the other.  Things such as this are real easy to do, and real easy todL > miss when looking for the error.  (Disclaimer, not sure of the syntax, but you H > get the idea.)  Not even going to get into null terminated strings and such.e  & A good workman never blames his tools.   int main(void) {a     int a = 1;     int b = 2;     if (a = b) a = 3;      if (a == b) b = 4;
     return 0;a }i   --- Module:   foo.c               _     if (a = b) a = 3;a/ foo.c(5) : Info 720: Boolean test of assignmentmA foo.c(5) : Info 774: Boolean within 'if' always evaluates to Truet               _a     if (a == b) b = 4;B foo.c(6) : Info 774: Boolean within 'if' always evaluates to False ---y  I There are diagnostic tools and utilities that will catch gaffes like thistH without any difficulty.  An experienced C programmer will spot somethingH like this with the eye immediately, but in case he should not, there are automatic solutions. --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_p I C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm=   ------------------------------  $ Date: Wed, 7 Jun 2000 16:43:10 -0700- From: "Dann Corbit" <dcorbit@solutionsiq.com>e8 Subject: Re: C bashing (was Re: VMS File Caching Futures' Message-ID: <m6B%4.140$Rp5.2397@client>-  7 "David A Froble" <davef@tsoft-inc.com> wrote in messaget& news:393EBF33.4319353@tsoft-inc.com... > Dann Corbit wrote: > >rG > > C bashing is done by people cringing in fear of something that theyr don'ti > > understand.  >-H > Or maybe it gets bashed more than others because it truly deserves it. > Possible?e  1 Anything is possible.  But you are wrong about C.u  F > Dave, who's not in fear of C, Unix, or NT, but reserves the right to despises > them.   L Anyone has the right to despise anything they like.  A carpenter can despise0 a hammer.  It won't make him a better carpenter. --0 C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html$  "The C-FAQ Book" ISBN 0-201-84519-91 C.A.P. Newsgroup   http://www.dejanews.com/~c_a_paI C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htmi   ------------------------------  % Date: Wed, 07 Jun 2000 19:15:06 -0500r* From: Keith Brown <kbrown780@usfamily.net>8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EE58A.1F856B63@usfamily.net>   Christopher Smith wrote: > ( > On Wed, 7 Jun 2000, Dann Corbit wrote: > D > > "Christopher Smith" <chriss@Mufasa.pubserv.com> wrote in messageO > > > Could you define "legacy programmer," please?  You wouldn't be using that- > C > > Someone maintaining code written long ago in whatever language.DP > > Languages can be better for particular applications.  Fortran is pretty goodG > > at number crunching because it does not allow aliased pointers, foriN > > instance.  You seem to think that there is such a thing as "the quality of0 > > C" which is a profoundly ignorant statement. > L > Good enough.  I'd say that there is a "quality of C," which contributes toD > the quality of the programs built with it.  Here's how I make that
 > connection:a > 8 > There is a quality of programs, I'm sure you'll agree. > F > I'm sure that you'd also agree that a compiler is a program like any3 > other, and subject to varying degrees of quality.y > I > Now, it follows that since a compiler can only follow specifications (I J > notice you stress proper design below -- think of this in the same way),C > the quality of the compiler, and hence the quality of the code istH > dependant on the quality of the specification which will produce them. > L > Note that I'm not saying that the quality of the C language is too low forL > one application or another, but that there is a quality factor involved in! > it, as with any other language.i > K > AFAIK that's pretty sound reasoning, but let me know if you see a problemm
 > with it. > J > Let me say that I'm not going to suggest that the quality of any programK > is completely dependant on the compiler -- that would be silly -- but thenK > quality of the compiler, and the language's spec, play an important part.a > + > > The quality of C code is a function of:s% > > 1. Proper mapping of requirementsr > > 2. Proper design > > 3. Proper implementation > > 4. Proper testingy > > 5. Proper documentatione > I > No argument there.  People should also realize that any language in uset@ > aside from assembly (which is certainly error proned) has beenH > developed using a lower-level (more error-proned...) language, so it'sJ > obviously not impossible to write airtight code in most languages, given( > the patience and understanding for it. > K > I don't think that this is what the argument is really about though.  ThenL > people arguing on the side of other languages simply believe that for someC > (or any?) purpose, the language(s) they use is/are better than C.  > I > Of course, C doesn't generate bad code by itself, it takes a programmer.G > who doesn't write good code, but the argument is that it's simpler to * > write good code in some other languages. > P > > Just like any other languge.  Furthermore, it is a function of the talent ofP > > the programmer and the entire software development team, just like any other
 > > langauge.r >  > err... yes, see above.. :) > P > > They should not be.  It is an unreasoning fear, but anyone who denies it has > > his head in the sand.B > H > I've seen a dislike, and a resentment sometimes, but I'm not sure I've; > seen fear -- perhaps you interpret it differently than I.t > P > > > Then, thirdly, please tell me why being afraid of unix would also make one > > > afraid of c programmers? > L > > It's an observation and a correct one.  People are afraid of Unix.  ThisP > > forum is absolutely struck with fear over it.  I can make a deja news search > I > Again, I've seen dislike and resentment, of course.  Honesly, though, IrH > think that it might be the other way around -- people who dislike unix > will sometimes blame it on C.o > K > Certainly people who dislike one are more prone to dislike the other, bute@ > I'm not sure that it's a 100% correlation, or even a majority. > M > > and prove the point, but someone who denies that doesn't care about facts0P > > anyway since they are ignoring them to start with.  People also fear WindowsK > > NT.  I think both types of fear are silly and unreasonable and yet theyoP > > exist.  I'm not saying that *you* have such fear.  But many, many people do. > J > Oh, I certainly fear windows nt, in that I'm afraid somebody will try to' > use it in a "smart" weapon one day ;)a > I > I think it's often not change that people are worried about.  For some,eJ > change might be the issue, but for myself (and several people I know whoL > do dislike unix and c -- I honestly like it for some things) it's an issue
 > of quality.- > J > It's unfair to say that language X has a larger number of buggy programsH > written in it, and therefor is a bad language; this is in the same wayD > it's unfair to say that people who dislike c also dislike unix. ;) >  > People still do this, though.  > P > > Historically, a lot of C programming is done on Unix systems.  They all comeP > > with a C compiler for free [at least in general] which may be a contributingL > > factor.  If you have not observed this connection, then you are not very& > > observant.  I see it all the time. > C > You'd be suprised -- or maybe not -- a lot of the commercial unixa8 > offerings charge lots of extra money for the compiler. > G > It makes sense, though, since the entire system -- more or less -- isfK > written in C, that a lot of new development on unix would also use C.  ItlB > doesn't, however, make sense that people would assume unix and C > inseperable. > P > > I'm not trying to bash OpenVMS.  It is one of my favorite operating systems.K > > I am not trying to bash COBOL or Fortran or BASIC or Bliss or Pascal or P > > PL/1.  Every tool has its place, and in the hands of a master it can produce > > a work of art. > F > No argument there, though you might want to be careful with the termD > "legacy" in this group ;) (Seriously, old doesn't necesarilly meanK > not-as-good.  I've seen too much incredible technology that was developedeK > years ago, never caught on, and just recently has had a weak attempt at a J > re-implimentation with "newer" methods that completely missed the mark.) > M > > C bashing is done by people cringing in fear of something that they don't- > > understand.- > F > I'm still not sure that a lot of this "bashing" is directed at the CJ > language, so much as it's directed (knowingly, or not) at the people whoJ > misuse it. In some cases, sure, people bash C for having null-terminatedE > strings, or the like -- but all that really means is that for their.H > purposes another language could be better.  I, for one, wouldn't use aH > null-terminated string to fetch a username or password ;) (not that itJ > couldn't be done in C, given the inclination, with a pascal string, or aK > character array of a fixed size)  For that, another language might be ther > better choice. > I > Of course, like nearly anything else related to computers, it sometimeseA > becomes a religeous issue, and then everything just goes crazy.I > G > Personally I try to choose the best language available to me for eachtK > given job, and not to fall into the trap of using language N because it'saH > the same language I've used the last six years.  Programming languagesL > are, in fact, so diverse, that I have no idea how anyone could help but do > this.i > 
 > Regards, >  > Chris- > Q > ===============================================================================cQ > "My two cents"                  (http://rootworks.com/twocentsworth.cgi?128562)FO > Christopher Smith(chriss@pubserv.com)                   Prgramer^W Programmer1! > Prime Synergy of Champaign, IL.@' > -------------------------------------SK > "Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and6J > weighs 30 tons, computers in the future may have only 1,000 vacuum tubes< > and weigh only 1.5 tons." -- Popular Mechanics, March 1949Q > -------------------------------------------------------------------------------   : I can only speak for myself so here's my 2 cents.  I don't? dislike C, rather I become quickly frustrated with the PHMs ando? kids who believe that using a standard language (they define as > C) is preferable to using the appropriate language. I wouldn't9 anymore recommend writing an accounting package in C thann> writing a compiler in COBOL, although I am sure both have been done before.   -- r Keith BrownD kbrown780@usfamily.net   ------------------------------  % Date: Wed, 07 Jun 2000 19:18:25 -0500D* From: Keith Brown <kbrown780@usfamily.net>8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EE651.892137A8@usfamily.net>  $ system@niuhep.physics.niu.edu wrote: > 1 > "Dann Corbit" <dcorbit@solutionsiq.com> writes: : > >"Peter LANGSTOEGER" <eplan@kapsch.net> wrote in message > 5 > >The Y2K problems (for instance) were mostly COBOL.t > L > And would have been in whatever language was used in the 60-70s (-80s?) by& > cost conscious mainframe programmers  ; And the rash of Y2K releated failures should says somethingl about this.e   -- e Keith Brown  kbrown780@usfamily.net   ------------------------------  % Date: Wed, 07 Jun 2000 19:30:13 -0500b* From: Keith Brown <kbrown780@usfamily.net>8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EE915.D1055291@usfamily.net>   Dann Corbit wrote: > L > "Wayne Sewell" <wayne@tachysoft.xxx.044962.killspam.0138> wrote in message* > news:mm0UqdQic5Db@tachxxsoftxxconsult... > [snip]I > > It had to be *ported*; it did not necessarily have to be *converted*.y > VeryK > > little of the vms source code written in macro had to be converted to aaG > > different language to support alpha, if any.   Some of it had to bei
 > modifiedJ > > because of vaxisms in the code, but it did *not* have to be converted.M > > Digital basically created a macro *compiler*, treating vax macro like anys > othereN > > computer language, including even optimization.   In other words, a single > vax N > > instruction produces many alpha machine instructions.  Contrast this to anK > > assembler, in which there is a one-to-one correspondence between sourcecG > > instruction and machine instruction.   From the viewpoint of a riscs > processor,L > > a cisc instruction set *is* a high level language.  In any case, much of > theeL > > vms source code is *still* macro, built with the vax *assembler* and the > alphap. > > *compiler* from basically the same source. > >r > >gI > > You accused someone else in this thread of pulling information out ofb > their K > > ass.  It appears you are doing the same.  You simply assumed all of the  > macro J > > code had to be converted to another language, and then you assumed theL > > conversion was to C.  If conversion had actually been required, it could > have! > > just as easily been to Bliss.r > H > Quite right, I have been taken to task on this one, and found wanting. > C > I do suspect (without proof) ;-) that Tru64 Unix is written in C. + > Would you describe this as a robust tool?0 > L > And consider the C that is used to construct OpenVMS. It is 1/3 C or 1/4 CI > or something like that, at any rate.  Is it (therefore) not robust?  IseN > OpenVMS not to be trusted as a secure operating system?  Should we switch to > IBM hardware and MVS?  > K > The point remains, bad software is written in every language.  It is trueFN > that you can do truly stupid things in C.  But that is not a language fault.K > It is a programmer fault.  Further, there are stupid programmers in every ) > discipline of software (unfortunately).o > L > Almost all of the backlash against C is (I think) due to unreasonable fearN > of UNIX by many legacy programmers who simply do not understand the language > and fear change. > --2 > C-FAQ: http://www.eskimo.com/~scs/C-faq/top.html& >  "The C-FAQ Book" ISBN 0-201-84519-93 > C.A.P. Newsgroup   http://www.dejanews.com/~c_a_ptK > C.A.P. FAQ: ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm-  = With all due respect I believe that most people who dislike Cb? simply believe that programmers (being human) make mistakes ande: prefer a language that will help them catch their mistakes@ earlier.  Yes, I agree that there are some very good programmers9 out there that can and do write excellent quality C code.;= Mortals like myself prefer something else.  There are expertsI9 after all that are very proficient at reading and writing > hieroglyphics, but does that mean that is the language I would# use to write my mother a letter? :)  -- l Keith Brownc kbrown780@usfamily.net   ------------------------------  % Date: Wed, 07 Jun 2000 19:32:39 -0500 * From: Keith Brown <kbrown780@usfamily.net>8 Subject: Re: C bashing (was Re: VMS File Caching Futures+ Message-ID: <393EE9A7.E4EE63D@usfamily.net>a   Larry Kilgallen wrote: > r > In article <009EB385.0A8C9D0B@SendSpamHere.ORG>, system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) writes: > I > > Clean your KLOK.  It really take very little effort to create quality I > > software.  It needs only a process and a desire to write quality s/w.C > G > No, it also requires time.  Some people are not given enough, perhapseD > because they are competing against those for whom quality is not a > significant priority.h  : It has been said that with software you have the optios of) "Fast, Good, and Cheap - Choose any two")    --   Keith Browno kbrown780@usfamily.net   ------------------------------  % Date: Wed, 07 Jun 2000 21:44:06 -0400h- From: JF Mezei <jfmezei.spamnot@videotron.ca> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures, Message-ID: <393EFA64.A75318DA@videotron.ca>   re:  if (a = b) c = 4 ;   I I am sorry, but the most serious problems I have encountered in C are not L these but rather very insidious ones that occur during the execution and areU not related to syntax. They are truly programming programs and not language specific.f  L For instance, I developped and tested a server on a Microvax II using DECNETI objects. The expected behaviour , and the one which testing on my machineo validated was :n   get connect requestb+ get a data packet (queue it for processing) E get a disconnect request from client and deallocate the structures ifa transaction was empty or done.  M However, at one customer, the disconnect request from the client was received I before the data packet. So the software would de-allocate the still empty H transaction packet, while the AST that would have filled the transactionG packet was next in the AST queue. The later executed fine except it was T writing in memory which was deallocated. And this would cause problems eventually...  F This was not a C problem, it was a behaviour problem because I had notI foreseen a case where the disconnect request would arrive before the datap/ packet that had been sent before disconnecting.   L Problems such as if (a=b) are fairly simple to isolate because you have code8 that executes even though the condition should be false.  N also, SEARCH foo.c if /window=3 does a pretty goo job of giving you a synopsys of potential IF problems.2   ------------------------------   Date: 7 Jun 2000 21:23:12 CDT = From: wayne@tachysoft.xxx.044962.killspam.0138 (Wayne Sewell)b8 Subject: Re: C bashing (was Re: VMS File Caching Futures. Message-ID: <mwWpwe7r0mNy@tachxxsoftxxconsult>  R In article <8hm90u$k1s$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:K > 'Dangerous' is clearly in the eye of the beholder.  I don't consider yournK > examples 'dangerous' for my own use of C, because they're not the kind of  > errors I make.    M Never?  Never ever?  Can you honestly sit there and say that not once in youriM professsional life you have ever made one of the errors shown?  No typos ("=" J when you meant "==") ever???  If that is truly the case, I am impressed asO hell.  comp.os.vms has been visited by deity and we must prostrate outselves in' awestruck wonder.  Amen.       [stuff deleted]/   -- iO ===============================================================================pM Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachysoft.xxx2: http://www.tachysoft.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)uO ===============================================================================DC Jake Blues: "Sell me your children!  How much for the little girl?"    ------------------------------  $ Date: Thu, 8 Jun 2000 01:41:41 -0400' From: "Bill Todd" <billtodd@foo.mv.com> 8 Subject: Re: C bashing (was Re: VMS File Caching Futures( Message-ID: <8hnbgj$rb8$1@pyrite.mv.net>  H Wayne Sewell <wayne@tachysoft.xxx.044962.killspam.0138> wrote in message( news:mwWpwe7r0mNy@tachxxsoftxxconsult...L > In article <8hm90u$k1s$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:oH > > 'Dangerous' is clearly in the eye of the beholder.  I don't consider yourJ > > examples 'dangerous' for my own use of C, because they're not the kind of > > errors I make. >iJ > Never?  Never ever?  Can you honestly sit there and say that not once in yourJ > professsional life you have ever made one of the errors shown?  No typos ("="L > when you meant "==") ever???  If that is truly the case, I am impressed asD > hell.  comp.os.vms has been visited by deity and we must prostrate outselves in > awestruck wonder.  Amen.  ' Really, that's OK:  you can get up now.e   - bill   >p >d >g > [stuff deleted]  >s > -- >aL ============================================================================ ===e: > Wayne Sewell, Tachyon Software Consulting  (281)812-0738 wayne@tachysoft.xxxt: > http://www.tachysoft.xxx/www/tachyon.html and wayne.htmlH > change .xxx to .com in addresses above, assuming you are not a spambot :-)l > L ============================================================================ ===bE > Jake Blues: "Sell me your children!  How much for the little girl?"    ------------------------------  % Date: Wed, 07 Jun 2000 16:03:16 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>-7 Subject: Re: Charon VAX on PC emulator - Asynch DECNET? + Message-ID: <393EAA84.568BDD7@videotron.ca>u   Alan Greig wrote:fB > Ok but is there actually code in DECNET Phase IV to specificallyB > break OPA0? or is it just not supported? In any case a colleagueA > has just told me that you can config the port to be txa0: (eveno3 > though the demo documentation suggest you can't).m  K To enable a port to be usable for DDMCP (or is it DDCMP ?) you need to load J the VTA driver and set the port to be disconnectable. Can you do that with OPA0: ?t   ------------------------------  # Date: Thu, 08 Jun 2000 00:52:58 GMT 2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com> Subject: Command Line question5 Message-ID: <K7C%4.296$ru5.95084@typhoon.aracnet.com>l  H At work I use csh on Unix for most of what I do.  It has one really nice< feature that I like, other than that I pretty much hate csh.  F Is there any kind of DCL equivalent of 'history' and the ability to doK something like a '!lynx' and have it run the last command that started with  'lynx'?t   			Zanet   ------------------------------  # Date: Thu, 08 Jun 2000 01:26:47 GMTg2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>" Subject: Re: Command Line question5 Message-ID: <rDC%4.297$ru5.95428@typhoon.aracnet.com>   + Chuck McCrobie <mccrobi@apl.jhu.edu> wrote:d > $ recall/all > $ recall lynx  > $ help recalle  5 Most cool, thanks!  That's pretty much what I wanted.s   	Zane    ------------------------------  % Date: Wed, 07 Jun 2000 21:30:38 -0400n- From: JF Mezei <jfmezei.spamnot@videotron.ca>n" Subject: Re: Command Line question, Message-ID: <393EF73C.A8CEBFD4@videotron.ca>   "Zane H. Healy" wrote:H > Is there any kind of DCL equivalent of 'history' and the ability to doM > something like a '!lynx' and have it run the last command that started withm	 > 'lynx'?r   $HELP RECALL  $ "RECALL/ALL" will give you a history1 "recall lynx" will recall the last lynx command. s  I In VMS 7.2 (andc possibly earlier), recall also has other options such asn, saving your history or reloading it etc etc.  K You can also use the up/down arrow keys to sequentially recall commands, orb the <CTRL>-B key.e   ------------------------------  % Date: Wed, 07 Jun 2000 21:10:14 -0400o* From: Chuck McCrobie <mccrobi@apl.jhu.edu>" Subject: Re: Command Line question+ Message-ID: <393EF276.BEA7F786@apl.jhu.edu>s   "Zane H. Healy" wrote: > J > At work I use csh on Unix for most of what I do.  It has one really nice> > feature that I like, other than that I pretty much hate csh. > H > Is there any kind of DCL equivalent of 'history' and the ability to doM > something like a '!lynx' and have it run the last command that started withv	 > 'lynx'?  >  >                         Zane   $ recall/all
 $ recall lynxt
 $ help recalli   ------------------------------  % Date: Wed, 07 Jun 2000 16:23:34 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>t1 Subject: Detecting memory leak in C with debugger , Message-ID: <393EAF45.A792C503@videotron.ca>  N I am running some tests on various modules to ensure quality control. In usingM a couple of modules together, I have noticed a memory leak. But when steppingcM through execution with the debugger, nothing obvious appears. (all the "free" + statements seems to execute as they should.e  E So obviously, I have coded something wrong. (and yes, this is full ofe# pointers, and dynamic structures).    J I am wondering, is there a way with the debugger to examine the content ofL memory that is used by malloc and strdup ?, or even better, to find the listH of cells currently allocated by malloc and strdup ? (I assume there is a> structure somewhere containing a pointer and length of cell ?)   VAX, DEC-C, VMS 7.2.   ------------------------------  " Date: Wed, 7 Jun 2000 21:04:43 GMT2 From: malmberg@eisner.decus.org (John E. Malmberg)5 Subject: Re: Detecting memory leak in C with debuggere' Message-ID: <FvsyJv.DpJ@news.decus.org>f  - In article <393EAF45.A792C503@videotron.ca>,  / JF Mezei <jfmezei.spamnot@videotron.ca> writes:c  F >I am running some tests on various modules to ensure quality control.E >In using a couple of modules together, I have noticed a memory leak.o  G >But when stepping through execution with the debugger, nothing obvious E >appears. (all the "free" statements seems to execute as they should.   J The debugger initializes some stack space to zero.  Look for uninitialized
 variables.  F >So obviously, I have coded something wrong. (and yes, this is full of$ >pointers, and dynamic structures).   H You may or may not have.  Some of the calls to the C runtime library canC expand your memory usage.  With recent versions, I would expect the-> memory usage to stabilize, since the storage should be reused.  M One approach would be to be to use a second process to watch the memory usagel< of the system under debug so that you can see the increases.  J IIRC the DEC C RTL uses the LIB$GET_VM_ZONE routines to get "large" memoryE blocks, and then further divides up the blocks for the malloc() stylet; requests.  This tends to make calling malloc() less costly.,  E It is possible for a program fragment the virtual memory pool so that F even though you have matched every malloc() with free(), the pool willH continue to expand, as none of the free cells will be big enough for the malloc() request.i  K >I am wondering, is there a way with the debugger to examine the content ofhM >memory that is used by malloc and strdup ?, or even better, to find the listvI >of cells currently allocated by malloc and strdup ? (I assume there is ai? >structure somewhere containing a pointer and length of cell ?)=  ; It might be useful to have an "Internals and Datastructurest, for the OpenVMS DEC C Runtime library" book.   >VAX, DEC-C, VMS 7.2   -Johnv wb8tyw@qsl.network   ------------------------------  $ Date: Wed, 7 Jun 2000 23:49:28 +0200. From: "Jesper Naur" <jesper.naur@post.tele.dk>5 Subject: Re: Detecting memory leak in C with debugger , Message-ID: <8hmfsc$rt3$1@news.inet.tele.dk>  8 JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message& news:393EAF45.A792C503@videotron.ca...J > I am running some tests on various modules to ensure quality control. In using F > a couple of modules together, I have noticed a memory leak. But when steppingH > through execution with the debugger, nothing obvious appears. (all the "free"- > statements seems to execute as they should.   I One technique, which I have found very useful on several occasions can beuI described as follows (you may or may not find this practical, but here ito is):  B 1) Arrange to have all 'malloc' and all 'free' done by two centralI functions. These two functions should have an additional string parameterrK for identification of the call location. It should be possible 'on the fly' L to have the functions print out the address, size and identification of eachG 'malloc'ed and 'free'd memory area to a text file (for the address, usemG format %08X at the beginning of the line, for reasons which will become 
 clear below).o  L 2) Change your modules to use the two central functions - either by actuallyK changing each specific malloc/free call - or by using a suitable C-macro tomF do it. The latter approach also allows the identification string to beB generated automatically using the __FILE__ and __LINE__ predefined C-preprocessor macros.  J 3) Run the program with the malloc/free output switched on for a while and" let the memory leak do its damage.  K 4) You now have a (probably quite large) file containing the record of eachaD malloc/free done by the program so far. Sort this using the command:  <     $ SORT/KEY=(POS:1,SIZ:8)/STABLE <raw file> <sorted file>  ? (recall, that the memory address is on the first 8 characters).l? This will bring together in chronological sequence each pair ofcL allocation/freeing of each memory area used by your program. When you have aE regular memory leak, browsing through the <sorted file> will probablyvG quickly show which area is being allocated and not appropriately freed.a       Good luck and best regards     Jesper Naurc   ------------------------------  % Date: Wed, 07 Jun 2000 17:49:49 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>t5 Subject: Re: Detecting memory leak in C with debugger , Message-ID: <393EC376.9891D2AB@videotron.ca>   "John E. Malmberg" wrote:tJ > You may or may not have.  Some of the calls to the C runtime library canE > expand your memory usage.  With recent versions, I would expect thel@ > memory usage to stabilize, since the storage should be reused.  J Well, since my test consists of peobably 20 lines of code that is repeatedM over and over again (any disk IO removed, and same values constantly given tosI the storage), any increase on memory requirements once it is in that loopaI indicates a memory leak to me. I want that program to be able to live fortF months without reboot, I have to make sure it doesn't leak ANY memory.  O > One approach would be to be to use a second process to watch the memory usaget> > of the system under debug so that you can see the increases.  K Not necessarily. by the time the working set increases, it is because of an)' operation some time before that failed.l    G > It is possible for a program fragment the virtual memory pool so that H > even though you have matched every malloc() with free(), the pool willJ > continue to expand, as none of the free cells will be big enough for the > malloc() request.y  K Even if the structures being assembled and disassembled in an infinite loopb always contain the same data ?    = > It might be useful to have an "Internals and Datastructures . > for the OpenVMS DEC C Runtime library" book.    H Does such a book exist, or are you suggesting that such a book should be	 written ?e   ------------------------------  % Date: Wed, 07 Jun 2000 18:10:39 -0500 ) From: "John E. Malmberg" <wb8tyw@qsl.net> 5 Subject: Re: Detecting memory leak in C with debuggerg7 Message-ID: <21cc01bfd0d5$9a2b3e80$020a0a0a@xile.realm>o  . JF Mezei <jfmezei.spamnot@videotron.ca> wrote:   > "John E. Malmberg" wrote:oL > > You may or may not have.  Some of the calls to the C runtime library canG > > expand your memory usage.  With recent versions, I would expect theaB > > memory usage to stabilize, since the storage should be reused. >mL > Well, since my test consists of peobably 20 lines of code that is repeatedL > over and over again (any disk IO removed, and same values constantly given toK > the storage), any increase on memory requirements once it is in that loopsK > indicates a memory leak to me. I want that program to be able to live fordH > months without reboot, I have to make sure it doesn't leak ANY memory.  I The guard bands that Jesper Naur posted could be used to get information.s  H If you can comment out functionality until you no longer see the growth, that could be usefull also.   G In the event that you have actually found an undesired feature, a smallF1 reproducer would probably be welcomed at the CSC.a  K > > One approach would be to be to use a second process to watch the memoryG usagee@ > > of the system under debug so that you can see the increases. >lJ > Not necessarily. by the time the working set increases, it is because of an) > operation some time before that failed.o  J And this failure was not detected or trapped?  Since you write the failureI occurs before the working set increases, it would seem that you have moreu than one possible problem.  L I would also assume that you meant virtual address space, as working set canJ expand and shrink as part of normal system memory management.  Working setK expansion does not indicate a memory leak if the virtual address space does  not also expand.   ><I > > It is possible for a program fragment the virtual memory pool so thatcJ > > even though you have matched every malloc() with free(), the pool willL > > continue to expand, as none of the free cells will be big enough for the > > malloc() request.a >rH > Even if the structures being assembled and disassembled in an infinite loop  > always contain the same data ?  L Are all structures freed before any new ones are assembled?  Are they always the same sizes?x  K And are there any calls to the DEC C RTL between them all being free()d andnL then malloc()d again?  Some of them will generate malloc() calls internally.J I do not know if they then keep the storage in case they are called again,# or if they are supposed to free it.m  L If you were to print out a listing of the values of the pointers returned byH malloc(), you would expect to see them being recycled.  If they are not,F then you may be able to determine from the addresses where the leak or) fragmentation of the memory pool occured.A  ? > > It might be useful to have an "Internals and Datastructures 0 > > for the OpenVMS DEC C Runtime library" book. >eJ > Does such a book exist, or are you suggesting that such a book should be > written ?o  L Unfortunately such a book does not exist.  Since C is becomming such a majorJ presence in programming on OpenVMS, such a book is probably needed.  It isG just one of those things that would be nice to have.  Or have the DEC C L Runtime library manual for OpenVMS completely document all possible outcomes of calling a routine.   K Sometimes I do not know when an observed but undesired behavior is a bug oreJ a feature.  And when I look it up on http://www.opengroup.org/ Single UnixC Specification, often it says that it is an implementation dependantr	 behavior.-   -Johnw wb8tyw@qsl.network   ------------------------------  % Date: Wed, 07 Jun 2000 21:48:46 -0400f- From: JF Mezei <jfmezei.spamnot@videotron.ca> 5 Subject: Re: Detecting memory leak in C with debuggerm, Message-ID: <393EFB7C.C4D9301C@videotron.ca>   Jesper Naur wrote:C  1) Arrange to have all 'malloc' and all 'free' done by two centralhK > functions. These two functions should have an additional string parameter * > for identification of the call location.  M I may have to do this. I was just hoping there was a debugger tool that would.N have helped doing this. But I guess I will have to bring out the big artillery and actually modify my code :-)t  I However, instead of trapping malloc and free (and strdup because I use it=K too). I may simply write a routine that dumps the contents of the structureaK including address and contents of each block of data. That would give me an ? idea of which block isn't being re-used the second time around.s   ------------------------------  " Date: Thu, 8 Jun 2000 04:47:12 GMT9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) 5 Subject: Re: Detecting memory leak in C with debugger-+ Message-ID: <fv1GeSo6ip59@eisner.decus.org>-  \ In article <393EFB7C.C4D9301C@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > Jesper Naur wrote:E >  1) Arrange to have all 'malloc' and all 'free' done by two central2L >> functions. These two functions should have an additional string parameter+ >> for identification of the call location.l > O > I may have to do this. I was just hoping there was a debugger tool that wouldbP > have helped doing this. But I guess I will have to bring out the big artillery! > and actually modify my code :-)   B Are you saying that the VMS C implementation is somehow impervious to the VMS Heap Analyzer ?   ------------------------------  # Date: Wed, 07 Jun 2000 17:56:40 GMTt3 From: Eric Dittman <dittman@narnia.int.dittman.net>s: Subject: Digital TCP/IP Services and Alternet Telnet Ports? Message-ID: <s1w%4.99177$v7.3642715@news-west.usenetserver.com>s  < I'm using Digital TCP/IP Services for OpenVMS V5.0A.  I need= to have a telnet server listening in on a port other than 23.=; According to Compaq Support, this isn't supported except by=@ writing a program to listen on the port, validate the connection= and then pass the connection over to LOGINOUT.  I could writew= this, but if anyone else has already written the code and cans< share it I'd appreciate saving the time so I can concentrate on some other issues.=   Thanks.= -- = Eric Dittman dittman@dittman.nete   ------------------------------  # Date: Thu, 08 Jun 2000 00:58:28 GMT- From: mccrobie@my-deja.com/ Subject: Files-11 ODS-2 Readability for FreeBSDM) Message-ID: <8hmr3d$thr$1@nnrp1.deja.com>c  G I've been working on a READ-ONLY Files-11 ODS2 file system for FreeBSD.6D This file system is an installable kernel module and provides access6 to files as a "true" mountable file system on FreeBSD.  7 Currently supported are the following FreeBSD versions:        FreeBSD 3.3o     FreeBSD 4.0p  D I am currently looking for beta testers who wish to read their ODS-2G file systems on FreeBSD - my motivation has been the optical world, buteF as long as FreeBSD recognizes the device, the file system should work.  F Sorry, automatic record type and file organization conversions are notH available, yet.  I'm guaging the interest in this product before I spend anymore time with it.n  D I should note that I plan this to be a commerical product which willE cost actual money.  However, I'm open to discussions on license fees, B terms, etc.  Beta testers will probably receive a free copy of the0 final production file system - some incentive...  C For you Linux and Solaris users, I may be persuaded to add support.mD Sorry for the Compaq Tru64 users - Compaq wants too much money for a development machine...   If you are interested:   Chuck McCrobie (** MAD VAX **) mccrobie@my-deja.com    & Sent via Deja.com http://www.deja.com/ Before you buy.4   ------------------------------   Date: 7 Jun 2000 20:26:46 +0200e* From: eplan@kapsch.net (Peter LANGSTOEGER)- Subject: Re: Hobbyist vms 7.2 alpha questions * Message-ID: <393e93e6$1@news.kapsch.co.at>  q In article <8hm1us$pk1$1@mailint03.im.hou.compaq.com>, hammond@not@peek.ppb.cpqcorp.net (Charlie Hammond) writes:c* >In article <393e727f@news.kapsch.co.at>, - >eplan@kapsch.net (Peter LANGSTOEGER) writes:y >...( >>Better go with V7.2-1 instead of V7.2. >'C >Yes, V7.2-1 is preferable to V7.2, but I ran my worstation at V7.2tC >for some time with no problems.  (PRODUCT SHOW HISTORY says it ran " >from 17-FEB-1999 to 05-Jan-2000.)  K Ok. Your decision. But this decision was done in FEB-1999 and at this time,iE there have been no V7.2-1. But today, my decision (and I think almostrI everyone's decision - if one has the choice) which version to upgrade to,q now, is V7.2-1.c  M >>So, you haven't got VMS V7, you run OpenVMS Alpha V6.2 (as you wrote above)0J >>First, upgrade to VMS V7 or do a fresh/new install of OpenVMS Alpha, ... >:D >Is there a reason for this advice?  The V7.2 manual says you can go< >directly from V6.2, and I know of no reason why you cannot.  L Did you see the list of the installed products on this system ? Some of themI are unknown to me. I can't tell, that they will run on V7.2 or will CRASHfI the system (just like POSIX did, if you forgot to deinstall it before thenJ OpenVMS upgrade and reinstall the new version after the upgrade - I think,I it was POSIX V2.0 and V3.0, then). You can decide to beg for problems ;-)-  E With a fresh install, you can also decide, which layered products you<F want to install at all, and not lurking around on how to deinstall now5 unwanted products installed with VMSINSTAL years ago.>  H But of course, if you do a fresh install, you should save the files fromG the old system you don't want to loose. But this is obvious, I think...h   -- p< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888h< FBFV/Information Services           E-mail  eplan@kapsch.netF <<< KAPSCH AG  Wagenseilgasse 1     PSImail PSI%(0232)281001141::EPLANH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"N "VMS is today what Microsoft wants Windows NT V8.0 to be!" Compaq, 22-Sep-1998   ------------------------------   Date: 7 Jun 2000 18:45:32 GMTt8 From: hammond@not@peek.ppb.cpqcorp.net (Charlie Hammond)- Subject: Re: Hobbyist vms 7.2 alpha questions-6 Message-ID: <8hm58c$r03$1@mailint03.im.hou.compaq.com>  + In article <393e93e6$1@news.kapsch.co.at>, j, eplan@kapsch.net (Peter LANGSTOEGER) writes:  L >Ok. Your decision. But this decision was done in FEB-1999 and at this time,F >there have been no V7.2-1. But today, my decision (and I think almostJ >everyone's decision - if one has the choice) which version to upgrade to, >now, is V7.2-1.  F I agree that if V7.2-1 is available, it is preferred.  BUT, if all youH have is V7.2, there is, in general, no reason to avoid it.  Particularly not as a "hobby".t  M >Did you see the list of the installed products on this system ? Some of them J >are unknown to me. I can't tell, that they will run on V7.2 or will CRASHJ >the system (just like POSIX did, if you forgot to deinstall it before theK >OpenVMS upgrade and reinstall the new version after the upgrade - I think,dJ >it was POSIX V2.0 and V3.0, then). You can decide to beg for problems ;-)  L (1) It is not clear that this problem is made much better by an intermediateM upgrade to V7.0.  Yes, he may have to remove some products and upgade others, L but I don't see this as an impediment to a V6.2 -> v7.2 (or V7.2-1) upgrade.D Whateer has to be removed or upgraded has to be done in either case.5 Remove the before the upgrade; re-install them after.e  E (2) The question asked was, I believe, how to upgrade WITHOUT loosingmF other things on his system disk.  So an new install is pretty much out of the question.   -- (K     Charlie Hammond -- Compaq Computer Corporation -- Pompano Beach  FL USAnF          (hammond@peek.ppb.cpqcorp.net -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Wed, 07 Jun 2000 17:55:26 -0400 * From: David A Froble <davef@tsoft-inc.com>- Subject: Re: Hobbyist vms 7.2 alpha questions - Message-ID: <393EC4CE.B4588296@tsoft-inc.com>e  
 Dan wrote: > G > Terminating is strongly recommended.  Do you want to terminate? [YES]E > Portion Done: 0%1 > %PCSI-E-S_OPCAN, operation cancelled by request-A > %PCSIUI-E-ABORT, fatal error encountered - operation terminatedj > $.  M Ok, trying to be helpful, not a smart-ass.  What happens if you reply 'NO' togM the above question.  I remember (it's been a while) that I ran into somethingaO similar (don't know if it was DECnet) and was a bit determined to continue, ande everything appeared to be fine.q  C It could have been a V7.2 field test distribution, cannot remember.c   Dave   -- k4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  % Date: Wed, 07 Jun 2000 20:44:00 +0200o2 From: martin@RADIOGAGA.HARZ.DE (Martin Vorlaender)D Subject: Re: is anyone able to get a DE500-BA working under 6.2 1h3?; Message-ID: <393e97f0.524144494f47414741@radiogaga.harz.de>l  * The Youngun (ftg_dts@earthlink.net) wrote:G : Is anyone able to get a DE500-BA working under Open VMS 6.2 1H3 on an  : AlphaServer 800? :r6 : Per COMPAQ, the following patches have been applied: :m
 : alpclusio01b : alpcliu03h
 : alplan05 : alpboot12   < I seem to remember that a sys* patch was also required - but+ perhaps it's contained in one of the above.    cu,g   Martin --D                        |  Martin Vorlaender  |  VMS & WNT programmer1   OpenVMS: When you    |  work: mv@pdv-systeme.detH   KNOW where you want  |        http://www.pdv-systeme.de/users/martinv/8   to go today.         |  home: martin@radiogaga.harz.de   ------------------------------   Date: 7 Jun 2000 19:46:06 GMTi2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)D Subject: Re: is anyone able to get a DE500-BA working under 6.2 1h3?6 Message-ID: <8hm8pu$sgi$1@mailint03.im.hou.compaq.com>  p In article <393e97f0.524144494f47414741@radiogaga.harz.de>, martin@RADIOGAGA.HARZ.DE (Martin Vorlaender) writes:+ :The Youngun (ftg_dts@earthlink.net) wrote:uH :: Is anyone able to get a DE500-BA working under Open VMS 6.2 1H3 on an :: AlphaServer 800?a ::7 :: Per COMPAQ, the following patches have been applied:l ..= :I seem to remember that a sys* patch was also required - buto, :perhaps it's contained in one of the above.  2   IIRC, AlphaServer 800 needs/uses ALPCPU1B02_062.M   DE500-BA requires an ECO specific to the CPU for OpenVMS V6.2 and V6.2-1Hx.i  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------   Date: 8 Jun 2000 01:22:42 GMTa2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)3 Subject: Microsoft loses, www.compaq.com disappearss, Message-ID: <8hmsh2$qj5@gap.cco.caltech.edu>  K Microsoft lost its court case, and now www.compaq.com does not even have a aH DNS entry (18:15 PST).  Guess Microsoft was right, rule against them and- the world as we know it comes tumbling down.    G But seriously folks, is it just the routing to us or has www.compaq.como" really gone away for some reason?    Thanks,k   David Mathog mathog@seqaxp.bio.caltech.edur? Manager, sequence analysis facility, biology division, Caltech s   ------------------------------  % Date: Wed, 07 Jun 2000 18:51:29 -0700 ! From: Shane.F.Smith@Healthnet.comg3 Subject: Microsoft loses, www.compaq.com disappearstC Message-ID: <OFF21AE167.E377F4BD-ON882568F8.000A3518@HEALTHNET.COM>   D >Microsoft lost its court case, and now www.compaq.com does not even@ > have a DNS entry (18:15 PST).  Guess Microsoft was right, rule? > against them and the world as we know it comes tumbling down.y > 9 >But seriously folks, is it just the routing to us or hasa2 > www.compaq.com really gone away for some reason? >  >Thanks, >s
 >David Mathog   I FWIW Dave, I'm seeing the same thing. However, www.openvms.digital.com is.K still there so all is right with the world! Hey, maybe Compaq realised thateJ the Digital and OpenVMS names are more associated with quality than Compaq: and Microsoft, and there's a corporate reshuffle going on?   ------------------------------  % Date: Wed, 07 Jun 2000 22:45:44 -0400d- From: JF Mezei <jfmezei.spamnot@videotron.ca><7 Subject: Re: Microsoft loses, www.compaq.com disappearsr, Message-ID: <393F08D2.95B46201@videotron.ca>   David Mathog wrote:r > L > Microsoft lost its court case, and now www.compaq.com does not even have aJ > DNS entry (18:15 PST).  Guess Microsoft was right, rule against them and. > the world as we know it comes tumbling down.  N And I still wonder why Compaq chose to side with Microsoft when all the others( either stayed quiet or sided against MS.  J As of 22:45, www.compaq.com and www.compaq.ca arte reacheable from canada.   ------------------------------  % Date: Wed, 07 Jun 2000 23:16:30 -0400 * From: Chuck Chopp <ChuckChopp@rtfmcsi.com>7 Subject: Re: Microsoft loses, www.compaq.com disappears + Message-ID: <393F100E.EB0647D3@rtfmcsi.com>s   David Mathog wrote:o  L > Microsoft lost its court case, and now www.compaq.com does not even have aJ > DNS entry (18:15 PST).  Guess Microsoft was right, rule against them and. > the world as we know it comes tumbling down. >CI > But seriously folks, is it just the routing to us or has www.compaq.comm# > really gone away for some reason?e >e	 > Thanks,  >  > David Mathog > mathog@seqaxp.bio.caltech.edu @ > Manager, sequence analysis facility, biology division, Caltech  I Interesting.  A quick look at Network Solutions' whois database shows thea following for "compaq.com":i   Registrant:f:                   Compaq Computer Corporation (COMPAQ-DOM)5                      20555 State Highway 249, M020303 &                      Houston, TX 77070  ,                      Domain Name: COMPAQ.COM  ,                      Administrative Contact:J                         Compaq Internet Domain Name Registrar  (CI746-ORG) domainreg@COMPAQ.COM-                         Compaq Computer Corp. /                         20555 State Highway 249 )                         Houston, TX 77070e                         US$                         281-518-8427)                         Fax- 281-514-4863t5                      Technical Contact, Zone Contact:s9                         Domain Tech-Contact  (DT1681-ORG)a domaintech@COMPAQ.COMr-                         Compaq Computer Corp.e8                         20555 State Highway 249, M020303)                         Houston, TX 77070e                         US$                         281-518-8427%                      Billing Contact:tH                         Compaq Internet Domain Name Billing  (CI745-ORG) domainbilling@COMPAQ.COM-                         Compaq Computer Corp.o/                         20555 State Highway 249x)                         Houston, TX 77070r                         US$                         281-518-4784)                         Fax- 281-514-4863o  8                      Record last updated on 15-May-2000.3                      Record expires on 21-Apr-2001.m3                      Record created on 20-Apr-1995.eF                      Database last updated on 7-Jun-2000 15:54:38 EDT.  4                      Domain servers in listed order:  ?                      NS1-PUBLIC.ZTX.COMPAQ.COM    161.114.1.204l?                      NS1-PUBLIC.ZMA.COMPAQ.COM    161.114.64.24h    N I ran NSLOOKUP on my WinNT laptop and directly interrogated each of those nameJ servers (via their IP addresses) and I got responses from both of them forJ "www.compaq.com".  The address that they both returned was "207.18.199.3".  H If I try to resolve the name via my ISP's DNS servers I cannot get an IPM address for www.compaq.com.  My ISP is Bell South and I have a DSL connectionoH via Greenville, SC.  It appears that the Bell South DNS servers that are1 assigned to me are ones based out of Atlanta, GA.   I If  I connect to some other DNS servers that I know are based on Internet7J connections via Verio and via MCI there appears to be no problem resolving www.compaq.com.e  @ It looks like a partial DNS snafu is in progress at this time...     Chuckn -- Chuck Choppe  8 ChuckChopp@rtfmcsi.com            http://www.rtfmcsi.com0                                   ICQ # 22321532@ RTFM Consulting Services Inc.     864 801 2795 voice & voicemail2 103 Autumn Hill Road              864 801 2774 fax4 Greer, SC  29651                  800 400 4935 pagerC                                   8004004935@alphapage.airtouch.comu   ------------------------------  # Date: Wed, 07 Jun 2000 20:04:55 GMTt From: Dirk Munk <munk@home.nl>! Subject: Re: MMOV 2.2 and VMS 7.1l' Message-ID: <393EAAEB.5CDB80F8@home.nl>r   Hoff Hoffman wrote:- > z > In article <7a1+fuLOFdUc@gaelic>, pmoreau@cenaath.cena.dgac.fr (Patrick MOREAU, CENA Athis, Tel: 01.69.57.64.40) writes:P > :Multimedia Services for OpenVMS is available on the June CD's with support of4 > :Ensoniq card and DSx0 workstations and servers !!  B And will it support the ESS (?) soundcards in the PWS series too ?     > :rP > :The minimum OVMS version is 7.1-2, but is there a chance to see it working onM > :7.1H1 ? (My Home Alpha 255 runs 7.1-1H1 with a good stability and if I cano > :avoid an update ...)  > E >   I would strongly encourage an upgrade to (at least) OpenVMS AlphaaG >   V7.1-2 for any system currently running V7.1, V7.1-1H1 or V7.1-1H2.tF >   PCSI is used for all ECO and update operations starting in V7.1-2,D >   and we also collected up a large variety of ECOs and rolled themF >   all into V7.1-2, and we have subsequently made various UPDATE kitsE >   available for V7.1-2.  All of which make the use of OpenVMS Alphac$ >   V7.1-2 preferable over V7.1-1H1. > E >   I would strongly discourage trying to use an OpenVMS Alpha V7.1-2c, >   kit on an earlier OpenVMS Alpha release. > E >   All EV6- and EV67-class boxes are also first supported on V7.1-2.r > F >   The AlphaStation 255 also runs rock-solid on OpenVMS Alpha V7.2-1,: >   with the current GRAPHICS and UPDATE ECO kits applied. > 2 >   Bottom line: are you feeling lucky today?  :-) > P >  --------------------------- pure personal opinion ---------------------------N >    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Wed, 07 Jun 2000 19:05:06 -0700-! From: Shane.F.Smith@Healthnet.com:2 Subject: Multithreading and event flags limitationC Message-ID: <OF439CC442.C0A63577-ON882568F8.000B745B@HEALTHNET.COM>.  F Background: I'm writing a multithreaded program using QIOW calls to doJ TCP/IP sends and receives. I'm using an asynch timer with a completion ASTG to cut these calls short if they take too long. I'm using the EFN$C_ENFlK value for the QIOW's event flag, and that's working fine. However, the docseH say you can only use that with "W" calls (like QIOW) that have their own thread-local IOSBs.o  B SYS$SETIMR isn't a "W" call, and doesn't take an IOSB, which to meH indicates I can't use EFN$C_ENF, and have to reserve a unique event flagB even though my code has no direct use for it. That would mean thatK regardless of my TQLM quota, I can't safely have more timers running than IiI have event flags available. Is this the case, or is there some way aroundh it?i   Shanea   ------------------------------  % Date: Wed, 07 Jun 2000 17:58:23 -0400c* From: David A Froble <davef@tsoft-inc.com>% Subject: Re: OpenVMS AND THE INTERNETt- Message-ID: <393EC57F.F2053AE1@tsoft-inc.com>a   Richard Jordan wrote:a > H > Amazon still lists it as purchasable on a pre-order basis, despite theH > fact that they informed me it was canceled some months ago.  Maybe has< > something to do with those new HP systems they're using :)  M Darn, I know I should resist this, but it's got the better of me.  Maybe it's  the Unix they're running?0   Dave   -- a4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  $ Date: Wed, 7 Jun 2000 15:23:47 -0400' From: "Bill Todd" <billtodd@foo.mv.com>>J Subject: Re: OpenVMS commentaries (was Re: Gartner commentary on Wildfire)( Message-ID: <8hm7a6$eg1$1@pyrite.mv.net>  5 Rob Young <young_r@eisner.decus.org> wrote in messagei% news:yYoH3e2ptHXB@eisner.decus.org...0L > In article <8hk5sg$29i$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:r   ...b  L > > Though not quite equal performance, since writes to the write-back cacheH > > (and reads from the controller cache, for that matter) take around aL > > millisecond (your own observation, IIRC) vs. well under 100 microseconds toH > > the Unix system cache:  you'll cut your default-settings performance penalty K > > from 10-to-1 or worse down to maybe around 2:1 (at least for writes; ifr you.K > > can use VIOC for reads, they may be closer to equal), which changes thet5 > > perception from 'damn slow' to merely 'sluggish'.w > >m >I= > Good observation... but the Unix trick to its own cache fortB > a write-back cache isn't very reassuring if the box takes a hit.  I It's fully as reassuring as the state of the records you've automaticallytJ accumulated in your RMS sequential file buffer but not yet written to diskI or hardware cache (or explicitly accumulated by using deferred write, buttL we're talking default operation here).  In both cases, the data weren't madeJ persistent because they presumably didn't have to be.  So what exactly areK you not reassured about:  that applications that had no idea they needed tooK get their data on disk *might* (or at least in the case of sequential files 6 might not) have gotten it there by accident using VMS?  A > The example I wrote about and am using and have used on several @ > different occasions, the write occurs to cache in two separate< > controllers, hardly an issue of stability or lack thereof.  J When discussing data storage, 'stable' is a technical term synonymous with! 'persistent' (or 'non-volatile').a   >-J > > Put as good a face as you can on it, it doesn't make VMS look great in this	 > > area.o > >k >bA > Depends on how you measure it.  End user response is fantastic.  > What are you measuring?m  K Default performance compared to Unix:  I really don't know how to make this.6 any more explicit than I have (several times) already.   >Y > >>E > >> Linux is great and Linux is fast, yada yada.   But Linux doesn'teC > >> have 4 or 5 nodes in a shared disk cluster.  Shared disk?  Whoe > >> needs shared disk.... > > 5 > > Good question:  most installations don't, at all.  > E > Last 3 VMS sites I have been involved in have.  How do you quantifya> > or measure most VMS installations?  Last I checked, this was, > comp.os.vms and I didnt' see a cross-post.  J I didn't say 'most *VMS* installations' (since your Linux comment appearedJ to be talking about general needs), but I'd guess that the statement 'mostI VMS installations don't "need" shared disk' is also true.  It may even beaJ true that a simple majority of VMS installations don't even *significantlyI benefit* from shared-disk access (as contrasted with, e.g., a served-file  approach, such as Tru64's).s  L As confirmation, consider the relative popularity of non-shared-disk clusterI (or other closely-knit networked) solutions, now that alternatives to VMS I clusters are starting to exist.  Or look at the database world, where theeH only distributed solutions that use shared-disk configurations are those: using Oracle Parallel Server (or DB2 on Parallel Sysplex).  H "Need" is a strong term, indicating the lack of a viable solution in theG absence of the feature.  Very few clustered solutions *need* disk-level-K sharing to be viable.  Some larger percentage, but still likely a minority,kH of clustered solutions may *benefit* noticeably from disk-level sharing,' either in performance of manageability.p   >' > > While shared-disk-G > > configurations are the *only* way to address some (extreme) scalingbI > > situations well, those situations aren't that common.  Lacking such aeJ > > scaling problem, if you can successfully partition your data such that it'sK > > usually local to the node that needs it (the common case, especially in H > > read-mostly web-server environments), your performance will often be betterL > > than in a shared-disk cluster where you *haven't* paid attention to suchK > > partitioning, because in the latter case the nodes will be caching data H > > redundantly (using up memory they could be using more effectively in othernL > > ways) and fighting each other over the occasional need to update it (and XFC C > > won't eliminate these problems, though it may ameliorate them).  > >e > D > Good solution.  You introduced partitioning of data as a solution.F > Unfortunately, the databases I am working with aren't partitionable.  F All databases are partitionable.  The only question is whether they'reK *efficiently* partitionable (and that can be turned into a hardware problem2G if you choose to do so, since faster hardware can always compensate forc inefficiency).  F Not to say that shared-disk isn't an important benefit for the systemsE you've been associated with:  that simply places you in the minority.t  C > What do you propose then?   Also, with partitioning of data , you = > better make real sure you have someone that knows what they * > are doing or that is a whole other pain.  H Yes, partitioning can be a real pain.  Or not (often it's easy, in whichJ case it's something you should do even in a shared-disk environment).  ButG it's never insurmountable, just one more consideration to factor into aM proposed solution.     Friend tells me a|G > of an installation where one DBA does that as a living.  PartitioningvC > Unix databases to make them usable.  Must be nice work if you cand	 > get it..  K I'd expect it to become tedious rather quickly.  But it probably pays well.o   >@ > >> > >> And on and on we go...) > >j > > Until you understand it. > >5 >B > Please...e >.D > My experience of VMS clusters and how they are being used tells me> > they work.  Very busy VMS clusters at that.  So what don't I7 > understand?  That Linux is great?  I understand that.n  < Of course they work, usually very well.  So do alternatives.  L What you don't seem to be able to understand is that, while VMS enjoys quiteH a few technical advantages over those alternatives in some areas, in theE particular case we're talking about (default, and even to some degreeeD non-default, file system performance) Unix-style mechanisms enjoy anI absolute, real, usable performance advantage over VMS facilities, withoutxG compensating drawbacks.  No ifs, ands, or buts:  VMS could, and should,-G improve in this area by emulating them (or if even improving upon them,l which is certainly possible).-  G [And I probably shouldn't have let you divert the conversation down the.I shared-disk rat-hole above, since it has virtually nothing to do with theeI area previously being discussed.  But, despite the fact that I'm a fan ofuJ shared-disk configurations, because in *rare* cases they *do* satisfy realL needs that partitioned configurations can't, it bothers me to see VMS bigotsE knock other environments in a *general* way simply because they don'toE support something so rarely needed:  that's not where VMS proponents'sK attention should be focused, because that's clearly not where the market assG a whole is focused - and if VMS is to thrive, the latter is important.]    - bill   >  > Rob0 >: > Yeee HAAAAAA!!!!!h >.   ------------------------------   Date: 7 Jun 2000 18:12:07 GMTs2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)+ Subject: OpenVMS V7.3 SDK Info; Freeware V5 6 Message-ID: <8hm39n$q9e$1@mailint03.im.hou.compaq.com>  K   Follow-ups to this message have been set to vmsnet.sdk.openvms.fieldtest.a  I   Attached is the announcement of the start of the general external test sI   for OpenVMS V7.3.  Additional updates and features and information are oJ   expected to be made available as the OpenVMS V7.3 field test progresses.  G   I am also now soliciting updates and new submissions for the OpenVMS  C   Freeware V5 release.  Submissions for the Freeware can be sent tonF   openvms.freeware@compaq.com -- please note this email address is forF   submissions only and cannot accept requests for the field test kits.  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   	--3   Dear Customer,  I Thank you for your interest in the OpenVMS Version 7.3 Software DeveloperdK Kit (SDK) Program.   We feel that this program has been very successful in r: the past, and we hope that you will decide to participate.  E The OpenVMS Version 7.3-EFT1 SDK will be available in July 2000 as an G orderable kit through our DECdirect channel.  You can order the kit by MH ordering part number QA-MT3AD-H8 at a cost of $39.95.   Once the kit is E available, it will be as simple as calling 1-800-ATCOMPAQ.  Business 6D Partners can order it on the Web through the Compaq Business Link at www.businesslink.compaq.com.  I The V7.3-EFT1 SDK will contain early releases of the OpenVMS V7.3 VAX andoJ Alpha operating systems, as well as DECwindows and network products.  The 2 SDK will also contain many new features including:  B o Write Bitmap and Shadow Minicopy -- Performance and availability&   enhancements to Shadowing full copy.  ) o Switched LAN as a cluster interconnect:.  A   - Support for Gigabit Ethernet and ATM as full switched clusteri     interconnects.  B   - Allows all LAN devices to participate in load distribution of      cluster traffic.  F o Lock Manager 2000 (Alpha only) -- Enhancements that improve cluster G   and application performance, SMP system performance, and scalability C   and application availability.   K o Extended File Cache (XFC) Version 1.0 support (Alpha only) -- Replacementu<   for VIOC.  Extends the capabilities of cluster file cache.  I o RMS Locking -- Reduces global buffer bucket locks by removing locks forpF   read operations.  Also allows RMS to read a record without taking a    record lock.  C o Kerberos Version 5 Security Client (Alpha only) -- Allows OpenVMSaD   applications to communicate and authenticate with NT5 domains and    most UNIX platforms.  C o Clusterwide Intrusion Detection -- Expands the current individual D   node-base intrusion detection in OpenVMS today to be cluster-wide    and cluster aware.  E o QIO Server (Alpha only) -- Provides improved performance over MSCP,qF   improved scaling and system management control, reduced maintenance D   cost, and a public device driver interface to support third party    driver writers.   F o SCSI Fastpath (Alpha only) -- Improves SCSI performance and scaling    in SMP systems.q  
 o And more.  N    J The OpenVMS Version 7.3-EFT2 SDK is presently scheduled to be available inK September/October  2000.  Customers can choose to either purchase both the  L OpenVMS Version 7.3-EFT1 SDK and the OpenVMS Version 7.3-EFT2 SDK at a cost # of $39.95 each, or just select one.   E Support for the SDK program will continue to be through the followingo Internet newsgroup:   "       vmsnet.sdk.openvms.fieldtest  I Please see the OpenVMS Web Page www.openvms.compaq.com  in late June 2000 ; timeframe for additional information for ordering this kit.   K Your involvement and support have been very important to the success of thePH overall SDK Program.  We hope that you will continue to support Version ' 7.3-EFT1 as this program moves forward.   
 Sincerely,   OpenVMS Product Management   ------------------------------  % Date: Wed, 07 Jun 2000 17:20:39 -0500s) From: "John E. Malmberg" <wb8tyw@qsl.net>oC Subject: Re: Part numbers for Qbus SCSI + ESDI controllers, anyone?t7 Message-ID: <21a801bfd0ce$9e6bf9a0$020a0a0a@xile.realm>g  8 Christopher Smith <chriss@Mufasa.pubserv.company> wrote: > E > I'm checking around to see about getting a SCSI controller, or ESDI $ > controller for my VAXStation 3200. >eK > I've been asked whether I can get specific part numbers of the boards I'mrH > interested in, so, since I have no idea which ones are good, here's my > question:o > C > Does anyone have a suggestion for a nice SCSI controller, or ESDI B > controller to look for -- and a part number to go along with it?  I I would check the CMD web site, and also see if the web sites for EMULEX, L and DILOG have any information.  This should get you some information on the third party controllers.  5 Try a search engine with q-bus or qbus and SCSI also.2  K The KZQSA should work in your unit providing it is running at least VAX/VMS4L 5.3, however I do not think that a VAXstation 3200 will be able to boot from it.u    H Good luck in your search.  I would expect that you could probably find aG faster used VAXstation 3100 or 4000 with memory and SCSI disks for lesse< money than what a dealer will sell you a Q-BUS SCSI adaptor.    L ESDI should be easier to find, but since those disks have not been made in aH while, I would be cautious on moving to it.  It will be faster than your RD5x series disks though.a  I > By "nice," I mean specifically that I'd like it to format disks, and if I > it's a scsi controller, maybe be able to use more than just one type ofaF > device (disk and tape would be great -- cdrom would be even better).  J Many third party SCSI q-bus adaptors will only support either disk or tape
 but not both.u   -Johnp wb8tyw@qsl.network   ------------------------------  # Date: Wed, 07 Jun 2000 22:43:52 GMT12 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>C Subject: Re: Part numbers for Qbus SCSI + ESDI controllers, anyone?r5 Message-ID: <IeA%4.289$ru5.93396@typhoon.aracnet.com>,  4 Christopher Smith <chriss@mufasa.pubserv.com> wrote: > Hi everybody,1  E > I'm checking around to see about getting a SCSI controller, or ESDIv$ > controller for my VAXStation 3200.  J Well, I'm using a KDA50 w/RA72's and RA73's on my MicroVAX 3.  However, onH my PDP-11/73 I use a Webster WQESD/04 ESDI controller, and a Viking SCSI controller.7  H The Webster is a very nice ESDI controller, and can do some fancy tricksJ with the disks that at least on the PDP-11 are very handy (it can make theF disk look like several disks to the OS, with each one being bootable).  I > By "nice," I mean specifically that I'd like it to format disks, and ifhI > it's a scsi controller, maybe be able to use more than just one type of-F > device (disk and tape would be great -- cdrom would be even better).  K With SCSI controllers you've got to watch out, some are Disk only, some are K Tape only, and some are both.  Also the versions of the ROMs will depend onrL what you can do with the board.  I don't believe the Viking for example willL work with CD-ROMS with most versions of the ROMS, but if you have new enough
 ones it will.h  L Something to be aware of is that you can expect a SCSI Q-Bus adapter to costK anywhere from $500-1000 on average.  Unless you get lucky and find one at acJ junk dealer, and they don't know what it is.  Of course the Tape only onesG are significantly cheaper and I've payed $50 each for the two tape onlysM controllers I have.  If you're real "lucky" you can also find a Q-Bus adaptero4 to a very early model non-SCSI CD-ROM that DEC made.  H For hobbyist use I'd recommend getting an ESDI Q-Bus controller and/or a# VAXstation 3100 or VAXstation 4000.r   				Zane   ------------------------------  % Date: Wed, 07 Jun 2000 23:39:02 -0400e2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>0 Subject: SMTP_RECV_MAIL from TCPIP$smtp_receiver7 Message-ID: <200006072339_MC2-A7ED-7DFA@compuserve.com>m  J         Just remember that UCX (TCPIP) had it's origins in Unix, not VMS!=  =  8 The *original* UCX was a port of the Berkeley Unix code.  J         TCPIP 5.0 is *not* an update or rewrite of UCX.  It's a port of t= he Digital Unix code!  5         You will find little Unixisms here and there!t  # Message text written by sol gongolaeD >I sometimes see this message in an smtp log file. It doesn't follow> the standard vms message format and i did not see it mentioned in the tcpip manuals:   + SMTP_RECV_MAIL: remote host not in hosts DBt   <.   ------------------------------  % Date: Wed, 07 Jun 2000 16:35:00 -0400-" From: Dan Sugalski <dan@sidhe.org>5 Subject: Re: Syntax-coloring editors (was: C bashing):: Message-ID: <4.3.2.7.0.20000607163331.01cfe3d0@24.8.96.48>  0 At 06:52 PM 6/7/00 +0000, Larry Kilgallen wrote:F >In article <Zgs%4.3227$2X2.129940@newsread2.prod.itd.earthlink.net>, 7 >"Michael D. Ober" <mdo.@.wakeassoc.com.nospam> writes:. >nI > > Agreed.  I tried to go through some Ada code.  After a few minutes, IqN > > started wishing for a program that would replace all comments with spaces,L > > leaving end of line markers so I could put the source and the documentedK > > source next to each other.  The comments were so heavy that I literallye* > > couldn't find the actual program code. >oD >The professional approach to this in recent years has been to use aB >syntax-coloring editor.  Perhaps the person who wrote the Ada youG >were reading had used one and presumed that readers would use one too.  >mI >One that I know of is called "Alpha" (strangely enough) and is sharewaret7 >or something close.  Ask in comp.lang.ada for details.i  K Emacs also does syntax coloring, and it's quite nice. (And, apropos to the rL ranting, it's only barely written in C--a good chunk is either in LISP or C F code so horribly mutated by the preprocessor that it might as well be 2 LISP... :) LSE might as well, though I'm not sure.   					Dan  I --------------------------------------"it's like this"-------------------d2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and even ;                                       teddy bears get drunkq   ------------------------------  # Date: Thu, 08 Jun 2000 03:07:03 GMT + From: "Jim Kozlowski" <jkozlow1@elp.rr.com>h Subject: system slow3 Message-ID: <r5E%4.137$244.2520@typhoon.elp.rr.com>w  ( We have an Alpha 200 running openVMS 6.1G System running fine until 2 months ago...at that time we replaced a 1gb I drive with a 2gb drive (i do not know drive model off hand) new drive was  not system disk.E We added the disk space to accomodate a software upgrade which has usa6 running two different version of synergy concurrently.H It seems, that since that time, the system runs much slower. At times itH will suspend all activity for 20-30 seconds at a time. Processes will be
 suspended.H I need some troubleshooting help. What can I monitor to determine if the9 hesitation is disk or software or something else related.d	 Thank your Jim K    ------------------------------  " Date: Thu, 8 Jun 2000 06:35:52 GMT* From: young_r@eisner.decus.org (Rob Young) Subject: Re: system slow+ Message-ID: <6L$nJg9RlTr8@eisner.decus.org>a  a In article <r5E%4.137$244.2520@typhoon.elp.rr.com>, "Jim Kozlowski" <jkozlow1@elp.rr.com> writes:e* > We have an Alpha 200 running openVMS 6.1I > System running fine until 2 months ago...at that time we replaced a 1gbaK > drive with a 2gb drive (i do not know drive model off hand) new drive wase > not system disk.G > We added the disk space to accomodate a software upgrade which has uso8 > running two different version of synergy concurrently.J > It seems, that since that time, the system runs much slower. At times itJ > will suspend all activity for 20-30 seconds at a time. Processes will be > suspended.J > I need some troubleshooting help. What can I monitor to determine if the; > hesitation is disk or software or something else related.  > Thank yous > Jim Ks >     E 	The problem for the hang may be that you are short on memory.. to a eG 	point... this is most likely happening during busy or peak periods of  	 	the day.   9 	Unfortunately, Deja is brain-dead these days... they arel= 	re-archiving everything or somesuch.  Otherwise, I would cutm< 	and paste this analysis in here.  As it is , I have to wing: 	it but hope to follow-up tomorrow as I believe I saved it 	to a file also (at work).  > 	I am going to be rather disjointed but not on purpose.  Sorry> 	for the secondhand effort but the work is done before and why> 	try to be accurate... you can do the verification tomorrow...     		1)  Memory is getting tight./ 		2)  Gets down to 2 times freegoal on freelist 6 		3)  Ticker kicks in trimming back 25% of mostly idle 			processes.  Problem:   - 				25% trim of large working sets in largishm, 				memory config (1 gig or more) results in/ 				60 MByte being freed up (more or less can'tn 				say without poking around).s  @ 		4)  Contents of Freed memory must be written to backing store, 			i.e. pagefile  Problem:  - 				Takes time to write out 60 MByte, this is + 				what the "hang" is all about (for thosee" 				that are effected, see below).   		Verifcation:   			Open up a window and run:   			$ monitor system   7 		When hang occurs , note number of processes in MWAIT.l 		What or why?  7 		Write a cheap DCL that finds and displays the MWAITs:l     			$ counter = 0
 			$ loop: 			$ counter = counter + 1  			$ show system/output=show.tmp 			$ search show.tmp "RWMPB" 			$ delete show.tmp;*, 			$ if counter .gt. 12 then exit ! Whatever 			$ wait 00:00:05 			$ goto loop    = 	That is my guess at your "hangs".  Very annoying.  Solution?fA 	Buy more memory.   Remember they are periodic hangs, most likelysA 	at peak when memory is tight.   My suggested solution posted outp? 	here about 1-2 years ago was to add bit values to MMG_CTLFLAGSo5 	to kick in 2 other algorithms.  The "kinder" ticker.e  > 	The current ticker and its 25% trim REGARDLESS of free memory@ 	is freeing up way too much in one swipe.  A "kinder" ticker mayF 	operate on the principal of when memory gets down to 4 times FREEGOAL? 	free, trim 25% of mostly idle processes UNTIL 8 times FREEGOALd= 	is free.  Maybe you get a bunch of oscillation but you avoid = 	the annoying system hangs.  THAT is *very* difficult to dealM? 	with when sitting in the system manager hot seat.  However, itmB 	does help to find the cause and let management know something can 	be done about it.  B 	Others will mention we need more information about your slowness.   		Are you maxed out on CPU?t 		etc.  A 	One other thing... it may be tricky to catch the hang in action.sA 	Be at the ready during peaks and you may just want to tweek thatuC 	pitiful DCL code above to save all the show systems and then latertG 	search the *.tmp files for "RWMPB"  (Resource Wait Modified Page Busy) # 	i.e. busy writing to the pagefile.a  > 	Also, who is "hung"?  Someone that is in the process of beingG 	paged out... one of the unlucky ones that is "mostly" idle.  That justfC 	so happens to try to do something when the paging is taking place.l@ 	I typically was seeing 7-10 in RWMPB out of a user community of@ 	600-700 interactive.   This may be like a tide (don't know, but? 	could be) whereby when their memory is written out they becomeuF 	responsive.  So the "hang" situation has a lot of potential customers= 	in that timeframe.  These are folks that would typically log-B 	on, read mail , run reports and turn their attention to somethingB 	else.  To be dubbed "mostly idle" by the ticker, you have to have? 	not used 1% CPU in the prior 30 seconds.  That is a wide fieldi< 	there.  So many potential candidates.  Since the ticker is @ 	going after everyone that meets that criteria, viola!  Many are? 	being paged right when they decide to run another report, maked 	another entry or whatever.n  ? 	That is why you may not be seeing it and others are... you mayrB 	have all active windows and be telling yourself and others: "hey,# 	I'm not hung up!"  True statement.   H 	You can be part of the fun at times too by opening up another session, G 	maybe editing a file and not touching it then when everyone is "hung"  G 	see if you can scroll up and down.. you may be in the middle of being 9C 	paged out and paged back in again... you too are "hung up."  I washD 	lucky a few times and I too was "hung up!"  So I could let everyoneB 	else know that I was "hung up" and we could share in the sadness.  > 	Look for a follow-up from the earlier analysis I did of this,A 	hopefully I still have it.  It is more straight forward than thes 	above.    				Robo   ------------------------------  # Date: Thu, 08 Jun 2000 04:40:44 GMT  From: bab@hotmail.com (bab)  Subject: tcp/ip printing0 Message-ID: <393f2396.30171764@nntp.hip.cam.org>   Hi all,s  @ got a problem last weekend, I installed tcp/ip services on alpha Openvmst$ 7.2. so decnet IV and tcp/ip are up.  	 problem 1i  B There is a xerox printer on the lan defined ip. I can ping it, but printrD queue remain starting. Do I have to enable routing? If so, should it beC gated or routing (it is connected to internet thru a firewall) ? IsS there aaE port setup I am missing? I used tcpip$ldr_smb.exe as processor. Tried  with dcps, it did not work either.a   problem2  = running pathworks 5 on that server, I switched transport from  decnet,netbeuiF to decnet,netbeui,tcpip. rebooted the machine. the last decserver 90 L wasr* not able to connect to the server anymore.   ------------------------------  # Date: Thu, 08 Jun 2000 04:41:56 GMTA From: bab@hotmail.com (bab)o Subject: tcp/ip printing0 Message-ID: <393f23f2.30263582@nntp.hip.cam.org>   Hi all,   @ got a problem last weekend, I installed tcp/ip services on alpha Openvmsl$ 7.2. so decnet IV and tcp/ip are up.  	 problem 1t  B There is a xerox printer on the lan defined ip. I can ping it, but print D queue remain starting. Do I have to enable routing? If so, should it beC gated or routing (it is connected to internet thru a firewall) ? Isn there a E port setup I am missing? I used tcpip$ldr_smb.exe as processor. Tried  with dcps, it did not work either.d   problem2  = running pathworks 5 on that server, I switched transport fromk decnet,netbeuiF to decnet,netbeui,tcpip. rebooted the machine. the last decserver 90 L wasc* not able to connect to the server anymore.   ------------------------------  % Date: Wed, 07 Jun 2000 23:39:04 -0400.2 From: "Richard B. Gilbert" <DRAGON@compuserve.com>A Subject: Re: What is the simplest way to network two VMS systems?-7 Message-ID: <200006072339_MC2-A7ED-7DFB@compuserve.com>p  ( Message text written by "Phil Tregoning"9 >JF Mezei <jfmezei.spamnot@videotron.ca> wrote in articleV# <393B366F.9F097B69@videotron.ca>...  > Dan wrote: > > =   H > > I installed DEC Net Plus instead of phase 4, compaq likes to suggest this.t! > > is this actually a good idea?e > =o  J > DECNET-PLUS is overly complex and a pain to manage. Avoid it unless you= rt > network requires it. > =n    J Lots of people seem to say this in comp.os.vms. I've been using DECnet/pl= us< for a few years now, and quite like it. Am I just perverse?<           Yes!   ------------------------------  % Date: Thu, 08 Jun 2000 06:55:28 +0800r- From: David B Sneddon <dbsneddon@bigpond.com>-  Subject: Re: X25 and psi$k_reset+ Message-ID: <393ED2E0.AEC4E272@bigpond.com>    JF Mezei wrote:s >  > moi_is_me wrote:A > >    I have been asked to port some X25 code to TCP. ThroughoutmG > >    the offending code, are calls to qiow to "reset the circuit".ie.KG > >    psi$k_reset (p4) The QIOW call itself, is wrapped within a timer  > >t& > >    Q ... Was does psi$k_reset do ? > O > I am not sure what the psi implementation of psi$k_reset does, but I stronglyeH > systect that it is an X.29 command which is sent to the other machine. > P > As such, its does a few things, such as asking the remote PAD to drop any dataP > not yet transmitted. (think of a 300 baud terminal connected to a PAD and fromL > there  a high speed line to the computer, the PAD will buffer data for theL > slower 300 baud terminal, a reset will force that buffer to be emptied and+ > bring the circuit to a "brand new state".0 > M > I am not sure what the implications are for a computer to computer (X.25 to: > X.25) connection thought.  > J > I would think that this would be equivalent to the "kill character" on a > telnet session.  > - > But I am not 100% sure of any of the above.    from the manual... -----------tF Use the IO$_NETCONTROL call, with a parameter of PSI$K_RESET, to reset9 a virtual circuit and confirm receipt of a reset request.rF The call resets the virtual circuit to its initial conditions (and all? pending messages are discarded) or confirms receipt of a reset.gB Note the a return status of SS$_NORMAL does not guarantee that theK remote DTE receives the diagnostic code.  Reception of the diagnostic code eD may be prevented if a collision of resets occurs within the network.
 ----------   -- d Regards, Dave.oI -------------------------------------------------------------------------oI David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.comsI DBS software at ...   http://www.users.bigpond.com/dbsneddon/software.htm2I "Life is what happens to you while you're busy making other plans" Lennone   ------------------------------   Date: 7 Jun 2000 19:54:06 GMTr* From: helbig@astro.rug.nl (Phillip Helbig)@ Subject: xxx.EDT to convert various forms of special characters?. Message-ID: <8hm98u$6t9$1@info.service.rug.nl>   Hi all,o    @ Despite Linus Torvald's crusade to the rights of the eighth bit,D seven-bit printable ASCII is often the lowest common denominator, atE least for things like email.  Even if an eighth bit is allowed, it is D not clear which character set is intended, unless one gets into MIME etc. a  A Having to deal daily with various formats---LaTeX or pseudo-LaTeXp5 (writing  as "A or  as "s, for example), eight-bit,oI X-MIME-auto-converted stuff (=E4,=DF), HTML (&Auml;, &szlig;) as well as -E real 8-bit characters from the DEC Multinational Character Set (HELP yE FORTRAN Character_Sets has some nice tables) I'm thinking of writing  G some .EDT files so that I can have a command procedure which will edit aG the file with /EDT/COMMAND=xxx.EDT with xxx one of 16 files which will eD convert between the proper formats with substitute/whole.  Before I * write these: has anyone already done this?  G I'd probably have one format for my own use---probably LaTeX, since it HG needs the fewest keystrokes to type, can be intuitively understood raw VF and easily incorporated into a LaTeX document---and convert to one of F the other 4 as needed, but I would need another four files to convert D stuff from elsewhere to my format, so I might as well do all 16 and  spare double conversions.t  D It appears to be easy to convert TO 8-bit characters with EDT; what  about FROM them?     --M Phillip Helbig                       Email .............. helbig@astro.rug.nlaM Kapteyn Instituut                    Email ................. helbig@man.ac.uk M Rijksuniversiteit Groningen          Tel. ................... +31 50 363 4067 M Postbus 800                          Fax .................... +31 50 363 6100bM NL-9700 AV Groningen                 Web ... http://www.astro.rug.nl/~helbig/   5 My opinions are not necessarily those of my employer.   N <A HREF=" http://gladia.astro.rug.nl:8000/helbig/hire/hire.html ">HIRE ME!</A>   ------------------------------   End of INFO-VAX 2000.318 ************************