From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Sun, 8 Dec 2002 04:43:03 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 392 ************************************************** Saturday 07 December 2002 Number 392 ************************************************** Subjects for today 1 Re: config.site : =?ISO-2022-JP?B?GyRCXzdFREBQISE9ZxsoQg==?= 2 Re: Make bug?, was: got lost : Stefan Neis 3 Re: config.site : Andreas Buening 4 Re: Installing autoconf : Andreas Buening 5 Re: Ncurses build error : John Poltorak 6 Re: Installing autoconf : John Poltorak 7 Re: mailman cgi : Ted Sikora 8 GCI Errors : Ted Sikora 9 Ncurses UnixOS/2 package : John Poltorak 10 Re: Ncurses build error : Thomas Dickey 11 Re: Installing autoconf : Thomas Dickey 12 Re: Ncurses UnixOS/2 package : Thomas Dickey 13 Re: GCI Errors : Thomas Dickey 14 Re: GCI Errors : Ted Sikora 15 Re: GCI Errors : Ted Sikora 16 Re: mailman cgi : John Poltorak 17 New NCURSES : John Poltorak 18 Re: New NCURSES : John Poltorak 19 Re: New NCURSES : John Poltorak 20 Re: GCI Errors : John Poltorak 21 Mailman: 'module' object has no attribute 'link' : Ted Sikora 22 Re: Mailman: 'module' object has no attribute 'link' : Ted Sikora 23 Re: GCI Errors : John Poltorak 24 Make v3.80 : John Poltorak 25 Re: Installing autoconf : Stefan Neis 26 Autoconf build error - multiple target patterns : John Poltorak **= Email 1 ==========================** Date: Sun, 08 Dec 2002 00:54:22 +0900 From: =?ISO-2022-JP?B?GyRCXzdFREBQISE9ZxsoQg==?= Subject: Re: config.site Ted Sikora wrote: >PATH=`cmd.exe /c "echo %PATH%" | sed -e 's at \\\\ at / at g'` >PATH_SEPARATOR=';' >Is this really needed? I say YES. Without converting '\' to '/', you will fail to run some "configure" scripts. # Jun Sawataishi # # My recent port: Bash 2.05b, Texinfo 4.3 **= Email 2 ==========================** Date: Sun, 8 Dec 2002 01:44:15 +0100 (CET) From: Stefan Neis Subject: Re: Make bug?, was: got lost On Sat, 7 Dec 2002, Ken Ames wrote: > hi Stefan, > > I am running emx 0.9d with fix4 and gcc 2.8.1 Same here. Maybe a difference in make or shell version... > - it does this on a fresh > reboot (only window open) hmmm, maybe file handles run out or something? But I don't have that problem. FILES is set to 80 in my config.sys, EMXOPT is not being set, GCCOPT=-pipe is being set and - surprise, at least to myself - GCCLOAD is not being set. Thinking about it, I seem to remember I had some problems with GCCLOAD being set, so if it's set on your system, that might cause the problem... :-( > > running > > for %1 in (*.d) do at (sed "s/\//\\/g" < %1 > dep.sed && copy dep.sed %1) > >when needed. > oh no! shell script! (I think) Nope. A plain command for cmd.exe... > I am completely lost in that It's doing a loop on all .d files, replacing all "/" with "\". > so could you > be a bit more specific as to where and how I use that? Well, just like so: [Prompt] make .... Lots of output .... [Prompt] for %1 in (*.d) do at (sed "s/\//\\/g" < %1 > dep.sed && copy dep.sed %1) ... Again lots of output... it does take a while (maybe some minutes, depending on the speed of your hard disk). [Prompt] make > oh, BTW, do you use the wxwindows irc channel? No, sorry, no irc over here. E-mail is both fast enough and distracting enough for my taste ... ;-) Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 3 ==========================** Date: Sun, 08 Dec 2002 01:47:26 +0100 From: Andreas Buening Subject: Re: config.site John Poltorak wrote: > > On Sat, Dec 07, 2002 at 12:38:29PM -0500, Thomas Dickey wrote: [whether PATH_SEPARATOR works] > > That would be nice; however at least one of the autoconf maintainers has > > stated that he refuses to make changes for path-separator. > > I've looked through various autom4te.cache/output.0 files and they seem to > contain code like this:- [some code] > I'm not exactly certain about how this works, but it seems to me that > $PATH_SEPARATOR is getting set appropriately if not already set, BICBW. That's true. Thomas, could you explain your statement, please? Who refuses to make which kind of changes? Bye, Andreas -- One OS to rule them all, One OS to find them, One OS to bring them all and in the darkness bind them In the Land of Mordor where the Shadows lie. **= Email 4 ==========================** Date: Sun, 08 Dec 2002 01:47:32 +0100 From: Andreas Buening Subject: Re: Installing autoconf John Poltorak wrote: > > On Sat, Dec 07, 2002 at 12:49:46AM +0100, Andreas Buening wrote: > > John Poltorak wrote: > > > > > > > [autoconf] > > > > > Is there any prospect of the OS/2 patches in 2.53bR2 being integrated into > > > the mainstream version? > > > > The "patches" aren't anything special. They set > > ac_executable_extensions, replace '\' by '/' in your PATH > > and change the behaviour of AC_CHECK_LIB slightly so that > > you don't get drowned by linker warnings if you use link386. > > It's great that there are so few patches, although I would have thought > some of the would also apply to Win32 so should be acceptable for > inclusion. This one for instance (general.m4) :- > > - [[\\/$]]* | ?:[[\\/]]* | NONE | '' ) ;; > + [[\\/$]]* | ?:[[\\/]]* | ?: | NONE | '' ) ;; Okay, this one I forgot to mention. I have no idea what the Win32 guys need. It look as if the cygwin people had their own autoconf. > > They don't affect the usability if you've set up your environment > > correctly. > > But how can anyone ensure that their environment has been set up > correctly in the first place? Can everything necessary be handled by > config.site? Not everything. > > To answer your question, I don't think they would > > be accepted as they are. > > The problem about long term maintenance remains. > > Are the patches applicable to every release of autoconf? I think so, unless the internal structure of autoconf is changed again. Then it could be difficult to find the locations in the code. ;-) > Wouldn't it be better to supply a standalone .m4 macro specifically for > OS/2 - a sort of ac-os2 or aclocal of some sort which could be maintained > indepedently of autoconf? I don't think so. The basic idea is to get every configure script working out of the box. Therefore it has to be compatible with the "official" relase. If you rerun autoconf on every package there is a risk that a) the OS/2 release is not compatible with the version used by the maintainer and b) the package required a special macro file which is undocumented and which means autoconf can only succeed on the maintainer's computer. Bye, Andreas -- One OS to rule them all, One OS to find them, One OS to bring them all and in the darkness bind them In the Land of Mordor where the Shadows lie. **= Email 5 ==========================** Date: Sun, 8 Dec 2002 10:35:22 +0000 From: John Poltorak Subject: Re: Ncurses build error On Sat, Dec 07, 2002 at 07:44:19PM -0500, Thomas Dickey wrote: > On Sat, Dec 07, 2002 at 10:49:25PM +0000, John Poltorak wrote: > > There is no ncurses.a, but a libncurses.a has been created. > > ncurses/Makefile is in charge of that - but the configure script > is supposed to set the name to ncurses.a > > It's done with this chunk in aclocal.m4: > > dnl --------------------------------------------------------------------------- > dnl Compute the library-prefix for the given host system > dnl $1 = variable to set > AC_DEFUN([CF_LIB_PREFIX], > [ > case $cf_cv_system_name in > OS/2*) LIB_PREFIX='' ;; > os2*) LIB_PREFIX='' ;; > *) LIB_PREFIX='lib' ;; > esac > ifelse($1,,,[$1=$LIB_PREFIX]) > AC_SUBST(LIB_PREFIX) > ])dnl Where is the source which contains this? My version has:- dnl --------------------------------------------------------------------------- dnl Compute the library-prefix for the given host system dnl $1 = variable to set AC_DEFUN([CF_LIB_PREFIX], [ case $cf_cv_system_name in os2) LIB_PREFIX='' ;; *) LIB_PREFIX='lib' ;; esac ifelse($1,,,[$1=$LIB_PREFIX]) AC_SUBST(LIB_PREFIX) ])dnl > When I last built on OS/2, I generated the configure script using autoconf > 2.13 + patches; the only reason the configure script is generated currently > with 2.52 + patches is to work around a defect in glibc which happens to > cancel a bug in autoconf 2.5x (sometime I'll work on a fix to 2.13 to keep > things in sync). I built it using 2.13 + patches. Does it also work with a newer version of autoconf? As far as the original error goes, is there a more recent version of ld.exe which can handle libncurses.a ? > -- > Thomas E. Dickey > http://invisible-island.net > ftp://invisible-island.net -- John **= Email 6 ==========================** Date: Sun, 8 Dec 2002 10:43:19 +0000 From: John Poltorak Subject: Re: Installing autoconf On Sun, Dec 08, 2002 at 01:47:32AM +0100, Andreas Buening wrote: > John Poltorak wrote: > > > > > The "patches" aren't anything special. They set > > > ac_executable_extensions, replace '\' by '/' in your PATH > > > and change the behaviour of AC_CHECK_LIB slightly so that > > > you don't get drowned by linker warnings if you use link386. > > > > It's great that there are so few patches, although I would have thought > > some of the would also apply to Win32 so should be acceptable for > > inclusion. This one for instance (general.m4) :- > > > > - [[\\/$]]* | ?:[[\\/]]* | NONE | '' ) ;; > > + [[\\/$]]* | ?:[[\\/]]* | ?: | NONE | '' ) ;; > > Okay, this one I forgot to mention. > I have no idea what the Win32 guys need. It look as if the > cygwin people had their own autoconf. I doubt whether there is a special cygwin version of autoconf... > > > They don't affect the usability if you've set up your environment > > > correctly. > > > > But how can anyone ensure that their environment has been set up > > correctly in the first place? Can everything necessary be handled by > > config.site? > > Not everything. I'd just like to know how we can ensure that your patches can be integrated into forthcoming releases of autoconf. > > > To answer your question, I don't think they would > > > be accepted as they are. > > > > The problem about long term maintenance remains. > > > > Are the patches applicable to every release of autoconf? > > I think so, unless the internal structure of autoconf is > changed again. Then it could be difficult to find the > locations in the code. ;-) Do you know if the patches can be applied to v2.57? > Bye, > Andreas > > -- > One OS to rule them all, One OS to find them, > One OS to bring them all and in the darkness bind them > In the Land of Mordor where the Shadows lie. -- John **= Email 7 ==========================** Date: Sun, 08 Dec 2002 10:44:27 -0500 From: Ted Sikora Subject: Re: mailman cgi John Poltorak wrote: > > On Sat, Dec 07, 2002 at 07:03:35PM -0500, Ted Sikora wrote: > > Ted Sikora wrote: > > > > > > Ted Sikora wrote: > > > > > > > > Well it seems it's not Mailman but Apache/2. No cgi script runs they all > > > > produce 'Internal Error: spawn child process'. Anyone know how to make > > > > unix cgi scripts work on OS/2? I renamed them tried changing permissions > > > > in unix then copying them. Is there a helper or variable I need to add > > > > for ScriptAlias on OS/2? > > > > > > > > -- > > > > Ted Sikora > > > > tsikora at ntplx.net > > > > > > I think we'll have to make wrapper for cgi. Does one exist? If I rename > > > the file .cmd it runs. I'll see what I can come up with. > > > > > > Here's what I got so far. > > > > > > http://powerusersbbs.net/cgi-bin/test2.cmd > > > > > > > Disregard my last few posts I was getting into left field there for a > > while. All cgi's now work. After playing with all these scripts my > > server locked solid. Now it only runs in VGA 16 colors. If I reinstall > > SDD704 or Gradd083 it locks solid on startup. Anyone get this before? > > Where does it lock up? The blue screen before WPS comes up? > > I thought you should be able to revert to VGA and then reinstall the video > driver. BTW SDD704 is quite old. I recently installed 711. > > Which FP are you on? None just plain Warp Server 4.5 for e-Business. I found the problem. I was using multiple 3Com nic cards. On the initial install I thought I had a bad card and put in another 3Com(905). NFS would lock the machine solid on a large transfer so I thought it was a bug in NFS and ignored it. Then the locking started again. I noticed during downloads on my internal network it would lock ocassionally too. So this morning I replaced one of the cards with a Kingston KNE110 and it's perfect. NFS works perfect. I guess Warp does not like 3Com or glitches with multiple same brand cards. I seen this on other platforms more likely it's poorly written drivers. I got Python to run as CGI with a small batch file(wrapper) I just make a file 'py.cmd' in /cgi-bin and use: at echo off cmd /c "bash -c ./test" Works fine http://powerusersbbs.net/cgi-bin/py.cmd Here it is with echo off http://powerusersbbs.net/cgi-bin/py2.cmd Funny thing I can run any #!/bin/sh script http://powerusersbbs.net/cgi-bin/test-cgi but #!/usr/bin/env python won't. Here it is without the wrapper: http://powerusersbbs.net/cgi-bin/test I tried configuring Apache every which way even going back to Apache 1.3.27 nothing. So I guess it's the wrapper for now. By the way in Unix they work perfectly without any changes to the server configs. Apache logs reports this error when trying to access the file normally. sh does not work either it appends a .exe to any file name so I had to use bash. [Sat Dec 7 21:24:12 2002] [error] [client 209.54.72.110] DosExecPgm( env python.exe d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - [Sat Dec 7 21:24:12 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:24:13 2002] [error] [client 209.54.72.110] DosExecPgm( env python.exe d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - [Sat Dec 7 21:24:13 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:24:14 2002] [error] [client 209.54.72.110] DosExecPgm( env python.exe d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - [Sat Dec 7 21:24:14 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:24:15 2002] [error] [client 209.54.72.110] DosExecPgm( env python.exe d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - [Sat Dec 7 21:24:15 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:24:16 2002] [error] [client 209.54.72.110] DosExecPgm( env python.exe d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - -- Ted Sikora tsikora at ntplx.net **= Email 8 ==========================** Date: Sun, 08 Dec 2002 10:53:56 -0500 From: Ted Sikora Subject: GCI Errors Here's some clearer log errors and why the compiled Mailman scripts fail. OME/MAILMAN/CGI-BIN/ADMIN ) failed, SYS0191: %1 cannot be run in an OS/2 session. - D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\ADMIN [Sat Dec 7 21:19:51 2002] [error] [client 209.54.72.110] Premature end of script headers: D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/ADMIN [Sat Dec 7 21:19:53 2002] [error] [client 209.54.72.110] DosExecPgm(D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO ) failed, SYS0191: %1 cannot be run in an OS/2 session. - D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\LISTINFO [Sat Dec 7 21:19:53 2002] [error] [client 209.54.72.110] Premature end of script headers: D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO [Sat Dec 7 21:19:54 2002] [error] [client 209.54.72.110] attempt to invoke directory as script: d:/unixos2/home/mailman/cgi-bin [Sat Dec 7 21:19:56 2002] [error] [client 209.54.72.110] DosExecPgm(D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO ) failed, SYS0191: %1 cannot be run in an OS/2 session. - D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\LISTINFO [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - \USR\BIN\ENV [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:20:28 2002] [error] [client 209.54.72.110] DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - \USR\BIN\ENV [Sat Dec 7 21:20:28 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:20:29 2002] [error] [client 209.54.72.110] DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, SYS0002: The system cannot find the file specified. - \USR\BIN\ENV [Sat Dec 7 21:20:29 2002] [error] [client 209.54.72.110] Premature end of script headers: d:/apache/cgi-bin/test.py [Sat Dec 7 21:20:31 2002] [error] [client 209.54.72.110] attempt to invoke directory as script: d:/unixos2/home/mailman/cgi-bin -- Ted Sikora tsikora at ntplx.net **= Email 9 ==========================** Date: Sun, 8 Dec 2002 10:58:41 +0000 From: John Poltorak Subject: Ncurses UnixOS/2 package I'd like to put together a UnixOS/2 package for ncurses, but am unsure of the directory structure it should contain. Should $PREFIX be /usr or /usr/local? SlackWare seems to use /usr. Where should includes go? I would prefer them under /usr/include, although SlackWare uses /usr/include/ncurses. What about man pages? Did we ever agree on the preferred location for them? -- John **= Email 10 ==========================** Date: Sun, 8 Dec 2002 11:15:13 -0500 From: Thomas Dickey Subject: Re: Ncurses build error On Sun, Dec 08, 2002 at 10:35:22AM +0000, John Poltorak wrote: > On Sat, Dec 07, 2002 at 07:44:19PM -0500, Thomas Dickey wrote: > > On Sat, Dec 07, 2002 at 10:49:25PM +0000, John Poltorak wrote: > > > > There is no ncurses.a, but a libncurses.a has been created. > > > > ncurses/Makefile is in charge of that - but the configure script > > is supposed to set the name to ncurses.a > > > > It's done with this chunk in aclocal.m4: > > > > dnl --------------------------------------------------------------------------- > > dnl Compute the library-prefix for the given host system > > dnl $1 = variable to set > > AC_DEFUN([CF_LIB_PREFIX], > > [ > > case $cf_cv_system_name in > > OS/2*) LIB_PREFIX='' ;; > > os2*) LIB_PREFIX='' ;; > > *) LIB_PREFIX='lib' ;; > > esac > > ifelse($1,,,[$1=$LIB_PREFIX]) > > AC_SUBST(LIB_PREFIX) > > ])dnl > > Where is the source which contains this? My version has:- I was quoting from current (5.3+patches). I recall changing this chunk a year or so ago, to cover some case where someone made the config.guess script incompatible (as you may see from the two flavors of OS/2). > dnl > --------------------------------------------------------------------------- > dnl Compute the library-prefix for the given host system > dnl $1 = variable to set > AC_DEFUN([CF_LIB_PREFIX], > [ > case $cf_cv_system_name in > os2) LIB_PREFIX='' ;; > *) LIB_PREFIX='lib' ;; > esac > ifelse($1,,,[$1=$LIB_PREFIX]) > AC_SUBST(LIB_PREFIX) > ])dnl > > > When I last built on OS/2, I generated the configure script using autoconf > > 2.13 + patches; the only reason the configure script is generated currently > > with 2.52 + patches is to work around a defect in glibc which happens to > > cancel a bug in autoconf 2.5x (sometime I'll work on a fix to 2.13 to keep > > things in sync). > > I built it using 2.13 + patches. Does it also work with a newer version of > autoconf? 2.52+patches, probably (there's a resync of my patch for 2.52). The ncurses 5.3 configure script is built that way. But OS/2 support in that version is incomplete, so I didn't try that. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 11 ==========================** Date: Sun, 8 Dec 2002 11:19:16 -0500 From: Thomas Dickey Subject: Re: Installing autoconf On Sun, Dec 08, 2002 at 10:43:19AM +0000, John Poltorak wrote: > On Sun, Dec 08, 2002 at 01:47:32AM +0100, Andreas Buening wrote: > > John Poltorak wrote: > > > > > > > The "patches" aren't anything special. They set > > > > ac_executable_extensions, replace '\' by '/' in your PATH > > > > and change the behaviour of AC_CHECK_LIB slightly so that > > > > you don't get drowned by linker warnings if you use link386. > > > > > > It's great that there are so few patches, although I would have thought > > > some of the would also apply to Win32 so should be acceptable for > > > inclusion. This one for instance (general.m4) :- > > > > > > - [[\\/$]]* | ?:[[\\/]]* | NONE | '' ) ;; > > > + [[\\/$]]* | ?:[[\\/]]* | ?: | NONE | '' ) ;; > > > > Okay, this one I forgot to mention. > > I have no idea what the Win32 guys need. It look as if the > > cygwin people had their own autoconf. > > I doubt whether there is a special cygwin version of autoconf... no, there's not. Instead what has happened is that some of the special tweaks needed are embedded into autoconf (the whole issue of how to decide whether ".exe" is needed explicitly is done to make it compatible with cygwin). > > > > They don't affect the usability if you've set up your environment > > > > correctly. > > > > > > But how can anyone ensure that their environment has been set up > > > correctly in the first place? Can everything necessary be handled by > > > config.site? > > > > Not everything. > > I'd just like to know how we can ensure that your patches can be > integrated into forthcoming releases of autoconf. (good luck ;-) -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 12 ==========================** Date: Sun, 8 Dec 2002 11:20:11 -0500 From: Thomas Dickey Subject: Re: Ncurses UnixOS/2 package On Sun, Dec 08, 2002 at 10:58:41AM +0000, John Poltorak wrote: > > I'd like to put together a UnixOS/2 package for ncurses, but am unsure of > the directory structure it should contain. > > Should $PREFIX be /usr or /usr/local? SlackWare seems to use /usr. the former. /usr/local should be used if there's more than one curses implementation. > Where should includes go? I would prefer them under /usr/include, although > SlackWare uses /usr/include/ncurses. the former if ncurses is the "real" curses implementation > What about man pages? Did we ever agree on the preferred location for > them? > > -- > John > > > -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 13 ==========================** Date: Sun, 8 Dec 2002 11:27:53 -0500 From: Thomas Dickey Subject: Re: GCI Errors On Sun, Dec 08, 2002 at 03:45:04PM +0000, John Poltorak wrote: > On Sun, Dec 08, 2002 at 10:53:56AM -0500, Ted Sikora wrote: > > Here's some clearer log errors and why the compiled Mailman scripts > > fail. > > > > [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] > > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, > > I can't get my head round what:- > > /usr/bin/env python > > is supposed to do. > > Can't it just be:- ? > > /usr/bin/python probably it could be - the use of env here is redundant: ENV(1) FSF ENV(1) NAME env - run a program in a modified environment SYNOPSIS env [OPTION]... [-] [NAME=VALUE]... [COMMAND [ARG]...] DESCRIPTION Set each NAME to VALUE in the environment and run COMMAND. -i, --ignore-environment start with an empty environment -u, --unset=NAME remove variable from the environment --help display this help and exit --version output version information and exit A mere - implies -i. If no COMMAND, print the resulting environment. REPORTING BUGS Report bugs to . SEE ALSO The full documentation for env is maintained as a Texinfo manual. If the info and env programs are properly installed at your site, the command info env should give you access to the complete manual. COPYRIGHT Copyright (C) 1999 Free Software Foundation, Inc. This is free software; see the source for copying condi- tions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. GNU sh-utils 2.0 August 1999 1 -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 14 ==========================** Date: Sun, 08 Dec 2002 11:30:41 -0500 From: Ted Sikora Subject: Re: GCI Errors John Poltorak wrote: > > On Sun, Dec 08, 2002 at 10:53:56AM -0500, Ted Sikora wrote: > > Here's some clearer log errors and why the compiled Mailman scripts > > fail. > > > > [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] > > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, > > I can't get my head round what:- > > /usr/bin/env python > > is supposed to do. > > Can't it just be:- ? > > /usr/bin/python > > > -- > > Ted Sikora > > tsikora at ntplx.net > > -- > John No Python uses the env command first the scripts use: #!/usr/bin/env python I'm not sure what running it without it would do. I'm assuming it uses your env variables to set itself up. It's the standard way as far as I can see. -- Ted Sikora tsikora at ntplx.net **= Email 15 ==========================** Date: Sun, 08 Dec 2002 11:35:27 -0500 From: Ted Sikora Subject: Re: GCI Errors Ted Sikora wrote: > > Here's some clearer log errors and why the compiled Mailman scripts > fail. > > OME/MAILMAN/CGI-BIN/ADMIN ) failed, SYS0191: %1 cannot be run in an OS/2 > session. - D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\ADMIN > > [Sat Dec 7 21:19:51 2002] [error] [client 209.54.72.110] Premature end > of script headers: D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/ADMIN > > [Sat Dec 7 21:19:53 2002] [error] [client 209.54.72.110] > DosExecPgm(D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO ) failed, SYS0191: > %1 cannot be run in an OS/2 session. - > > D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\LISTINFO > > [Sat Dec 7 21:19:53 2002] [error] [client 209.54.72.110] Premature end > of script headers: D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO > > [Sat Dec 7 21:19:54 2002] [error] [client 209.54.72.110] attempt to > invoke directory as script: d:/unixos2/home/mailman/cgi-bin > > [Sat Dec 7 21:19:56 2002] [error] [client 209.54.72.110] > DosExecPgm(D:/UNIXOS2/HOME/MAILMAN/CGI-BIN/LISTINFO ) failed, SYS0191: > %1 cannot be run in an OS/2 session. - > D:\UNIXOS2\HOME\MAILMAN\CGI-BIN\LISTINFO > > [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, > SYS0002: The system cannot find the file specified. - \USR\BIN\ENV > > [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] Premature end > of script headers: d:/apache/cgi-bin/test.py > > [Sat Dec 7 21:20:28 2002] [error] [client 209.54.72.110] > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, > SYS0002: The system cannot find the file specified. - \USR\BIN\ENV > > [Sat Dec 7 21:20:28 2002] [error] [client 209.54.72.110] Premature end > of script headers: d:/apache/cgi-bin/test.py > > [Sat Dec 7 21:20:29 2002] [error] [client 209.54.72.110] > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, > SYS0002: The system cannot find the file specified. - \USR\BIN\ENV > > [Sat Dec 7 21:20:29 2002] [error] [client 209.54.72.110] Premature end > of script headers: d:/apache/cgi-bin/test.py > > [Sat Dec 7 21:20:31 2002] [error] [client 209.54.72.110] attempt to > invoke directory as script: d:/unixos2/home/mailman/cgi-bin > I was just thinking what if I use SET OS2_SHELL=D:\bin\bash.exe or sh.exe would it make a difference. Any other consequences system-wide? -- Ted Sikora tsikora at ntplx.net **= Email 16 ==========================** Date: Sun, 8 Dec 2002 12:33:31 +0000 From: John Poltorak Subject: Re: mailman cgi On Sat, Dec 07, 2002 at 07:03:35PM -0500, Ted Sikora wrote: > Ted Sikora wrote: > > > > Ted Sikora wrote: > > > > > > Well it seems it's not Mailman but Apache/2. No cgi script runs they all > > > produce 'Internal Error: spawn child process'. Anyone know how to make > > > unix cgi scripts work on OS/2? I renamed them tried changing permissions > > > in unix then copying them. Is there a helper or variable I need to add > > > for ScriptAlias on OS/2? > > > > > > -- > > > Ted Sikora > > > tsikora at ntplx.net > > > > I think we'll have to make wrapper for cgi. Does one exist? If I rename > > the file .cmd it runs. I'll see what I can come up with. > > > > Here's what I got so far. > > > > http://powerusersbbs.net/cgi-bin/test2.cmd > > > > Disregard my last few posts I was getting into left field there for a > while. All cgi's now work. After playing with all these scripts my > server locked solid. Now it only runs in VGA 16 colors. If I reinstall > SDD704 or Gradd083 it locks solid on startup. Anyone get this before? Where does it lock up? The blue screen before WPS comes up? I thought you should be able to revert to VGA and then reinstall the video driver. BTW SDD704 is quite old. I recently installed 711. Which FP are you on? > -- > Ted Sikora > tsikora at ntplx.net -- John **= Email 17 ==========================** Date: Sun, 8 Dec 2002 14:39:26 +0000 From: John Poltorak Subject: New NCURSES I've just discovered that NCURSES 5.3 has been released... It's available from:- ftp://ftp.gnu.org/pub/gnu/ncurses/ncurses-5.3.tar.gz -- John **= Email 18 ==========================** Date: Sun, 8 Dec 2002 15:01:27 +0000 From: John Poltorak Subject: Re: New NCURSES On Mon, Dec 09, 2002 at 01:54:15AM +1100, IanM wrote: > Hi John > > >I've just discovered that NCURSES 5.3 has been released... > > >It's available from:- > >ftp://ftp.gnu.org/pub/gnu/ncurses/ncurses-5.3.tar.gz > > And ..... > > ftp://hobbes.nmsu.edu/pub/incoming > http://www.os2site.com/sw/new > http://unixos2.com/pub/incoming > > With much thanks to Patrick Ash AFAICT, NCURSES should be buildable on OS/2 straight out of the box. Now that I know there is a new release, I'll see if my build system will handle it. In theory, all I need to do is run:- build ncurses > Cheers > IanM > http://www.os2site.com/ > > > Those who are without sine shall be transformed... -- John **= Email 19 ==========================** Date: Sun, 8 Dec 2002 15:27:18 +0000 From: John Poltorak Subject: Re: New NCURSES On Mon, Dec 09, 2002 at 12:03:52AM +0900, Jun Sawataishi wrote: > At Sun, 8 Dec 2002 14:39:26 +0000, > John Poltorak wrote: > > I've just discovered that NCURSES 5.3 has been released... > Yesterday, I tested this way. > > sh configure && make SHELL=sh > > All library files (FOO.a not libFOO.a), and executables were compliled > out of the box. Good! Yes, this is how it ought to be, unfortunately I got:- ** Configuration summary for NCURSES 5.3 20021012: bin directory: /usr/local/bin lib directory: /usr/local/lib include directory: /usr/local/include/ncurses man directory: /usr/local/man terminfo directory: /usr/local/share/terminfo ** Include-directory is not in a standard location make cd man && make DESTDIR="" all make[1]: Entering directory `/unixos2/workdir/ncurses-5.3/man' make[1]: *** No rule to make target `u:UNIXOS2WORKDIRNCURSES-5.3/man/terminfo.head', needed by `terminfo.5'. Stop. make[1]: Leaving directory `/unixos2/workdir/ncurses-5.3/man' make: *** [all] Error 2 cd man && make DESTDIR="" install make[1]: Entering directory `/unixos2/workdir/ncurses-5.3/man' make[1]: *** No rule to make target `u:UNIXOS2WORKDIRNCURSES-5.3/man/terminfo.head', needed by `terminfo.5'. Stop. make[1]: Leaving directory `/unixos2/workdir/ncurses-5.3/man' make: *** [install] Error 2 end This was without using autoconf first. I'll try adding that step and see if it makes any difference. > > # OS/2 is not a question, it's a solution. > # SAWATAISHI Jun -- John **= Email 20 ==========================** Date: Sun, 8 Dec 2002 15:45:04 +0000 From: John Poltorak Subject: Re: GCI Errors On Sun, Dec 08, 2002 at 10:53:56AM -0500, Ted Sikora wrote: > Here's some clearer log errors and why the compiled Mailman scripts > fail. > > [Sat Dec 7 21:20:25 2002] [error] [client 209.54.72.110] > DosExecPgm(/usr/bin/env python d:/apache/cgi-bin/test.py) failed, I can't get my head round what:- /usr/bin/env python is supposed to do. Can't it just be:- ? /usr/bin/python > -- > Ted Sikora > tsikora at ntplx.net -- John **= Email 21 ==========================** Date: Sun, 08 Dec 2002 15:54:44 -0500 From: Ted Sikora Subject: Mailman: 'module' object has no attribute 'link' Getting closer. If we could somehow spoof the S bit right from the onset I think we would have no problems. patch error: missing header for context diff at line 3 of patch MailList.py error: [dhcp-209-54-72-109|/unixos2/home/mailman/bin]./newlist Enter the name of the list: test Enter the email of the person running the list: ted at powerusersbbs.net Initial test password: Traceback (most recent call last): File "./newlist", line 220, in ? main() File "./newlist", line 169, in main mlist.Create(listname, owner_mail, pw) File "/unixos2/home/mailman/Mailman/MailList.py", line 793, in Create self.__lock.lock() File "/unixos2/home/mailman/Mailman/LockFile.py", line 232, in lock os.link(self.__tmpfname, self.__lockfile) AttributeError: 'module' object has no attribute 'link' -- Ted Sikora tsikora at ntplx.net **= Email 22 ==========================** Date: Sun, 08 Dec 2002 16:01:57 -0500 From: Ted Sikora Subject: Re: Mailman: 'module' object has no attribute 'link' Ted Sikora wrote: > > Getting closer. If we could somehow spoof the S bit right from the onset > I think we would have no problems. > > patch error: > > missing header for context diff at line 3 of patch > > MailList.py error: > > [dhcp-209-54-72-109|/unixos2/home/mailman/bin]./newlist > Enter the name of the list: test > Enter the email of the person running the list: ted at powerusersbbs.net > Initial test password: > Traceback (most recent call last): > File "./newlist", line 220, in ? > main() > File "./newlist", line 169, in main > mlist.Create(listname, owner_mail, pw) > File "/unixos2/home/mailman/Mailman/MailList.py", line 793, in Create > self.__lock.lock() > File "/unixos2/home/mailman/Mailman/LockFile.py", line 232, in lock > os.link(self.__tmpfname, self.__lockfile) > AttributeError: 'module' object has no attribute 'link' > Anyway to have the program equate If system flag is set = Sticky bit is set? Sounds like a good approach to me. -- Ted Sikora tsikora at ntplx.net **= Email 23 ==========================** Date: Sun, 8 Dec 2002 17:05:08 +0000 From: John Poltorak Subject: Re: GCI Errors On Sun, Dec 08, 2002 at 11:35:27AM -0500, Ted Sikora wrote: > I was just thinking what if I use SET OS2_SHELL=D:\bin\bash.exe or > sh.exe would it make a difference. Any other consequences system-wide? I don't have an answer to your question, but since upgrading to the sh.exe contained here:- http://www.math.ohio-state.edu/~ilya/software/os2/pdksh-5.2.14-bin-2.zip I no longer have a need for BASH in shell scripts. > -- > Ted Sikora > tsikora at ntplx.net -- John **= Email 24 ==========================** Date: Sun, 8 Dec 2002 17:14:30 +0000 From: John Poltorak Subject: Make v3.80 Has anyone tried building Make v3.80:- ? ftp://ftp.gnu.org/pub/gnu/make/make-3.80.tar.gz I just noticed that a recent version of Make on Hobbes:- http://hobbes.nmsu.edu/pub/os2/dev/util/make-3_79_2a1-bin.zip was built straight from the CVS source without any patches. It looks as though Andreas Buening may have got all his fixes included in the mainstream version so we may finally stop having to do catch up with GNU Make... -- John **= Email 25 ==========================** Date: Sun, 8 Dec 2002 19:25:25 +0100 (CET) From: Stefan Neis Subject: Re: Installing autoconf On Sun, 8 Dec 2002, John Poltorak wrote: > > > - [[\\/$]]* | ?:[[\\/]]* | NONE | '' ) ;; > > > + [[\\/$]]* | ?:[[\\/]]* | ?: | NONE | '' ) ;; > > > > Okay, this one I forgot to mention. > > I have no idea what the Win32 guys need. It look as if the > > cygwin people had their own autoconf. > > I doubt whether there is a special cygwin version of autoconf... cygwin is living with /cygrive/d/something, it doesn't understand d:/something or d:\something at all, normally. So, when it comes to drive letter support, we have at most "mingw" on our side... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 26 ==========================** Date: Sun, 8 Dec 2002 19:33:28 +0000 From: John Poltorak Subject: Autoconf build error - multiple target patterns When trying to install Autoconf (2.53), I get the error:- *** multiple target patterns Anyone know what this means and how to get round it? -- John