From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Sat, 18 Jan 2003 04:48:36 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 17 ************************************************** Friday 17 January 2003 Number 17 ************************************************** Subjects for today 1 Re: How to identify OS/2 from shell script : Andreas Buening 2 Re: Autoconf & Make 3.76.1 : Thomas Dickey 3 Re: BYACC : Thomas Dickey 4 Re: KSH : Thomas Dickey 5 BYACC : John Poltorak 6 Re: How to identify OS/2 from shell script : John Poltorak 7 Autoconf & Make 3.76.1 : John Poltorak 8 ZLIB install : John Poltorak 9 Variable argument prototypes : Jeff Robinson 10 Re: How to identify OS/2 from shell script : Yuri Dario" 11 KSH : John Poltorak 12 Re: Variable argument prototypes : Jeff Robinson 13 Re: How to identify OS/2 from shell script : Adrian Gschwend" 14 Re: Variable argument prototypes : Jeff Robinson 15 Re: DIALOG and drive letters : Thomas Dickey 16 Re: Variable argument prototypes : Thomas Dickey 17 Re: KSH : Thomas Dickey 18 Re: How to identify OS/2 from shell script : Adrian Gschwend" 19 Re: Variable argument prototypes : Thomas E. Dickey" 20 Re: KSH : John Poltorak 21 GZIP patch : John Poltorak 22 Re: Testing for LINK386 on path : John Poltorak 23 Re: Testing for LINK386 on path : John Poltorak 24 Re: How to identify OS/2 from shell script : Yuri Dario" 25 Re: How to identify OS/2 from shell script : Adrian Gschwend" 26 Re: Testing for LINK386 on path : Yuri Dario" 27 DIALOG and drive letters : John Poltorak 28 Re: isc DHCP client : Michael Zolk 29 Re: -Zomf (was: Testing for LINK386 on path) : Illya Vaes 30 Re: DIALOG and drive letters : Thomas Dickey 31 Re: -Zomf (was: Testing for LINK386 on path) : Stefan Neis 32 Re: Variable argument prototypes : Stefan Neis 33 Re: ZLIB install : Stefan Neis 34 Re: DIALOG and drive letters : John Poltorak 35 Re: ZLIB install : John Poltorak **= Email 1 ==========================** Date: Sat, 18 Jan 2003 00:30:53 +0100 From: Andreas Buening Subject: Re: How to identify OS/2 from shell script Adrian Gschwend wrote: > > On Thu, 16 Jan 2003 21:01:22 +0000, John Poltorak wrote: > > >One suggestion is uname. Anything else? > > > >How about using something like > > > >if `ver`="Operating System/2" ? > > why not report UnixOS2 in uname( )? Because a) UnixOS2 is not an OS, and b) if everybody has his preferred OS string (OS/2, eCS, UnixOS2, UnixOS/2, OS2) we had complete chaos. > Then it doesn't matter if it's eCS or OS/2 and it does not give > headache with / in the name. 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 2 ==========================** Date: Sat, 18 Jan 2003 07:18:32 -0500 From: Thomas Dickey Subject: Re: Autoconf & Make 3.76.1 On Sat, Jan 18, 2003 at 09:36:04AM +0000, John Poltorak wrote: > > I realise the README for Autoconf says that Make v3.79.1+ is required to > build Autoconf, but I'm trying to understand why... > > With fixed versions of Make 3.76.1, building Autoconf goes quite well up > to a certain point and part of the app actually gets installed, but then > Make goes into some strange loop. ... > Is there any way to trace what is happening and why? "make -d" is supposed to show what/why "make" does something. But the trace can be hard to interpret. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 3 ==========================** Date: Sat, 18 Jan 2003 07:18:53 -0500 From: Thomas Dickey Subject: Re: BYACC On Sat, Jan 18, 2003 at 08:47:07AM +0000, John Poltorak wrote: > I've managed to build BYACC straight from the source code, but am unsure > about the destination of the output from build... > > It looks as though the two files which need to be installed are the > binary - yacc.exe and the man page - yacc.1. > > Can they be installed using 'make install'? > > I guess not, because the install target looks as though it simply moves > $PROGRAM which is yacc rather than yacc.exe to $DEST which is defined in > the Makefile as '.' > > > If I can't use the Makefile to install it, I propose to copy yacc.exe to > /usr/bin and yacc.1 to /usr/share/man/man1. Does this sound correct? yes. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 4 ==========================** Date: Sat, 18 Jan 2003 07:46:05 -0500 From: Thomas Dickey Subject: Re: KSH On Sat, Jan 18, 2003 at 12:28:31PM +0000, John Poltorak wrote: > > The Makefile for PDKSH has SHELL set to ksh. > > Am I likely to hit any problems if I simply copy sh.exe as ksh.exe simply > for the purposes of building PDKSH? perhaps. what does the makefile look like? -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 5 ==========================** Date: Sat, 18 Jan 2003 08:47:07 +0000 From: John Poltorak Subject: BYACC I've managed to build BYACC straight from the source code, but am unsure about the destination of the output from build... It looks as though the two files which need to be installed are the binary - yacc.exe and the man page - yacc.1. Can they be installed using 'make install'? I guess not, because the install target looks as though it simply moves $PROGRAM which is yacc rather than yacc.exe to $DEST which is defined in the Makefile as '.' If I can't use the Makefile to install it, I propose to copy yacc.exe to /usr/bin and yacc.1 to /usr/share/man/man1. Does this sound correct? -- John **= Email 6 ==========================** Date: Sat, 18 Jan 2003 09:09:28 +0000 From: John Poltorak Subject: Re: How to identify OS/2 from shell script On Sat, Jan 18, 2003 at 12:30:53AM +0100, Andreas Buening wrote: > Adrian Gschwend wrote: > > > > On Thu, 16 Jan 2003 21:01:22 +0000, John Poltorak wrote: > > > > >One suggestion is uname. Anything else? > > > > > >How about using something like > > > > > >if `ver`="Operating System/2" ? > > > > why not report UnixOS2 in uname( )? > > Because a) UnixOS2 is not an OS, and b) if everybody has > his preferred OS string (OS/2, eCS, UnixOS2, UnixOS/2, OS2) > we had complete chaos. Yes, I agree. Does anyone knows how eCS reports itself? If it isn't OS/2 then this is likely to cause a lot of problems. Also, if OSFree ever gets completed someone needs to think very carefully about the output it will produce when 'uname -s' is run. If it isn't OS/2 then there is a huge amount of work that would be required for it to build any Unix apps. > 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: Sat, 18 Jan 2003 09:36:04 +0000 From: John Poltorak Subject: Autoconf & Make 3.76.1 I realise the README for Autoconf says that Make v3.79.1+ is required to build Autoconf, but I'm trying to understand why... With fixed versions of Make 3.76.1, building Autoconf goes quite well up to a certain point and part of the app actually gets installed, but then Make goes into some strange loop. Can anyone explain what the problem is? I've tried to follow what happens but get hopelessly tied up in knots. The build sequence goes fine until:- Making install in bin Making install in tests Making install in . and then it goes into a continuous loop:- Making install in bin Making install in tests Making install in . Making install in bin Making install in tests Making install in . Making install in bin Making install in tests Making install in . Making install in bin Is there any way to trace what is happening and why? -- John **= Email 8 ==========================** Date: Sat, 18 Jan 2003 09:58:51 +0000 From: John Poltorak Subject: ZLIB install Is there any way to override $prefix when installing ZLIB? I guess the Makefile sets the value to /usr/local irrespective of any existing environment variable... Is there any way for it to check the environment first and only set a value if nothing is already set? -- John **= Email 9 ==========================** Date: Sat, 18 Jan 2003 10:19:23 -0600 From: Jeff Robinson Subject: Variable argument prototypes Facing the risk of overlooking the obvious (and having it pointed out), I'm wondering if there are any known differences in how C prototypes with variable numbers of arguments is different on OS/2's gcc (3.0.3) than it is on other platforms? I'm currently recompiling gbash with both JUN SAWATAISHI patch and a "new" tmpdir fix. (Jun's version has built-ins that respect/maintain OS/2 drive letters). Having successfully patched the code and gotten compilation to go part of the way through I'm running into the following error message: error.c: In function `programming_error': error.c:113: number of arguments doesn't match prototype error.h:33: prototype declaration Okay, simple right... not enough arguments... except the prototype is supposed to take a variable number of arguments... The prototype: /* Report a programmer's error, and abort. Pass REASON, and ARG1 ... ARG5. */ extern void programming_error __P((const char *, ...)) __attribute__((__format__ (printf, 1, 2))); The method: void programming_error (reason, arg1, arg2, arg3, arg4, arg5) const char *reason; { char *h; ... (and so on) So is the answer right infront of my nose and I can't see it, or is something afoot? If nothing else, at least I can appreciate the irony that this is occurring in the file named error.c Thanks in advance for any advice, Jeff -- ---------------- Whatza JamochaMUD? http://jamochamud.anecho.mb.ca Or other stuff: http://www.anecho.mb.ca/~jeffnik ----------------------------------------------------------- **= Email 10 ==========================** Date: Sat, 18 Jan 2003 10:44:30 +0100 (CET) From: "Yuri Dario" Subject: Re: How to identify OS/2 from shell script Hi John, >Does anyone knows how eCS reports itself? If it isn't OS/2 then this is >likely to cause a lot of problems. uname -s OS/2 the same as IBM OS/2. Bye, Yuri Dario /* * member of TeamOS/2 - Italy * http://www.quasarbbs.net/yuri * http://www.teamos2.it * http://www.opera.com/os2/ */ **= Email 11 ==========================** Date: Sat, 18 Jan 2003 12:28:31 +0000 From: John Poltorak Subject: KSH The Makefile for PDKSH has SHELL set to ksh. Am I likely to hit any problems if I simply copy sh.exe as ksh.exe simply for the purposes of building PDKSH? -- John **= Email 12 ==========================** Date: Sat, 18 Jan 2003 12:30:49 -0600 From: Jeff Robinson Subject: Re: Variable argument prototypes Stefan Neis wrote: > On Sat, 18 Jan 2003, Jeff Robinson wrote: > > >>The method: >>void >>programming_error (reason, arg1, arg2, arg3, arg4, arg5) >>const char *reason; >>{ >> char *h; >>... (and so on) > > > Huh? What kind of function header syntax is that supposed to be? > Is that supposed to indicate the arguments are const char *reason > and 5 more of unknown type? Whereever this comes from, it sure > doesn't look like valid C. :-( > > Regards, > Stefan Yeah, it looked pretty strange to me, but apparently it is a construct specific to gcc...! ( http://www.unixwiz.net/techtips/gnu-c-attributes.html ). Originally (post patch) the code read: programming_error (reason, arg1, arg2, arg3, arg4, arg5) char *reason; { Before changing *reason to a const char I would get: error.c: In function `programming_error': error.c:113: argument `reason' doesn't match prototype error.h:33: prototype declaration error.c:113: number of arguments doesn't match prototype error.h:33: prototype declaration So the change implies to me (by the errors being reduced) that the compiler does indeed see const char *reason as part of the method. I dunno, either...! This code comes from GNU bash, so I felt pretty confident going in that even if it wasn't standard C, that gcc would be able to handle it. Of course, I can't find any other __attribute__ example where a variable is decalred before the bracket... Jeff -- ---------------- Whatza JamochaMUD? http://jamochamud.anecho.mb.ca Or other stuff: http://www.anecho.mb.ca/~jeffnik ----------------------------------------------------------- **= Email 13 ==========================** Date: Sat, 18 Jan 2003 12:53:29 +0100 (CET) From: "Adrian Gschwend" Subject: Re: How to identify OS/2 from shell script On Sat, 18 Jan 2003 10:44:30 +0100 (CET), Yuri Dario wrote: >>Does anyone knows how eCS reports itself? If it isn't OS/2 then this is >>likely to cause a lot of problems. > >uname -s >OS/2 where is uname implemented, in IBM libc? cu Adrian -- Adrian Gschwend at netlabs.org ktk [a t] netlabs.org ------- Free Software for OS/2 and eCS http://www.netlabs.org **= Email 14 ==========================** Date: Sat, 18 Jan 2003 12:53:59 -0600 From: Jeff Robinson Subject: Re: Variable argument prototypes Thomas E. Dickey wrote: > On Sat, 18 Jan 2003, Jeff Robinson wrote: > > >>>Huh? What kind of function header syntax is that supposed to be? >>>Is that supposed to indicate the arguments are const char *reason >>>and 5 more of unknown type? Whereever this comes from, it sure >>>doesn't look like valid C. :-( >>> >>> Regards, >>> Stefan >> >>Yeah, it looked pretty strange to me, but apparently it is a construct >>specific to gcc...! ( >>http://www.unixwiz.net/techtips/gnu-c-attributes.html ). > > > It's not specific to gcc - though as a matter of fact, people writing > GNU software were slow to adapt to ANSI C. It was around 1995 or 1996 > when they stopped advising people to write code this way. > > Oopse... you are quite correct. I misread the quote on the link that I supplied. Jeff -- ---------------- Whatza JamochaMUD? http://jamochamud.anecho.mb.ca Or other stuff: http://www.anecho.mb.ca/~jeffnik ----------------------------------------------------------- **= Email 15 ==========================** Date: Sat, 18 Jan 2003 13:22:04 -0500 From: Thomas Dickey Subject: Re: DIALOG and drive letters On Sat, Jan 18, 2003 at 04:09:15PM +0000, John Poltorak wrote: > > > Is it possible to build DIALOG when the source and include dir are under > different drive letters. It works fine when they are both on the same > drive. The configure script doesn't look for that, but you should be able to persuade it by setting $CPPFLAGS to use -I options for the appropriate drive. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 16 ==========================** Date: Sat, 18 Jan 2003 13:26:27 -0500 From: Thomas Dickey Subject: Re: Variable argument prototypes On Sat, Jan 18, 2003 at 06:47:19PM +0100, Stefan Neis wrote: > On Sat, 18 Jan 2003, Jeff Robinson wrote: > > > The method: > > void > > programming_error (reason, arg1, arg2, arg3, arg4, arg5) > > const char *reason; > > { > > char *h; > > ... (and so on) > > Huh? What kind of function header syntax is that supposed to be? it's K&R code (pre-standard). For at least a few years after the C standard, people were still writing code like that, because they didn't want to rely on (or even ) being available. > Is that supposed to indicate the arguments are const char *reason > and 5 more of unknown type? Whereever this comes from, it sure > doesn't look like valid C. :-( > > Regards, > Stefan > -- > Micro$oft is not an answer. It is a question. The answer is 'no'. > -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 17 ==========================** Date: Sat, 18 Jan 2003 13:29:24 -0500 From: Thomas Dickey Subject: Re: KSH On Sat, Jan 18, 2003 at 01:56:58PM +0000, John Poltorak wrote: > On Sat, Jan 18, 2003 at 07:46:05AM -0500, Thomas Dickey wrote: > > On Sat, Jan 18, 2003 at 12:28:31PM +0000, John Poltorak wrote: > > > > > > The Makefile for PDKSH has SHELL set to ksh. > > > > > > Am I likely to hit any problems if I simply copy sh.exe as ksh.exe simply > > > for the purposes of building PDKSH? > > > > perhaps. what does the makefile look like? > > Here is part of it:- I don't see any obvious problems in this fragment. Some things to look for: embedded shell script commands, especially those that are complicated and may attempt to cd to a subdirectory. "make" may be confused if timestamps are rounded up to the next second. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 18 ==========================** Date: Sat, 18 Jan 2003 13:34:56 +0100 (CET) From: "Adrian Gschwend" Subject: Re: How to identify OS/2 from shell script On Sat, 18 Jan 2003 00:30:53 +0100, Andreas Buening wrote: >Because a) UnixOS2 is not an OS, and b) if everybody has >his preferred OS string (OS/2, eCS, UnixOS2, UnixOS/2, OS2) >we had complete chaos. but it would show that there is a common base used (UnixOS2) so we can assume if this is returned a lot of other things will be available. Anyway, uname( ) is Possix IIRC (no Stevens near me) so if we implement it *we* have to decide what we want to return. And I think reporting different things for OS/2 and eCS is really not a good idea. Also Yuri told it does report OS/2 but I didn't yet got the point which uname reports that. cu Adrian -- Adrian Gschwend at netlabs.org ktk [a t] netlabs.org ------- Free Software for OS/2 and eCS http://www.netlabs.org **= Email 19 ==========================** Date: Sat, 18 Jan 2003 13:40:04 -0500 (EST) From: "Thomas E. Dickey" Subject: Re: Variable argument prototypes On Sat, 18 Jan 2003, Jeff Robinson wrote: > > Huh? What kind of function header syntax is that supposed to be? > > Is that supposed to indicate the arguments are const char *reason > > and 5 more of unknown type? Whereever this comes from, it sure > > doesn't look like valid C. :-( > > > > Regards, > > Stefan > > Yeah, it looked pretty strange to me, but apparently it is a construct > specific to gcc...! ( > http://www.unixwiz.net/techtips/gnu-c-attributes.html ). It's not specific to gcc - though as a matter of fact, people writing GNU software were slow to adapt to ANSI C. It was around 1995 or 1996 when they stopped advising people to write code this way. > Originally (post patch) the code read: > programming_error (reason, arg1, arg2, arg3, arg4, arg5) > char *reason; > { -- T.E.Dickey http://invisible-island.net ftp://invisible-island.net **= Email 20 ==========================** Date: Sat, 18 Jan 2003 13:56:58 +0000 From: John Poltorak Subject: Re: KSH On Sat, Jan 18, 2003 at 07:46:05AM -0500, Thomas Dickey wrote: > On Sat, Jan 18, 2003 at 12:28:31PM +0000, John Poltorak wrote: > > > > The Makefile for PDKSH has SHELL set to ksh. > > > > Am I likely to hit any problems if I simply copy sh.exe as ksh.exe simply > > for the purposes of building PDKSH? > > perhaps. what does the makefile look like? Here is part of it:- # # OS2_Makefile # srcdir = . VPATH = . CC = gcc CPP = cpp INSTALL = /bin/install -c INSTALL_PROGRAM = $(INSTALL) INSTALL_DATA = $(INSTALL) -m 644 DEFS = -DHAVE_CONFIG_H LIBS = -los2 CPPFLAGS = CFLAGS = -O -DOS2 LDSTATIC = LDFLAGS = -O -s $(LDSTATIC) SHELL_PROG = sh prefix = c:/usr exec_prefix = c:/usr transform= at program_transform_name at bindir = $(exec_prefix)/bin mandir = $(prefix)/man/man$(manext) manext = 1 # Suffix for executables: nothing for unix, .exe for os/2. exe_suffix=.exe SHELL = ksh SRCS = alloc.c c_ksh.c c_sh.c c_test.c c_ulimit.c edit.c emacs.c \ eval.c exec.c expr.c history.c io.c jobs.c lex.c mail.c \ main.c misc.c missing.c path.c shf.c sigact.c syn.c table.c trap.c \ tree.c tty.c var.c version.c vi.c OBJS = os2.o alloc.o c_ksh.o c_sh.o c_test.o c_ulimit.o edit.o emacs.o \ eval.o exec.o expr.o history.o io.o jobs.o lex.o mail.o \ main.o misc.o missing.o path.o shf.o sigact.o syn.o table.o trap.o \ tree.o tty.o var.o version.o vi.o HDRS = c_test.h edit.h expand.h ksh_dir.h ksh_limval.h ksh_stat.h ksh_time.h \ ksh_times.h ksh_wait.h lex.h proto.h sh.h shf.h sigact.h \ table.h tree.h tty.h DISTFILES = $(SRCS) $(HDRS) ksh.Man Makefile.in configure.in \ config.h.in conf-end.h acconfig.h aclocal.m4 emacs-gen.sh \ mkinstalldirs install-sh new-version.sh siglist.in siglist.sh mkman \ check-fd.c check-pgrp.c check-sigs.c \ README NEWS CONTRIBUTORS LEGAL PROJECTS INSTALL NOTES BUG-REPORTS \ IAFA-PACKAGE ChangeLog ChangeLog.0 configure stamp-h.in # ETCFILES also disted, but handled differently ETCFILES = etc/ksh.kshrc etc/profile etc/sys_config.sh # MISCFILES also disted, but handled differently MISCFILES = misc/ChangeLog.sjg misc/Changes.jrm misc/Changes.mlj \ misc/Changes.pc misc/README.sjg misc/ReadMe.eg misc/ReadMe.emacs \ misc/ReadMe.jrm misc/Bugs # TESTFILES also disted, but handled differently TESTFILES = tests/README tests/th tests/th-sh tests/alias.t tests/arith.t \ tests/bksl-nl.t tests/brkcont.t tests/cdhist.t tests/eglob.t \ tests/glob.t tests/heredoc.t tests/history.t tests/ifs.t \ tests/integer.t tests/lineno.t tests/read.t tests/regress.t \ tests/syntax.t tests/unclass1.t tests/unclass2.t \ tests/version.t # OS2FILES also disted, but handled differently OS2FILES = os2/Makefile os2/config.h os2/config.status os2/configure.cmd \ os2/emacs.out os2/kshrc.ksh os2/make.sed os2/os2.c os2/os2siglist.out \ os2/README.os2 os2/NEWS.os2 os2/os2bugs os2/th.cmd os2/config.cache all: $(SHELL_PROG)$(exe_suffix) $(SHELL_PROG).1 # This shouldn't be first - some makes don't know about PRECIOUS and assume it # is the default target. .PRECIOUS: configure config.h.in Makefile config.status .c.o: $(CC) -c $(CPPFLAGS) $(DEFS) -I. -I$(srcdir) $(CFLAGS) $< > -- > Thomas E. Dickey > http://invisible-island.net > ftp://invisible-island.net -- John **= Email 21 ==========================** Date: Sat, 18 Jan 2003 14:21:21 +0000 From: John Poltorak Subject: GZIP patch Has anyone come across an OS/2 patch for GZIP v1.3.3 ? I'm not sure if one exists, but I have been unable to build the program straight out of the box. -- John **= Email 22 ==========================** Date: Sat, 18 Jan 2003 14:26:00 +0000 From: John Poltorak Subject: Re: Testing for LINK386 on path On Fri, Jan 17, 2003 at 03:33:02PM +0100, Stefan Neis wrote: > On Fri, 17 Jan 2003, John Poltorak wrote: > > > How do I know if I want to build OMF based stuff or a.out? > > Well, if you're using -Zomf, you're building OMF based stuff. > _The_ argument in favour of OMF is that it usually results in > smaller executables, that AFAIK, that's it. I guess we ought to have a standard for UnixOS/2 apps... Should CFLAGS include -Zomf in all builds if possible ? > Regards, > Stefan > -- > Micro$oft is not an answer. It is a question. The answer is 'no'. > -- John **= Email 23 ==========================** Date: Sat, 18 Jan 2003 14:57:27 +0000 From: John Poltorak Subject: Re: Testing for LINK386 on path On Sat, Jan 18, 2003 at 03:43:17PM +0100, Yuri Dario wrote: > Hi, > > >I guess we ought to have a standard for UnixOS/2 apps... Should CFLAGS > >include -Zomf in all builds if possible ? > > with -Zomf, executables doesn't use the emx runtime, so many functions are not available > (e.g. fork()). Do you mean they get built as standalone apps? I would have thought that this would mean they would be larger apps, but Stefan said that -Zomf would create smaller apps... > > Bye, > > Yuri Dario > > /* > * member of TeamOS/2 - Italy > * http://www.quasarbbs.net/yuri > * http://www.teamos2.it > * http://www.opera.com/os2/ > */ -- John **= Email 24 ==========================** Date: Sat, 18 Jan 2003 15:13:16 +0100 (CET) From: "Yuri Dario" Subject: Re: How to identify OS/2 from shell script Hi Adrian, >>uname -s >>OS/2 > >where is uname implemented, in IBM libc? no, it is part of gnu shell utilities. Mine is from sh-utils 2.12 Bye, Yuri Dario /* * member of TeamOS/2 - Italy * http://www.quasarbbs.net/yuri * http://www.teamos2.it * http://www.opera.com/os2/ */ **= Email 25 ==========================** Date: Sat, 18 Jan 2003 15:36:33 +0100 (CET) From: "Adrian Gschwend" Subject: Re: How to identify OS/2 from shell script On Sat, 18 Jan 2003 15:13:16 +0100 (CET), Yuri Dario wrote: >>where is uname implemented, in IBM libc? > >no, it is part of gnu shell utilities. Mine is from sh-utils 2.12 that's the binary but there is also an according syscall called uname. I wonder where this one is implemented. cu Adrian -- Adrian Gschwend at netlabs.org ktk [a t] netlabs.org ------- Free Software for OS/2 and eCS http://www.netlabs.org **= Email 26 ==========================** Date: Sat, 18 Jan 2003 15:43:17 +0100 (CET) From: "Yuri Dario" Subject: Re: Testing for LINK386 on path Hi, >I guess we ought to have a standard for UnixOS/2 apps... Should CFLAGS >include -Zomf in all builds if possible ? with -Zomf, executables doesn't use the emx runtime, so many functions are not available (e.g. fork()). Bye, Yuri Dario /* * member of TeamOS/2 - Italy * http://www.quasarbbs.net/yuri * http://www.teamos2.it * http://www.opera.com/os2/ */ **= Email 27 ==========================** Date: Sat, 18 Jan 2003 16:09:15 +0000 From: John Poltorak Subject: DIALOG and drive letters Is it possible to build DIALOG when the source and include dir are under different drive letters. It works fine when they are both on the same drive. -- John **= Email 28 ==========================** Date: Sat, 18 Jan 2003 16:44:07 +0100 From: Michael Zolk Subject: Re: isc DHCP client On Mon, Jan 13, 2003 at 11:55:13AM +0000, John Poltorak wrote: > > but I run into missing library functions, I've looked into the libc sources > > and found out that the missing functions are really simple, > > compiled them created my own lib and it 'worked' as in I got the stuff > > compiled, I'm not ready for testing yet, but its a start.. > > but my question now, how is adding of missing libc functions coordinated ? > > because like this we end up having private implementations of missing > > functions which is not a good idea I think. > > This is a major problem when we have to many devlopers working in > isolation. > > My personal view is that we should try to incorporate missing libc > functions into Posix/2... Of course that puts the burden of maintaining > these functions on Stefan, and it may well be a time consuming task. Isn't this was Andreas Buening is working on with libunixos2? M. **= Email 29 ==========================** Date: Sat, 18 Jan 2003 18:02:55 +0100 From: Illya Vaes Subject: Re: -Zomf (was: Testing for LINK386 on path) ** Reply to note from "Yuri Dario" Sat, 18 Jan 2003 15:43:17 +0100 (CET) >>I guess we ought to have a standard for UnixOS/2 apps... Should CFLAGS >>include -Zomf in all builds if possible ? >with -Zomf, executables doesn't use the emx runtime, so many functions are >not available (e.g. fork()). That's not true, it only switches to OMF format output and the usual extensions (.obj and .lib instead of .o and .a). You're confused with -Zso (StandalOne) (and -Zsys). -- Illya Vaes (illya at vaeshiep.demon.nl) "Do...or do not, there is no 'try'" - Yoda **= Email 30 ==========================** Date: Sat, 18 Jan 2003 18:41:18 -0500 From: Thomas Dickey Subject: Re: DIALOG and drive letters On Sat, Jan 18, 2003 at 10:49:28PM +0000, John Poltorak wrote: > On Sat, Jan 18, 2003 at 01:22:04PM -0500, Thomas Dickey wrote: > > On Sat, Jan 18, 2003 at 04:09:15PM +0000, John Poltorak wrote: > > > > > > > > > Is it possible to build DIALOG when the source and include dir are under > > > different drive letters. It works fine when they are both on the same > > > drive. > > > > The configure script doesn't look for that, but you should be able to > > persuade it by setting $CPPFLAGS to use -I options for the appropriate > > drive. > > Can you give me an example? > > Are CPPFLAGS documented anywhere? more than one place. Here's a quote from GNU standards (which in turn influences the autoconf stuff): Each program-name variable should come with an options variable that is used to supply options to the program. Append `FLAGS' to the program-name variable name to get the options variable name--for example, `BISONFLAGS'. (The names `CFLAGS' for the C compiler, `YFLAGS' for yacc, and `LFLAGS' for lex, are exceptions to this rule, but we keep them because they are standard.) Use `CPPFLAGS' in any compilation command that runs the preprocessor, and use `LDFLAGS' in any compilation command that does linking as well as in any direct use of `ld'. The -I and -D options are preprocessor options, btw. -- Thomas E. Dickey http://invisible-island.net ftp://invisible-island.net **= Email 31 ==========================** Date: Sat, 18 Jan 2003 18:42:22 +0100 (CET) From: Stefan Neis Subject: Re: -Zomf (was: Testing for LINK386 on path) On Sat, 18 Jan 2003, Illya Vaes wrote: > ** Reply to note from "Yuri Dario" Sat, 18 Jan 2003 > 15:43:17 +0100 (CET) > >>I guess we ought to have a standard for UnixOS/2 apps... Should CFLAGS > >>include -Zomf in all builds if possible ? > >with -Zomf, executables doesn't use the emx runtime, so many functions are > >not available (e.g. fork()). > > That's not true, it only switches to OMF format output and the usual > extensions (.obj and .lib instead of .o and .a). > You're confused with -Zso (StandalOne) (and -Zsys). While it's still using the emx runtime, there are some limitations which apply as soon as you're using -Zomf (specifically, such programs can't use fork(), according to the documentation...). Regards, Stefan **= Email 32 ==========================** Date: Sat, 18 Jan 2003 18:47:19 +0100 (CET) From: Stefan Neis Subject: Re: Variable argument prototypes On Sat, 18 Jan 2003, Jeff Robinson wrote: > The method: > void > programming_error (reason, arg1, arg2, arg3, arg4, arg5) > const char *reason; > { > char *h; > ... (and so on) Huh? What kind of function header syntax is that supposed to be? Is that supposed to indicate the arguments are const char *reason and 5 more of unknown type? Whereever this comes from, it sure doesn't look like valid C. :-( Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 33 ==========================** Date: Sat, 18 Jan 2003 18:49:52 +0100 (CET) From: Stefan Neis Subject: Re: ZLIB install On Sat, 18 Jan 2003, John Poltorak wrote: > > Is there any way to override $prefix when installing ZLIB? The "usual way" to override it, would be to do it at configure time, via e.g. ./configure --prefix=/usr Once you have a Makefile, it's to late to change your mind, normally. Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 34 ==========================** Date: Sat, 18 Jan 2003 22:49:28 +0000 From: John Poltorak Subject: Re: DIALOG and drive letters On Sat, Jan 18, 2003 at 01:22:04PM -0500, Thomas Dickey wrote: > On Sat, Jan 18, 2003 at 04:09:15PM +0000, John Poltorak wrote: > > > > > > Is it possible to build DIALOG when the source and include dir are under > > different drive letters. It works fine when they are both on the same > > drive. > > The configure script doesn't look for that, but you should be able to > persuade it by setting $CPPFLAGS to use -I options for the appropriate > drive. Can you give me an example? Are CPPFLAGS documented anywhere? > -- > Thomas E. Dickey > http://invisible-island.net > ftp://invisible-island.net -- John **= Email 35 ==========================** Date: Sat, 18 Jan 2003 22:58:10 +0000 From: John Poltorak Subject: Re: ZLIB install On Sat, Jan 18, 2003 at 06:49:52PM +0100, Stefan Neis wrote: > On Sat, 18 Jan 2003, John Poltorak wrote: > > > > > Is there any way to override $prefix when installing ZLIB? > > The "usual way" to override it, would be to do it at configure time, > via e.g. ./configure --prefix=/usr > Once you have a Makefile, it's to late to change your mind, normally. I don't think ZLIB uses a configure script built by autoconf. It isn't usable by OS/2, anyway. For OS/2, a Makefile.os2 is supplied. Maybe getting the ZLIB developers to use autoconf would make it easier for us in the long run... > Regards, > Stefan > -- > Micro$oft is not an answer. It is a question. The answer is 'no'. > -- John