From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Sun, 16 Nov 2003 14:15:41 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 242 ************************************************** Saturday 15 November 2003 Number 242 ************************************************** Subjects for today 1 Re: expat : Stefan Neis 2 Re: Makeinfo and sed.texi : Stefan Neis 3 Re: Use of depcomp : Dave and Natalie" 4 Re: Automake problem : Dave and Natalie" 5 std::min? : Dave and Natalie" 6 Re: magic : John Poltorak 7 Re: Use of depcomp : John Poltorak 8 Re: Use of depcomp : John Poltorak 9 Re: Makeinfo and sed.texi : John Poltorak 10 Re: Automake problem : John Poltorak 11 Any shell experts? : John Poltorak 12 Re: Use of depcomp : Dave and Natalie" 13 Re: Any shell experts? : John Poltorak 14 Texinfo at sc attribute : John Poltorak 15 Re: Any shell experts? : Andreas Buening 16 Re: Use of depcomp : Andreas Buening 17 Re: Any shell experts? : John Poltorak 18 Re: Any shell experts? : John Poltorak 19 Building GNU Text Utils : John Poltorak 20 Re: Any shell experts? : Andreas Buening 21 Re: Use of depcomp : John Poltorak 22 Re: Any shell experts? : John Poltorak **= Email 1 ==========================** Date: Sun, 16 Nov 2003 01:51:43 +0100 (CET) From: Stefan Neis Subject: Re: expat On Sat, 15 Nov 2003, John Poltorak wrote: > There is a new port of EXPAT (v1.95.6) on Hobbes. > > Can anyone make any sense of the included README-expat which shows how it > can be built from source? Actually, it's being used by wxWindows and the original expat package does apparently build just fine after regenerating the configure script with autoconf-2.57 (IIRC, the one in the expat distribution is genrated by 2.53 and caused some problems to me ...). Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 2 ==========================** Date: Sun, 16 Nov 2003 01:59:04 +0100 (CET) From: Stefan Neis Subject: Re: Makeinfo and sed.texi On Sat, 15 Nov 2003, John Poltorak wrote: > makeinfo -I . \ > -o sed.info `test -f 'sed.texi' || echo './'`sed.texi > ./version.texi:20: Misplaced {. > ./version.texi:20: Misplaced }. > ./version.texi:36: Unknown command `(#)PD'. > sed.texi:99: warning: at sc argument all uppercase, thus no effect. (snipp) > sed.texi:243: warning: at sc argument all uppercase, thus no effect. > makeinfo: Removing output file `sed.info' due to errors; use --force to preserve. > make[1]: *** [sed.info] Error 2 (snipp) > > Is this a problem with MAKEINFO? That's one possible explanation - you might be using a makeinfo that's either too old or too new. Or sed.texi might just be broken in lines 20 and 36. That kind of accident happens occasionally. Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 3 ==========================** Date: Sun, 16 Nov 2003 09:16:18 -0800 From: "Dave and Natalie" Subject: Re: Use of depcomp On Sun, 16 Nov 2003 11:12:33 +0000, John Poltorak wrote: >> and tried but couldn't find >> ltmain.sh > >Would you expect to find this? Is that part of libtool? Yes, I just installed libtool and now ltmain.sh is installed. Blackbox-cvs only just started needing libtool Dave **= Email 4 ==========================** Date: Sun, 16 Nov 2003 09:22:33 -0800 From: "Dave and Natalie" Subject: Re: Automake problem On Sun, 16 Nov 2003 11:54:23 +0000, John Poltorak wrote: >I haven't had any problems with ftp.gnu.org recently... If you go to eg ftp.gnu.org/gnu/texinfo you will see a lot of files ending in back-RSN.README -> ../../MISSING-FILES Which start out with The following files are not available because their authenticity is being confirmed. We expect to have them uploaded Real Soon Now (RSN). and a big list of files. I guess they got cracked and don't know if any files have been compromised Dave **= Email 5 ==========================** Date: Sun, 16 Nov 2003 09:58:10 -0800 From: "Dave and Natalie" Subject: std::min? Hi, trying to build blackbox-cvs I get this error with both pgcc2.95.3 and gcc 3.2.1 Menu.cc: In method `unsigned int bt::Menu::insertItem(const bt::MenuItem &, unsigned int = 0ffffffff, unsigned int = 0ffffffff)': Menu.cc:355: no matching function for call to `min (unsigned int &, size_t)' Shouldn't std::min be defined in -lgcc or somesuch? I am getting around this with a macro but it doesn't seem like an elegant solution plus the blackbox team are receptive to keeping blackbox building on OS/2 Dave **= Email 6 ==========================** Date: Sun, 16 Nov 2003 11:05:49 +0000 From: John Poltorak Subject: Re: magic On Sat, Nov 15, 2003 at 09:52:54PM -0800, Dave and Natalie wrote: > On Sat, 15 Nov 2003 22:06:22 +0000, John Poltorak wrote: > > > > > > >What is the preferred location for MAGIC? > > > My Debian partition has it in /usr/share/misc > > >That is the reference file used by the FILE utility for identifying many > >different file types. > > > >This is a very useful utility and am pleased to have found that the most > >recent release, v4.06, builds with UX2BS straight out of the box without > >the need for any patches. > > Does it work? Sort of... > I vaguely remember some previous builds having problems with LF vs CR-LF That appears to be happening here with text files, binaries appear to be identified OK though. > Dave > Dave > > -- John **= Email 7 ==========================** Date: Sun, 16 Nov 2003 11:12:33 +0000 From: John Poltorak Subject: Re: Use of depcomp On Sat, Nov 15, 2003 at 10:42:59PM -0800, Dave and Natalie wrote: > On Sat, 15 Nov 2003 20:15:33 +0000, John Poltorak wrote: > > > > >In some old instructions I have for building a couple of apps there is a > >requirement to copy DEPCOMP to the working directory from the AUTOMAKE > >installation. > > > >Can this now be replaced by using:- ? > > > >automake --add-missing > > Well I just ran automake -a on blackbox-cvs and it installed > ./install-sh > ./mkinstalldirs > ./config.guess > ./config.sub > ./depcomp > and tried but couldn't find > ltmain.sh Would you expect to find this? Is that part of libtool? > Dave > ps -a is short for --add-missing > -- John **= Email 8 ==========================** Date: Sun, 16 Nov 2003 11:12:33 +0000 From: John Poltorak Subject: Re: Use of depcomp **= Email 9 ==========================** Date: Sun, 16 Nov 2003 11:27:13 +0000 From: John Poltorak Subject: Re: Makeinfo and sed.texi On Sun, Nov 16, 2003 at 01:59:04AM +0100, Stefan Neis wrote: > On Sat, 15 Nov 2003, John Poltorak wrote: > > > makeinfo -I . \ > > -o sed.info `test -f 'sed.texi' || echo './'`sed.texi > > ./version.texi:20: Misplaced {. > > ./version.texi:20: Misplaced }. > > ./version.texi:36: Unknown command `(#)PD'. > > sed.texi:99: warning: at sc argument all uppercase, thus no effect. > (snipp) > > sed.texi:243: warning: at sc argument all uppercase, thus no effect. > > makeinfo: Removing output file `sed.info' due to errors; use --force to preserve. > > make[1]: *** [sed.info] Error 2 > (snipp) > > > > Is this a problem with MAKEINFO? > > That's one possible explanation - you might be using a makeinfo that's > either too old or too new. I've tried the oldest I can find, along with the newest and both have errors, albeit different ones. > Or sed.texi might just be broken in lines 20 and 36. That kind of accident > happens occasionally. I can't figure out what this version.texi is supposed to be. The original supplied file contains:- at set UPDATED 28 June 1998 at set EDITION 3.02 at set VERSION 3.02 but after Make has run it has the current environment prepended. Very strange! I can't begin to understand what has happened there. I guess it must have something to do with this part of the Makefile:- sed.info: sed.texi version.texi sed.dvi: sed.texi version.texi sed.pdf: sed.texi version.texi version.texi: stamp-vti stamp-vti: sed.texi $(top_srcdir)/configure at (dir=.; test -f ./sed.texi || dir=$(srcdir); \ set `$(SHELL) $(srcdir)/mdate-sh $$dir/sed.texi`; \ echo " at set UPDATED $$1 $$2 $$3"; \ echo " at set UPDATED-MONTH $$2 $$3"; \ echo " at set EDITION $(VERSION)"; \ echo " at set VERSION $(VERSION)") > vti.tmp at cmp -s vti.tmp version.texi \ || (echo "Updating version.texi"; \ cp vti.tmp version.texi) - at rm -f vti.tmp at cp version.texi $ at > > Regards, > Stefan > -- > Micro$oft is not an answer. It is a question. The answer is 'no'. > -- John **= Email 10 ==========================** Date: Sun, 16 Nov 2003 11:54:23 +0000 From: John Poltorak Subject: Re: Automake problem On Sat, Nov 15, 2003 at 09:42:17PM -0800, Dave and Natalie wrote: > On Sat, 15 Nov 2003 22:01:25 +0000, John Poltorak wrote: > > >> I'll try fping > > > >Please do. It looks as though it should be really easy to build, but > >automake seems to mess up the Makefile pretty badly. > > > >If anyone wants to give it a try and you have UX2BS installed you should > >be able to add the following line to build.table:- > > > >fping;;ftp://ftp.mirrors.wiretapped.net/pub/security/network-mapping/fping/fping-2.4b2.tar.gz;;;;;. > > > >and run:- > > > >build fping > > > >It would be nice to figure out why it is failing. > > I don't have UX2BS installed right now (waiting for ftp.gnu.org to replace files) but I did get a correct makefile by running I haven't had any problems with ftp.gnu.org recently... > sh aclocal > sh autoheader (gives warning about deprecated functions etc) > automake -a (copies missing) > sh autoconf Do you get warnings like these:- WARNING: Using auxiliary files such as `acconfig.h', `config.h.bot' WARNING: and `config.h.top', to define templates for `config.h.in' WARNING: is deprecated and discouraged. WARNING: Using the third argument of `AC_DEFINE' and WARNING: `AC_DEFINE_UNQUOTED' allows to define a template without WARNING: `acconfig.h': WARNING: AC_DEFINE([NEED_MAIN], 1, WARNING: [Define if a function `main' is needed.]) WARNING: More sophisticated templates can also be produced, see the WARNING: documentation. I also get the following errors:- configure.in: no proper invocation of AM_INIT_AUTOMAKE was found. configure.in: You should verify that configure.in invokes AM_INIT_AUTOMAKE, configure.in: that aclocal.m4 is present in the top-level directory, configure.in: and that aclocal.m4 was recently regenerated (using aclocal). configure.in: installing `./missing' Makefile.am: installing `./depcomp' /usr/local/share/automake-1.7/am/depend2.am: am__fastdepCC does not appear in AM_CONDITIONAL /usr/local/share/automake-1.7/am/depend2.am: AMDEP does not appear in AM_CONDITIONAL Looks like an important part of Automake is missing from my setup... > Then running make. I get a lot of errors, about 60 lines worth starting with > fping.c:418: conflicting types for `prog' > fping.c:255: previous declaration of `prog' > fping.c:418: `argv' undeclared here (not in a function) > fping.c:418: warning: data definition has no type or storage class > fping.c:419: initializer element is not constant I only get:- make Makefile:15: *** missing separator. Stop. Makefile:15: *** missing separator. Stop. but suspect this is bacause the Makefile wasn't generated correctly. > After touching a source file I have to run > make distclean > sh configure > or make dies here > cd . && at AUTOHEADER at > f:/usr/bin/sh: at AUTOHEADER at : not found > make: *** [stamp-h.in] Error 127 > > Anyways don't have much time to play with fping now > Dave > > > -- John **= Email 11 ==========================** Date: Sun, 16 Nov 2003 12:25:08 +0000 From: John Poltorak Subject: Any shell experts? If we have any shell experts around, could you suggest what this should do:- ? at (dir=.; test -f ./sed.texi || dir=$(srcdir); \ set `$(SHELL) $(srcdir)/mdate-sh $$dir/sed.texi`; \ echo " at set UPDATED $$1 $$2 $$3"; \ echo " at set UPDATED-MONTH $$2 $$3"; \ echo " at set EDITION $(VERSION)"; \ echo " at set VERSION $(VERSION)") > vti.tmp mdate-sh is included in SED but it's probably a general purpose GNU utility which gets modification time of a file or directory and pretty-prints it. -- John **= Email 12 ==========================** Date: Sun, 16 Nov 2003 12:30:21 -0800 From: "Dave and Natalie" Subject: Re: Use of depcomp On Sun, 16 Nov 2003 21:20:15 +0000, John Poltorak wrote: > >I'm still trying to get my head around how libtool works... > >There is a libtool package, yet ltmain.sh seems to be included with a >number of apps these days, so how do you specify the seperate libtool >package rather than the one provided with the app? By running libtoolize. Most likely have to run the rest of the auto tools after. This is the script that comes with blackbox #!/bin/sh libtoolize -c aclocal autoheader automake --foreign -a -c autoconf Dave **= Email 13 ==========================** Date: Sun, 16 Nov 2003 12:31:30 +0000 From: John Poltorak Subject: Re: Any shell experts? On Sun, Nov 16, 2003 at 12:25:08PM +0000, John Poltorak wrote: > mdate-sh is included in SED but it's probably a general purpose GNU > utility which gets modification time of a file or directory and > pretty-prints it. It looks as though this mdate-sh doesn't work correctly on OS/2 because of the output from LS. Removing a couple of 'shifts' from the script gets round the problem. -- John **= Email 14 ==========================** Date: Sun, 16 Nov 2003 12:55:35 +0000 From: John Poltorak Subject: Texinfo at sc attribute Does anyone know what the at sc attribute is for in Texinfo files? The errors I get when running makeinfo on sed.texi are all related to the at sc argument being all uppercase. -- John **= Email 15 ==========================** Date: Sun, 16 Nov 2003 14:40:08 +0100 From: Andreas Buening Subject: Re: Any shell experts? John Poltorak wrote: > > On Sun, Nov 16, 2003 at 12:25:08PM +0000, John Poltorak wrote: > > > mdate-sh is included in SED but it's probably a general purpose GNU > > utility which gets modification time of a file or directory and > > pretty-prints it. > > It looks as though this mdate-sh doesn't work correctly on OS/2 because of > the output from LS. > > Removing a couple of 'shifts' from the script gets round the problem. run 'automake --add-missing --force-missing'. This will hopefully update _all_ automake files in your package, including mdate-sh. If you encounter any makeinfo problems, just 'touch' all existing *.info and *.inf* files before running 'make' the first time (but after autoconf). That's the simplest solution. Bye, Andreas **= Email 16 ==========================** Date: Sun, 16 Nov 2003 14:40:31 +0100 From: Andreas Buening Subject: Re: Use of depcomp John Poltorak wrote: > > On Sat, Nov 15, 2003 at 10:42:59PM -0800, Dave and Natalie wrote: > > and tried but couldn't find > > ltmain.sh > > Would you expect to find this? Is that part of libtool? Yes. Bye, Andreas **= Email 17 ==========================** Date: Sun, 16 Nov 2003 15:39:23 +0000 From: John Poltorak Subject: Re: Any shell experts? On Sun, Nov 16, 2003 at 02:40:08PM +0100, Andreas Buening wrote: > John Poltorak wrote: > > > > On Sun, Nov 16, 2003 at 12:25:08PM +0000, John Poltorak wrote: > > > > > mdate-sh is included in SED but it's probably a general purpose GNU > > > utility which gets modification time of a file or directory and > > > pretty-prints it. > > > > It looks as though this mdate-sh doesn't work correctly on OS/2 because of > > the output from LS. > > > > Removing a couple of 'shifts' from the script gets round the problem. > > run 'automake --add-missing --force-missing'. This will hopefully update > _all_ automake files in your package, including mdate-sh. Thanks for that suggestion. mdate-sh is indeed updated, but it still produces the wrong output. > If you encounter any makeinfo problems, just 'touch' all existing > *.info and *.inf* files before running 'make' the first time (but after > autoconf). That's the simplest solution. Whilst that is the simplest solution for manual intervention is does not easily lend itself to a fully automated build. I propose to put a patch into doc/sed.texi to replace at sc{GNU} with at acronym{GNU} unless someone can suggest an alternative. > > Bye, > Andreas -- John **= Email 18 ==========================** Date: Sun, 16 Nov 2003 16:15:24 +0000 From: John Poltorak Subject: Re: Any shell experts? On Sun, Nov 16, 2003 at 03:39:23PM +0000, John Poltorak wrote: > On Sun, Nov 16, 2003 at 02:40:08PM +0100, Andreas Buening wrote: > > run 'automake --add-missing --force-missing'. This will hopefully update > > _all_ automake files in your package, including mdate-sh. > > Thanks for that suggestion. mdate-sh is indeed updated, but it still > produces the wrong output. Looking at it a little more closely, mdate-sh works correctly when I run it by itself, but when it run as part of the Makefile, it produces:- Making all in doc make[2]: Entering directory `U:/unixos2/workdir/sed-3.02/doc' ./mdate-sh[79]: shift: nothing to shift Updating version.texi make[2]: Leaving directory `U:/unixos2/workdir/sed-3.02/doc' It's not clear to me why it does that. > > Bye, > > Andreas -- John **= Email 19 ==========================** Date: Sun, 16 Nov 2003 17:12:22 +0000 From: John Poltorak Subject: Building GNU Text Utils If someone wants to rebuild the GNU Text Utils, is the procedure still as described in INSTALL.OS2? ie. after applying patches:- aclocal -I m4 automake autoconf autoheader ./configure --prefix=x:/usr --without-included-gettext make -- John **= Email 20 ==========================** Date: Sun, 16 Nov 2003 19:04:15 +0100 From: Andreas Buening Subject: Re: Any shell experts? John Poltorak wrote: > Looking at it a little more closely, mdate-sh works correctly when I run > it by itself, but when it run as part of the Makefile, it produces:- > > Making all in doc > make[2]: Entering directory `U:/unixos2/workdir/sed-3.02/doc' > ./mdate-sh[79]: shift: nothing to shift > Updating version.texi > make[2]: Leaving directory `U:/unixos2/workdir/sed-3.02/doc' > > It's not clear to me why it does that. Maybe you have two mdate-sh in your package? Otherwise I have no idea. For sed.texi, have you checked the latest sources, sed 4.x? Bye, Andreas **= Email 21 ==========================** Date: Sun, 16 Nov 2003 21:20:15 +0000 From: John Poltorak Subject: Re: Use of depcomp On Sun, Nov 16, 2003 at 09:16:18AM -0800, Dave and Natalie wrote: > On Sun, 16 Nov 2003 11:12:33 +0000, John Poltorak wrote: > > >> and tried but couldn't find > >> ltmain.sh > > > >Would you expect to find this? Is that part of libtool? > > Yes, I just installed libtool and now ltmain.sh is installed. Blackbox-cvs only just started needing libtool I'm still trying to get my head around how libtool works... There is a libtool package, yet ltmain.sh seems to be included with a number of apps these days, so how do you specify the seperate libtool package rather than the one provided with the app? > Dave -- John **= Email 22 ==========================** Date: Sun, 16 Nov 2003 22:09:53 +0000 From: John Poltorak Subject: Re: Any shell experts? On Sun, Nov 16, 2003 at 07:04:15PM +0100, Andreas Buening wrote: > John Poltorak wrote: > > > Looking at it a little more closely, mdate-sh works correctly when I run > > it by itself, but when it run as part of the Makefile, it produces:- > > > > Making all in doc > > make[2]: Entering directory `U:/unixos2/workdir/sed-3.02/doc' > > ./mdate-sh[79]: shift: nothing to shift > > Updating version.texi > > make[2]: Leaving directory `U:/unixos2/workdir/sed-3.02/doc' > > > > It's not clear to me why it does that. > > Maybe you have two mdate-sh in your package? It is using the correct version. Originally the error was:- Making all in doc make[2]: Entering directory `U:/unixos2/workdir/sed-3.02/doc' ./mdate-sh[60]: shift: nothing to shift Updating version.texi make[2]: Leaving directory `U:/unixos2/workdir/sed-3.02/doc' Using the updated version of mdate-sh produces:- Making all in doc make[2]: Entering directory `U:/unixos2/workdir/sed-3.02/doc' ./mdate-sh[79]: shift: nothing to shift Updating version.texi make[2]: Leaving directory `U:/unixos2/workdir/sed-3.02/doc' I realise that you have made a change to mdate-sh specifically to get round incorrect parsing of file dates using ls on OS/2, but something is messing things up. > For sed.texi, have you checked the latest sources, sed 4.x? sed.texi from sed v4 uses at acronym{GNU} instead of at sc{GNU}. > > Bye, > Andreas -- John