Date: Sun, 4 Dec 2005 00:04:23 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 659 ************************************************** Saturday 03 December 2005 Number 659 ************************************************** Subjects for today 1 Re: Subversion : Dave Bamford 2 Re: strange behaviour with file read : Jon Saxton 3 Re: strange behaviour with file read : Thomas Dickey 4 Re: strange behaviour with file read : Dave Bamford 5 Re: strange behaviour with file read : Thomas Dickey 6 Re: strange behaviour with file read : Thomas Dickey 7 Zope : John Poltorak 8 Building Python in a UnixOS/2 environment : John Poltorak 9 DosLoadModule problem : Paul Smedley 10 ZLIB : John Poltorak 11 Build notes for Apache v2 : John Poltorak **= Email 1 ==========================** Date: Fri, 02 Dec 2005 19:32:20 +0000 From: Dave Bamford Subject: Re: Subversion Franz Bakan wrote: >On Thu, 01 Dec 2005 21:31:32 +0000, Dave Bamford wrote: > > > >>My problem is now that the autogenerated makefiles have a construct that >>is not >>understood by the make I am using 3.76.1 >>it has a -D switch to pass a dependancy to sub make files (RELEASE or DEBUG) >>so I have to edit all the Makefiles to make it work. >> >> > >or you might try to change to a newer make version: > >http://unix.os2site.com/sw/pub/source/make/make-3.81beta3-r2.zip > >Franz > > > Tried to build this make but after running configure I get when I make AR=emxomfar make: *** Recursive variable `AUTOCONF' references itself (eventually). Stop. I have seen this before but can't remember what the cause was? Dave. **= Email 2 ==========================** Date: Fri, 02 Dec 2005 16:53:21 -0500 From: Jon Saxton Subject: Re: strange behaviour with file read Stefan.Neis at t-online.de wrote: >Yes, unix platforms use a simple "LF" for line break in text files, >not the insane "CRLF" combination used by (Win)DOS and OS/2. > > I wouldn't use the term "insane" to describe the CR-LF convention. It makes a lot of sense when viewed from a historical perspective. OS/2 inherited it from DOS which inherited it from CP/M which in turn inherited it from some DEC predecessor. One of the early I/O devices suited to interaction with humans was the teletype and that required both control codes to advance to the next line. The carriage return moved the print head to the first position and the line feed rolled the platen up. The machinery involved in each operation was completely different and quite independent. By embedding both control characters in a file, the file could be sent to the teletype "as is" and it would display correctly. With just a linefeed at the end of each line a series of single-word lines would display something like this where you see the platen advance but the print head remains in the same position. If those early systems had adopted a UNIX-like convention of having a single end-of-line marker then raw dumping to a teletype would not have been possible. Operating system or (more likely) application programs would have to intercept the data stream and when an end-of-line marker was detected, send a CR and a LF to the teletype. On those early systems with tiny amounts of memory, that sort of thing would have been an imposition. The CR LF ordering is a hardware optimization. The carriage return was typically the more time consuming operation but because (as mentioned above) the machinery was independent, the platen could roll up while the print head was still moving. With the advent of cursor-addressable character-mode CRT displays the need for separate control codes for carriage return and line feed vanished and those systems which still used teletypes could relegate the translation of EOL to CR/LF to the output device driver. Meanwhile DOS had inherited the CP/M convention a wealth of ported software which expected dual-character EOL markers. -- Jon Saxton Shopkeeper, Numismatist-in-training Developer of cross-platform software for UNIX, OS/2 and Windows U.S. Agent for Triton Technologies International Ltd http://www.triton.vg/ **= Email 3 ==========================** Date: Fri, 2 Dec 2005 17:13:14 -0500 (EST) From: Thomas Dickey Subject: Re: strange behaviour with file read On Fri, 2 Dec 2005, Jon Saxton wrote: > where you see the platen advance but the print head remains in the same > position. If those early systems had adopted a UNIX-like convention of It would have been unlikely for the "DEC predecessor" to have adopted a UNIX-like convention, simply because UNIX came later. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 4 ==========================** Date: Sat, 03 Dec 2005 00:07:09 +0000 From: Dave Bamford Subject: Re: strange behaviour with file read Thomas Dickey wrote: > On Fri, 2 Dec 2005, Jon Saxton wrote: > >> where you see the platen advance but the print head remains in the >> same position. If those early systems had adopted a UNIX-like >> convention of > > > It would have been unlikely for the "DEC predecessor" to have adopted > a UNIX-like convention, simply because UNIX came later. > Having been brought up on dec. I seem to remember that vms character files are represented differently. The EOL is still 2 bytes but the number of characters in the line is also stored somewhere. The word insane came from the developer of the code I am porting, he is FreeBSD or OpenBSD. I remember teletypes too, I was at univercity and they were a great improvement on punched cards :-) Dave. **= Email 5 ==========================** Date: Fri, 2 Dec 2005 20:16:12 -0500 (EST) From: Thomas Dickey Subject: Re: strange behaviour with file read On Fri, 2 Dec 2005, Thomas Dickey wrote: > VMS came out toward the end of the 70's. I was using RT-11 from around 1972. > UNIX was in-house only til the mid 70's. hmm - no, I was using paper tape in 1972. RT-11 came along in early 1975 (there were some other variations in between - 8-track cassettes ;-) Now that I stop to think, RT-11's date-origin was iirc January 1975 (will google around to refresh memory ;-) -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 6 ==========================** Date: Fri, 2 Dec 2005 20:12:35 -0500 (EST) From: Thomas Dickey Subject: Re: strange behaviour with file read On Sat, 3 Dec 2005, Dave Bamford wrote: > Thomas Dickey wrote: > >> On Fri, 2 Dec 2005, Jon Saxton wrote: >> >>> where you see the platen advance but the print head remains in the same >>> position. If those early systems had adopted a UNIX-like convention of >> >> >> It would have been unlikely for the "DEC predecessor" to have adopted >> a UNIX-like convention, simply because UNIX came later. >> > Having been brought up on dec. I seem to remember that vms character files > are represented differently. The EOL is still 2 bytes but the number of > characters VMS came out toward the end of the 70's. I was using RT-11 from around 1972. UNIX was in-house only til the mid 70's. VMS's I/O is supposed to be based on RSX-11M (one of my associates was working with that in 1975-1976), but what I saw of it didn't resemble VMS, so I don't know how much was "based on". My initial impression of the newlines was that it was removing some capability from applications - though the ugly stty interface does sort of work around _that_. > in the line is also stored somewhere. > The word insane came from the developer of the code I am porting, he is > FreeBSD or OpenBSD. I remember teletypes too, I was at univercity and they > were a great improvement on punched cards :-) > > Dave. > > -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 7 ==========================** Date: Sat, 3 Dec 2005 09:34:42 +0000 From: John Poltorak Subject: Zope Is anyone using Zope on OS/2? I've had it working and I know there was a Zope-OS2 mailing list, but I'm not sure how many other people ever used it. I'm hoping to get the latest version - 2.8.4 working and then hopefully install zForum. Is anyone else interested in using Zope? For anyone who doesn't know Zope is a very powerful, fully integrated web publishing environment which includes its own web server. See:- http://zope.org -- John **= Email 8 ==========================** Date: Sat, 3 Dec 2005 11:27:59 +0000 From: John Poltorak Subject: Building Python in a UnixOS/2 environment What changes do I need to make when building Python in a Unix environment on OS/2? Is it just a matter of changing some paths in the Makefile? -- John **= Email 9 ==========================** Date: Sat, 03 Dec 2005 22:21:05 +1030 From: Paul Smedley Subject: DosLoadModule problem Hi All, As many would we aware, I'm working on building Apache2 using Innotek Libc. Getting it to compile was pretty straightforward, largely due to Brian Havard's efforts in getting it working with EMX. I'm currently grappling with loading Apache2 modules dynamically. The way Apache2 works, it loads all modules, then unloads them, then loads them again (don't ask me why!!). DosLoadModule works fine the first time, but on the second call returns a return code of 87 - ie incorrect parameter. For the life of me I can't work out why it works OK the first time and not the second. The code in question is: APR_DECLARE(apr_status_t) apr_dso_load(apr_dso_handle_t **res_handle, const char *path, apr_pool_t *ctx) { char failed_module[200]; HMODULE handle; int rc; *res_handle = apr_pcalloc(ctx, sizeof(**res_handle)); (*res_handle)->cont = ctx; (*res_handle)->load_error = APR_SUCCESS; (*res_handle)->failed_module = NULL; if ((rc = DosLoadModule(failed_module, sizeof(failed_module), path, &handle)) != 0) { (*res_handle)->load_error = APR_FROM_OS_ERROR(rc); (*res_handle)->failed_module = apr_pstrdup(ctx, failed_module); return APR_FROM_OS_ERROR(rc); } (*res_handle)->handle = handle; apr_pool_cleanup_register(ctx, *res_handle, dso_cleanup, apr_pool_cleanup_null); return APR_SUCCESS; } The Apache2 errorlog states: Cannot load E:/apache2/modules/access.dll into server: The parameter is incorrect. ((\xfb\t) Any ideas? Cheers, Paul. **= Email 10 ==========================** Date: Sat, 3 Dec 2005 12:09:31 +0000 From: John Poltorak Subject: ZLIB Has anyone built the latest ZLIB on OS/2? -- John **= Email 11 ==========================** Date: Sat, 3 Dec 2005 12:39:36 +0000 From: John Poltorak Subject: Build notes for Apache v2 If anyone has any build notes for Apache v2, I'd appreciate a copy of them. -- John