/ INFO-VAX	Mon, 03 Jan 2005	Volume 2005 : Issue 5       Contents:= Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images = Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images  Dell Replaces Carly? Re: Dell Replaces Carly? Re: Supported Itanium hardware Re: Supported Itanium hardware  F ----------------------------------------------------------------------  % Date: Sun, 02 Jan 2005 14:06:54 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large ImagesB Message-ID: <1104692227.9dff164eda9fe32ee73623d475eecb08@teranews>   Neil Rieck wrote: K > I agree and I suggested this to the folks at HP. Put the OpenVMS-specific E > routines in an optional loadable Apache module called "MOD_RMS" (or  > whatever).    L How does Apache run on IBM mainframes ?   IBM also has strange file formats.  N Seems to me that DEC could simply have its own IO routines to read any/all VMSQ files, and those routines would then have the smarts to do what needs to be done.   0 > For people who desire the UNIX style speed up,  H Does Apache have a cache ? If so, once the VMS file has been read, there should be no performance hit.     J File size (content bytes) should still be something which is stored in theN file header. Modern applications such as web, IMAP etc all want a file's size.   ------------------------------  $ Date: Sun, 2 Jan 2005 17:01:50 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images6 Message-ID: <ib_Bd.86$7n1.14581@news20.bellglobal.com>  : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message< news:1104692227.9dff164eda9fe32ee73623d475eecb08@teranews... > Neil Rieck wrote:  [...snip...] > E > How does Apache run on IBM mainframes ?   IBM also has strange file 
 > formats.   Don't know.    > L > Seems to me that DEC could simply have its own IO routines to read any/all > VMS J > files, and those routines would then have the smarts to do what needs to
 > be done. >   L They seem to be getting bogged down in this area every time they port ApacheH to OpenVMS. That's why I thought it might be better to move this OpenVMSI stuff into a loadable Apache module. (of course I'm not the guy doing the + coding so I may be over simplifying things)    > J > Does Apache have a cache ? If so, once the VMS file has been read, there > should be no performance hit.  > L > File size (content bytes) should still be something which is stored in theJ > file header. Modern applications such as web, IMAP etc all want a file's > size.  >   J Remember that HTTP 1.0 is connectionless. (HTTP 1.1 persistent connectionsH are possible for a short period of time if client and server conditions  allow L it, but that is another issue). That said, it is possible that the first 2KBK of a file will be sent to your browser by Apache's child process #1 and the I next 2KB by child process #2 and so on. So the question remains "how does < the second process find byte #2001 of an OpenVMS text file?"  H With variable length records, the precise RMS structure of a file is notK known to the basic file system and so there is no way to easily compute the H location of byte #2001 begins without invoking RMS routines like readingK records until you get to #2001. Click here for a very simple hacker view of  RMS-based text files. = http://www3.sympatico.ca/n.rieck/docs/openvms_text_files.html H So you will see a small amount of file overhead is required to store RMSL data (each line in this case) including some padding bytes to word-align theL length fields. The 32-bit length bits need to be dropped for each record andK replaced with 16-bits containing <cr> + <lf> while the padding bytes would  L need to be dropped entirely. This could only be done by reading through the 4 RMS records and computing the byte count on the fly.  
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  % Date: Sun, 02 Jan 2005 19:44:54 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large ImagesB Message-ID: <1104712485.78cef4d519b89cd47b5de1441daa96ba@teranews>   Neil Rieck wrote: N > They seem to be getting bogged down in this area every time they port Apache > to OpenVMS.   J Which is the wrong thing to do. They should have adopted open sourcing andK make VMS part of the real Apache source code so that thety wouldn't have to  re-port every new version.  J They wanted to make a proprietary version of Apache (will it be renamed to4 HPSWS ?), and now they are reaping the higher costs.  L > Remember that HTTP 1.0 is connectionless. (HTTP 1.1 persistent connectionsI > are possible for a short period of time if client and server conditions  > allow   J And re-usable connections do require to provide the client with exact byteN count that you will be feeding it so that it knows when the data has been sent# and a new HTTP request can be made.     N > it, but that is another issue). That said, it is possible that the first 2KBM > of a file will be sent to your browser by Apache's child process #1 and the * > next 2KB by child process #2 and so on.   J There is a difference between re-using connections (requires you send fullL length before sending data) and content-range requests where client requestsN only portion of a file (byte range). The PDF plug ins will request byte ranges
 for instance.   I However, for an HTML file, requesting byte range doesn't make much sense.   M > need to be dropped entirely. This could only be done by reading through the 6 > RMS records and computing the byte count on the fly.  J But if you are going to feed the file anyways, then reading it in one shot makes sense.  E Of course, of apache uses separate processes, it makes caching rather M impossible. This is perhaps an area where the OSU server is far superior with  its threaded approach.   ------------------------------   Date: 2 Jan 2005 17:10:10 -0800 ) From: "Rich Faust" <faustrj@fastspot.net> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large ImagesC Message-ID: <1104714610.085742.109390@f14g2000cwb.googlegroups.com>   G In my original post I asked if anyone was successfully running CSWS 2.0 F with MultiNet 5.0.  So far there are no replies in the affirmative.  IG may not have a chance to try your suggestion and install TCPIP services D until next weekend.  Maybe someone will step forward in the meantimeG and confirm that the CSWS 2.0/MultiNet 5.0 combination will work and/or 8 provide some additional hints to cajole it into service.  B Anyone from Process Software following this thread with first-handF knowledge that CSWS V2.0 functions without incident with MultiNet 5.0?  ! Thanks for all the posts to date. 
 Rich Faust OpenVMS Hobbyist Richmond, TX   ------------------------------  # Date: Mon, 03 Jan 2005 01:13:24 GMT L From: winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing)F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images6 Message-ID: <00A3D476.D47CECFA@SSRL.SLAC.STANFORD.EDU>  r In article <1104712485.78cef4d519b89cd47b5de1441daa96ba@teranews>, JF Mezei <jfmezei.spamnot@teksavvy.com> writes: >Neil Rieck wrote:O >> They seem to be getting bogged down in this area every time they port Apache  >> to OpenVMS.   > K >Which is the wrong thing to do. They should have adopted open sourcing and L >make VMS part of the real Apache source code so that thety wouldn't have to >re-port every new version.   E I understand they're working on this.  As I understand it, The STM-LF M requirement comes in with  Apache 2.0 and the way it's now mapping files into I memory for performance.  They'll have to write some kind of work-around - O either, as Neil suggested, an add-on module, or some conditionalized additional 5 code in the file access routines - to make this work.   K >They wanted to make a proprietary version of Apache (will it be renamed to 5 >HPSWS ?), and now they are reaping the higher costs.   L No, they wanted to make a version that would run on VMS.  You have the issueL exactly backward; the STM-LF pain comes from their _not_ redoing it for VMS.N (And they said at the VMS Boot Camp that they _would_ redo it for VMS once the  Itanium porting work is done.)     > F >Of course, of apache uses separate processes, it makes caching ratherN >impossible. This is perhaps an area where the OSU server is far superior with >its threaded approach.   N WASD also has a tuneable cache, and seems to generally go significantly faster than OSU or SWS.  M Actually, if Apache development had been particularly interested in cacheing, = they could have used shared memory among different processes.   M Most Apache development takes place on Unix/Linux/xBSD, which have for a long G time had heavily-cached file systems.  The developers didn't think the  E incremental gain to Apache performance by doing cacheing on top of a  N fully-cached file system was worth the development effort.  XFC addresses this
 issue on VMS.    -- Alan    ------------------------------  % Date: Sun, 02 Jan 2005 23:16:49 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large ImagesB Message-ID: <1104725192.fe09f0edb9fd228326e3d630dbdee93c@teranews>  , Alan Winston - SSRL Central Computing wrote:O > Most Apache development takes place on Unix/Linux/xBSD, which have for a long ( > time had heavily-cached file systems.   & Fair enough. I hadn't thought of that.  K OK so lets turn the table. What would be the impact of changing VMS to have J STMLF as default file format for DCL commands such as OPEN, CREATE and for9 TPU/EDT ? What other utilities would need to be changed ?   I Or are there too many utilities that are "mature" and which HP refuses to N allocate development money on that such a change is really impossible on VMS ?   ------------------------------  $ Date: Mon, 3 Jan 2005 00:30:24 -0500) From: "Neil Rieck" <n.rieck@sympatico.ca> F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images9 Message-ID: <RL4Cd.2438$7n1.138028@news20.bellglobal.com>   ; "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message  < news:1104725192.fe09f0edb9fd228326e3d630dbdee93c@teranews.... > Alan Winston - SSRL Central Computing wrote:L >> Most Apache development takes place on Unix/Linux/xBSD, which have for a  >> long ( >> time had heavily-cached file systems. > ( > Fair enough. I hadn't thought of that. > I > OK so lets turn the table. What would be the impact of changing VMS to   > haveL > STMLF as default file format for DCL commands such as OPEN, CREATE and for; > TPU/EDT ? What other utilities would need to be changed ?  > K > Or are there too many utilities that are "mature" and which HP refuses to K > allocate development money on that such a change is really impossible on   > VMS ?   ' In some ways, OpenVMS is already there.   I I do not know how to create a stream_lf file directly from EDT. However,  L once you go through the little ritual of converting your file to stream_lf, L EDT will allow you to modify the file and the new file version will also be 
 stream_lf.  E Other apps that generate static HTML web pages to be served up (like  4 reports?) would need their open statements modified.    
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.9 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html     ------------------------------  % Date: Mon, 03 Jan 2005 07:40:38 +0100  From: Dirk Munk <munk@home.nl>F Subject: Re: CSWS 2.0 + MultiNet 5.0 - Problem Displaying Large Images2 Message-ID: <crapd6$9t1$1@news4.zwoll1.ov.home.nl>   Neil Rieck wrote: 7 > "Rich Faust" <faustrj@fastspot.net> wrote in message  ? > news:1104626045.060255.211490@z14g2000cwz.googlegroups.com...  > E >>I wondered about that, so before I posted my original note I took a I >>closer look at APACHE$COMMON:[000000]APACHE$CONVERT_STREAMLF.COM.  This I >>tool is called by APACHE$MENU to convert a directory tree to Stream_LF. ; >>The convert procedure references the following data file:  >> >># APACHE$CVT_TYPES.DAT >># * >>#       File types that get converted by >>APACHE$CONVERT_STREAMLF.COM  >>#  >>> >>.HTM*           #All HTML files (.HTM, .HTML, .HTML.FR, etc)' >>.SHTML          #Server-side includes   >>.TXT            #All TXT files >>H >>I take the contents of this file to mean that .JPG files should not beF >>converted to Stream_LF, so I don't believe the format of the file is >>the problem. >> >>Rich Faust >>OpenVMS Hobbyist >>Richmond, TX >> >  > L > Your analysis is correct. The script doesn't look for anything other than B > VMS text files which means JPG files don't need to be converted.  Q Any attempt to *convert* the .JPG files would result in corrupted files. So even  N if .JPG files should be in Stream-LF format, this procedure could not be used.     > M > Have you checked your Apache error log? (You'll need to restart the server   > to look at it) >  > Last resort test: K > I see that you are a hobbyist. Try shutting down MultiNet and installing  M > HP's "TCPIP Services for OpenVMS". If your problem goes away then you know  O > that there is a bug with the stack. (this is how I solved my TCPware related  " > problem which I documented here I > http://www3.sympatico.ca/n.rieck/docs/csws_tips.html#intermittent-cgi )  >  > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.: > http://www3.sympatico.ca/n.rieck/links/cool_openvms.html >  >    ------------------------------  % Date: Sun, 02 Jan 2005 15:23:53 -0800 # From: "Tom Linden" <tom@kednos.com>  Subject: Dell Replaces Carly? ( Message-ID: <opsjz0l3aazgicya@hyrrokkin>  
 read this....   / http://moneycentral.msn.com/content/P101826.asp    i really like this part!  K   The Hewlett-Packard (HPQ, news, msgs) board ousts Chief Executive Carly   K Fiorina and replaces her with a former top executive at Dell (DELL, news,   K msgs). The new management team pares down the production and distribution   H infrastructure, revamps sales and marketing and produces a comeback in  F just six quarters. Growing acknowledgment that the change is working  K catches investors by surprise, but the stock ultimately becomes a leading   3 name in the Dow Jones industrials ($INDU) for 2005.    --  C Using Opera's revolutionary e-mail client: http://www.opera.com/m2/    ------------------------------  $ Date: Sun, 2 Jan 2005 18:48:53 -0500# From: "John Smith" <a@nonymous.com> ! Subject: Re: Dell Replaces Carly? , Message-ID: <FcSdnRcO1qfDFUXcRVn-jg@igs.net>   Tom Linden wrote:  > read this....  > 1 > http://moneycentral.msn.com/content/P101826.asp  >  > i really like this part! > E >   The Hewlett-Packard (HPQ, news, msgs) board ousts Chief Executive D > Carly Fiorina and replaces her with a former top executive at DellG > (DELL, news, msgs). The new management team pares down the production B > and distribution infrastructure, revamps sales and marketing andG > produces a comeback in just six quarters. Growing acknowledgment that D > the change is working catches investors by surprise, but the stock@ > ultimately becomes a leading name in the Dow Jones industrials > ($INDU) for 2005.     ) I know that I could do a better job.  :-)    ------------------------------   Date: 2 Jan 2005 16:42:32 -0800  From: "Ben" <cellctl@lycos.com> ' Subject: Re: Supported Itanium hardware C Message-ID: <1104712952.669867.281510@c13g2000cwb.googlegroups.com>   F > The E8.2 (external field test) kit does contain the needed bits, but I don't D > know if the i2000 will boot on E8.2.  It clearly is not supported. >   E External Field Test kit?  Is that different from the FT/SDK available G on the OVMS/Itanium page on HP's website?  If yes, how would one obtain  it?  Thanks for the tip!!    ------------------------------   Date: 2 Jan 2005 18:51:58 -0600 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ' Subject: Re: Supported Itanium hardware 3 Message-ID: <wRd8Njcjiywr@eisner.encompasserve.org>   e In article <1104712952.669867.281510@c13g2000cwb.googlegroups.com>, "Ben" <cellctl@lycos.com> writes: G >> The E8.2 (external field test) kit does contain the needed bits, but 	 > I don't E >> know if the i2000 will boot on E8.2.  It clearly is not supported.  >> > G > External Field Test kit?  Is that different from the FT/SDK available + > on the OVMS/Itanium page on HP's website?    No.    ------------------------------   End of INFO-VAX 2005.005 ************************