Date: Sun, 23 Feb 2003 02:35:33 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [Ux2bs_Archive] No. 86 ************************************************** Saturday 22 February 2003 Number 86 ************************************************** Subjects for today 1 Cool or what? : Ted Sikora 2 Re: Perl & Posix/2 : John Poltorak 3 Re: Perl build on JFS : John Poltorak 4 Re[2]: Private archives ? : Csaba 5 Re: Perl & Posix/2 : Henry Sobotka 6 Re: Perl & Posix/2 : Stefan.Neis at t-online.de 7 Re: resolv or 2 : Maynard" 8 Re: Perl & Posix/2 : Henry Sobotka 9 Re: Perl build on JFS : Henry Sobotka 10 Re: Perl build on JFS : John Poltorak 11 Re: Perl & Posix/2 : John Poltorak 12 Re: Perl & Posix/2 : Maynard" 13 Re: Perl & Posix/2 : Stefan.Neis at t-online.de 14 Build munges filesystem : James Moe" 15 Re: Perl & Posix/2 : Stefan.Neis at t-online.de 16 perl build issues : brpms at earthlink.net 17 Re: Perl & Posix/2 : John Poltorak 18 Re: Perl & Posix/2 : Stefan.Neis at t-online.de **= Email 1 ==========================** Date: Sun, 23 Feb 2003 09:57:53 -0500 From: Ted Sikora Subject: Cool or what? http://os2ports.com/content/sections/ux2bs http://os2ports.com/mailman/listinfo/ux2bs All set just waiting for MX records tomorrow. -- Ted Sikora tsikora at ntplx.net _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 2 ==========================** Date: Sun, 23 Feb 2003 10:09:55 +0000 From: John Poltorak Subject: Re: Perl & Posix/2 On Sat, Feb 22, 2003 at 11:55:01PM +0100, Stefan.Neis at t-online.de wrote: > Hi, > > Just to give everyone something to compare too, here are the final > statistics of my Perl build with Posix/2 enabled. Congratulations, Stefan. I think this is a major milestone. > Not quite as good > as before, so this gives some hints for debugging/improving Posix/2. What you probably haven't considered is that this Perl build performs an additional 153 subtests compared with the non-Posix/2 build, so whilst there may be some scope for debugging/improving Posix/2 there is equally as much scope for doing the same with the Perl tests, some of whih may never have been done on OS/2 before and may prove to be invalid in some way. > > Regards, > Stefan > ---------------------------------- > Failed 15/726 test scripts, 97.93% okay. 433/68803 subtests failed, 99.37% okay. > Failed Test Stat Wstat Total Fail Failed List of Failed > ------------------------------------------------------------------------------- > ./ext/DB_File/t/db-btree.t 22 5632 163 145 88.96% 19-163 > ./ext/DB_File/t/db-hash.t 22 5632 117 103 88.03% 15-117 > ./ext/DB_File/t/db-recno.t 2 512 160 144 90.00% 17-160 I had major problems with these tests initially until Sebastian came up with his 'malloc' patch which solved things. Maybe that patch needs to be removed when using Posix/2... > ./ext/IO/lib/IO/t/io_poll.t 10 1 10.00% 3 > ./ext/Time/HiRes/HiRes.t 3 768 25 6 24.00% 20-25 These may be Posix/2 related. > ./lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14 Perl bug. > ./lib/File/Temp/t/posix.t 255 65280 7 5 71.43% 3-7 > ./lib/File/Temp/t/tempfile.t 255 65280 20 11 55.00% 10-20 Posix/2...? > ./lib/Memoize/t/errors.t 11 1 9.09% 4 > ./lib/Memoize/t/tie.t 22 5632 4 4 100.00% 1-4 I had these previosly and they may also be related to the malloc patch. > lib/os2_process.t 3 768 227 3 1.32% 90 174 209 > lib/os2_process_kid.t 227 3 1.32% 90 174 209 > lib/rx_cmprt.t 255 65280 18 3 16.67% 16-18 > op/stat.t 22 5632 73 2 2.74% 72-73 These are standard errors which we are all familiar with. > run/fresh_perl.t 97 1 1.03% 91 Another fail which seemed to disappear after the malloc patch. > 60 tests and 554 subtests skipped. Normally 64 tests and 559 subtests get skipped, so it looks as though this build is tested more rigorously. I would like to be able to repeat these results, but have been unable to get Perl to build. My latest attempt ends with:- perlio.c: In function `PerlIOStdio_set_ptrcnt': perlio.c:2898: structure has no member named `ptr' perlio.c:2918: structure has no member named `rcount' perlio.c: In function `PerlIOStdio_fill': perlio.c:2988: structure has no member named `ptr' perlio.c:2988: structure has no member named `buffer' perlio.c:2988: structure has no member named `ptr' perlio.c:2988: structure has no member named `ptr' perlio.c:2988: structure has no member named `rcount' perlio.c:2988: warning: left-hand operand of comma expression has no effect perlio.c:2988: warning: left-hand operand of comma expression has no effect make: *** [perlio.obj] Error 1 -- John _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 3 ==========================** Date: Sun, 23 Feb 2003 10:43:24 +0000 From: John Poltorak Subject: Re: Perl build on JFS On Sat, Feb 22, 2003 at 06:50:29PM -0500, Henry Sobotka wrote: > John Poltorak wrote: > > > > That is not much use on OS/2. > > > > Should we get this changed to %ETC%\RESOLV(2) ? > > Something like: > > my $resolv = "/etc/resolv.conf"; > $resolv = $ENV{"ETC"}."/resolv" if $^O eq "os2"; > > if (open (RES, $resolv)) { > > would be more acceptable than simply changing it. The ideal solution would be if IBM's resolver routines could actually use resolv.conf, but I don't think they can. However given that this is a test environment for Perl, we could easily create resolv.conf... I wonder if Perl itself would use it if it existed, or does it rely on MPTS for host name resolution? > But here "domain" is > only in resolv; resolv2 just lists nameservers; and there's no > resolv(2). I have never been very clear about the difference in usage between RESOLV and RESOLV2. Originally (TCP/IP v1.2) there was only RESOLV, but at some point RESOLV2 was introduced and no mention is made of RESOLV although it is used if available. > So trying "every conceivable way" would require something > like: > > if ($^O eq "os2") { > my at resolv = ("/resolv", "/resolv2", "/resolv(2)"); > foreach my $r ( at resolv) { > $domain = read_res($ENV{"ETC"}.$r); > next unless defined $domain; > } > } > else { > $domain = read_res("/etc/resolv.conf"); > } > > with the read process: > > sub read_res { > if (open (RES, shift)) { > > etc. returning $domain if defined. > > h~ -- John _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 4 ==========================** Date: Sun, 23 Feb 2003 11:58:40 +0000 From: Csaba Subject: Re[2]: Private archives ? Hello John, Friday, February 21, 2003, 10:31:26 PM, you wrote: [snip] JP> Is that the real Csaba? Or an imposter? :-)... The one and only :-) I have recently switched mail clients and the Bat! has threading of emails so I added another mailing list. -- Csaba mailto:adwx88 at uk.uumail.com Ceci n'est pas un .signature! _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 5 ==========================** Date: Sun, 23 Feb 2003 11:59:19 -0500 From: Henry Sobotka Subject: Re: Perl & Posix/2 Stefan.Neis at t-online.de wrote: > > Any idea how to find out what perl is actually doing in this one line of code? Running "perl -d [scriptname]" puts you into the Perl debugger. For usage details see perldebug.pod and perldebtut.pod (tutorial). If you want to step into Perl itself, though, you'll have to do a debug build and use an OMF-capable debugger. And of course there's always good old "print '\$foo=$foo\n'". > > (my $fh, $tmpnam) = tmpnam(); > Is that executing just the C function or is there a perl wrapper around it, > that is called in this line? Perl scripts never execute C functions directly; it's calling the tmpnam on line 2722 of POSIX.c, which is generated from 1454 of POSIX.xs. This particular call is further complicated by the fact that tmpnam is mapped to my_tmpnam in os2ish.h; my_tmpnam is implemented in os2.c. So that's very likely what it's really calling. h~ _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 6 ==========================** Date: Sun, 23 Feb 2003 13:03:04 +0100 From: Stefan.Neis at t-online.de Subject: Re: Perl & Posix/2 Hi, > What you probably haven't considered is that this Perl build performs an > additional 153 subtests compared with the non-Posix/2 build, That's something I was hoping, but I didn't have numbers to compare with at hand... ;-) > > Failed 15/726 test scripts, 97.93% okay. 433/68803 subtests failed, 99.37% okay. > > Failed Test Stat Wstat Total Fail Failed List of Failed > > ------------------------------------------------------------------------------- > > ./ext/DB_File/t/db-btree.t 22 5632 163 145 88.96% 19-163 > > ./ext/DB_File/t/db-hash.t 22 5632 117 103 88.03% 15-117 > > ./ext/DB_File/t/db-recno.t 2 512 160 144 90.00% 17-160 > > I had major problems with these tests initially until Sebastian came up > with his 'malloc' patch which solved things. Maybe that patch needs to be > removed when using Posix/2... Looking through the detailed error messages, this looks more like a real error in DB code. BTW, it's db-1.85 which is in Posix/2. :-( > > ./ext/IO/lib/IO/t/io_poll.t 10 1 10.00% 3 > > ./ext/Time/HiRes/HiRes.t 3 768 25 6 24.00% 20-25 > > These may be Posix/2 related. For HiRes, it looks like something goes wrong with signal handling, the process apparently dies on a SIGALRM, that should tell it to continue its work. :-( > > ./lib/File/Temp/t/posix.t 255 65280 7 5 71.43% 3-7 > > ./lib/File/Temp/t/tempfile.t 255 65280 20 11 55.00% 10-20 > > Posix/2...? Error messages as displayed in the log is (for posix.t): ./lib/File/Temp/t/posix...............Error in mkstemp using E:/MPTN/ETC/XXXXXXXXXX: Parent directory (E:/MPTN/ETC/) is not a directory at ../lib/File/Temp/t/posix.t line 32 dubious Test returned status 255 (wstat 65280, 0xff00) DIED. FAILED tests 3-7 Failed 5/7 tests, 28.57% okay First, it's _not_ a drive letter issue as I get the same kind of error messagen when doing set TEMP=/tmp (which exists!) and rerunning the test, only that now it claims that /tmp/ is no directory. Second, tempnam and mkstemp are EMX function, no redefinition in Posix/2! :-( Any idea how to find out what perl is actually doing in this one line of code? I'm not really familiar with perl... > (my $fh, $tmpnam) = tmpnam(); Is that executing just the C function or is there a perl wrapper around it, that is called in this line? Also note that EMX manual says, that the string passed to mkstemp should have 6 trailing 'X' characters, while the above has ten. So whatever really is wrong here, I don't understand at all, what's the difference to the working (i.e. pure EMX) case. The problem for tempfile.t is rather similar, and I wonder if that problem affects the temporary files created for the db-tests above as well. > I would like to be able to repeat these results, but have been unable to > get Perl to build. My latest attempt ends with:- > > perlio.c: In function `PerlIOStdio_set_ptrcnt': > perlio.c:2898: structure has no member named `ptr' > perlio.c:2918: structure has no member named `rcount' I had the same kind of problem quite some time ago, now if I only could remember, how I got around that .... .. One upacking of original sources and "diff --recursive" later: There's one more fix required in hints/os2.sh: Replace stdstdunder=`echo "#include " | cpp | egrep -c "char +\* +_ptr"` by stdstdunder=`echo "#include " | cpp | egrep -c "char *\* *_ptr"` (i.e. replace two '+' by '*'). There is no reason, why a header file shouldn't be allowed to write "char* _ptr" or "char *_ptr" or even "char*_ptr", it's nonsens to enforce at least one blank before and after the "*". Also you might want to #include and not #include in os2/os2ish.h, but I believe this change is no longer necessary after the setmode issue is fixed in Posix/2. Regards, Stefan _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 7 ==========================** Date: Sun, 23 Feb 2003 13:18:09 -0600 (CST) From: "Maynard" Subject: Re: resolv or 2 Hi Henry, >I noticed last night that resolv is created when I go online with >dial-up, and deleted when I hang up. resolv2 persists. with InJoy at least, and likely other dialers, the dialer maintains configuration for domain and DNS servers; these it puts into resolv when it controls the net connection. InJoy first copies resolv to resolv.inj before writing its configuration to resolv. This is quite practical and reasonable when considered that a dialer can make several different connections, representing different domains and certainly different nameservers. None of this however explains the distinctions between resolv and resolv2 which I had always understood to be resolv for DOS and resolv2 for OS/2 apps; but I certainly don't have any confidence in this or other distinctions, and have always just maintained them the same. -- Maynard _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 8 ==========================** Date: Sun, 23 Feb 2003 13:26:24 -0500 From: Henry Sobotka Subject: Re: Perl & Posix/2 Stefan.Neis at t-online.de wrote: > > Now my problem is that it apparently does not end up calling that function.. > I should see something like "pltmp" in the filename quoted in the error > message, if it is going through that routine, shouldn't I? Very confusing... (When your brain's a pretzel, might as well have a beer to go with it.) It just occurred to me that when building POSIX, "probably harmless" warnings about -lposix not found occur; so this tmpnam may well be directly calling the one in -lposix (or Posix/2 in this case), i.e. bypassing the one in os2.c, because with POSIX::tmpnam() you would want the POSIX and not Perl function. h~ _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 9 ==========================** Date: Sun, 23 Feb 2003 13:42:20 -0500 From: Henry Sobotka Subject: Re: Perl build on JFS John Poltorak wrote: > > However given that this is a test environment for Perl, we could easily > create resolv.conf... I wonder if Perl itself would use it if it existed, > or does it rely on MPTS for host name resolution? It uses the gethost* C functions which probably use MPTS on OS/2. > I have never been very clear about the difference in usage between RESOLV > and RESOLV2. Originally (TCP/IP v1.2) there was only RESOLV, but at some > point RESOLV2 was introduced and no mention is made of RESOLV although it > is used if available. I noticed last night that resolv is created when I go online with dial-up, and deleted when I hang up. resolv2 persists. The former appears to contain the data from dial-up configuration, the latter from TCP/IP config. h~ _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 10 ==========================** Date: Sun, 23 Feb 2003 14:16:58 +0000 From: John Poltorak Subject: Re: Perl build on JFS On Mon, Feb 24, 2003 at 12:41:48AM +1100, IanM wrote: > Hi > > I've just build perl on a bootable JFS partition, PIII-600, 512Mb RAM Well done, Ian. I was wondering whether you had ever tried it... As for a *bootable* JFS partition... I have to ask how you have managed to create such a beast since my understanding is that JFS is not bootable at this point in time... Are you using some sort of eCS enhancement to make it such? > Failed 6/726 test scripts, 99.17% okay. 130/68687 subtests failed, 99.81% okay. Was one of the fails in lib/os2_ea.t ? You seem to have performed more subtests than normal. I wonder why... > Sun Feb 23 23:37:25 GMT 2003 > elapsed time: 2701 secs > > Must say John, your ux2_bootstrap.cmd works wonders, and > rsync worked perfectly through my firewall this time. That's good to know. Now that you have a working Perl, can you try running update_base.cmd? > Cheers > IanM > http://www.os2site.com/ > > This is your brain on Windows: Any questions? -- John _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 11 ==========================** Date: Sun, 23 Feb 2003 14:52:21 +0000 From: John Poltorak Subject: Re: Perl & Posix/2 On Sun, Feb 23, 2003 at 01:03:04PM +0100, Stefan.Neis at t-online.de wrote: > > > > Failed 15/726 test scripts, 97.93% okay. 433/68803 subtests failed, 99.37% okay. > > > Failed Test Stat Wstat Total Fail Failed List of Failed > > > ------------------------------------------------------------------------------- > > > ./ext/DB_File/t/db-btree.t 22 5632 163 145 88.96% 19-163 > > > ./ext/DB_File/t/db-hash.t 22 5632 117 103 88.03% 15-117 > > > ./ext/DB_File/t/db-recno.t 2 512 160 144 90.00% 17-160 > > > > I had major problems with these tests initially until Sebastian came up > > with his 'malloc' patch which solved things. Maybe that patch needs to be > > removed when using Posix/2... > > Looking through the detailed error messages, this looks more like a real > error in DB code. BTW, it's db-1.85 which is in Posix/2. :-( I'd really like Sebastian have a look at this and maybe figure out a way of removing his patches which were applied last June/July wrt to malloc. My instinct tells me it is somehow related... > > I would like to be able to repeat these results, but have been unable to > > get Perl to build. My latest attempt ends with:- > > > > perlio.c: In function `PerlIOStdio_set_ptrcnt': > > perlio.c:2898: structure has no member named `ptr' > > perlio.c:2918: structure has no member named `rcount' > > I had the same kind of problem quite some time ago, now if I only > could remember, how I got around that .... > .. > One upacking of original sources and "diff --recursive" later: > There's one more fix required in hints/os2.sh: > Replace > stdstdunder=`echo "#include " | cpp | egrep -c "char +\* +_ptr"` > by > stdstdunder=`echo "#include " | cpp | egrep -c "char *\* *_ptr"` > (i.e. replace two '+' by '*'). > There is no reason, why a header file shouldn't be allowed to > write "char* _ptr" or "char *_ptr" or even "char*_ptr", it's nonsens > to enforce at least one blank before and after the "*". > > Also you might want to #include and not #include > in os2/os2ish.h, but I believe this change is no longer necessary > after the setmode issue is fixed in Posix/2. I've made the change above to hints/os2.h and do get a build of some sort but a lot more errors. I suspect I have missed out one or more of your changes. Can you summarise what changes you made? I'm using p2-gcc.exe and have Posix/2 on INCLUDE and LIBRARY paths. There is a sed script for adjusting the the test for c:/. Patches for Configure (add sed '/\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\/' :-)) os2/os2.c Change as mentioned above to hints/os2.h There's something else I've overlooked, but can't think what... In my recent build I get this error msg:- Can't open OS2/Process/Const/os2emx.h: No such file or directory Can't build module with contants, falling back to no constants at Makefile.PL line 30. I suspect it is something I've overlooked from hints/os2.h. > Regards, > Stefan -- John _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 12 ==========================** Date: Sun, 23 Feb 2003 16:22:05 -0600 (CST) From: "Maynard" Subject: Re: Perl & Posix/2 John, >I'm currently stumped. I started again from scratch with a reboot. I'm >building everything on a RAMFS drive It is possible that the ram drive is the problem. -- Maynard _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 13 ==========================** Date: Sun, 23 Feb 2003 17:47:46 +0100 From: Stefan.Neis at t-online.de Subject: Re: Perl & Posix/2 Hi, > > Looking through the detailed error messages, this looks more like a real > > error in DB code. BTW, it's db-1.85 which is in Posix/2. :-( > > I'd really like Sebastian have a look at this and maybe figure out a way > of removing his patches which were applied last June/July wrt to malloc. > My instinct tells me it is somehow related... I'd expect to get some SYSsomething popup in this case, like I did prior to removing \usr\bin\perl.exe und \usr\lib\perl*dll from the previous (non-Posix/2) build (Thanks for that hint, BTW ;-) ). > > Also you might want to #include and not #include > > in os2/os2ish.h, but I believe this change is no longer necessary > > after the setmode issue is fixed in Posix/2. BTW, I meanwhile tested it without those changes to os2/os2ish.h, they really are no longer necessary. :-) > I've made the change above to hints/os2.h and do get a build of some sort > but a lot more errors. You're sure, there's no older exe or dll that could interfere like it did over here? ;-) > I suspect I have missed out one or more of your changes. > Can you summarise what changes you made? > I'm using p2-gcc.exe and have Posix/2 on INCLUDE and LIBRARY paths. .. while I'm still using standard gcc and adding "-lcExt -lstcExt" explicitly in hints/os2.sh (as I started this text adventure prior to p2-gcc's existence). This results in a difference in linking order (you'll end up with p2-gcc doing "-lsocket -lm -lbsd -lcExt -lstcExt" while I am using "-lcExt -lstcExt -lsocket -lm -lbsd") - this might make a difference (in that case, cleaning the libs in hints/os2.sh should help, socket library is linked implicitly anyway by p2-gcc and "m" and "bsd" shouldn't be needed any more with Posix/2. I'll give p2-gcc a try and see myself what happens. For the rest, I don't see anything missing in your list as compared to what diff --recursive tells me about my changes over here. Maybe send me your log file (preferably compressed) and I'll give it a quick check if I see something that is especially suspicious... > In my recent build I get this error msg:- > > Can't open OS2/Process/Const/os2emx.h: No such file or directory > Can't build module with contants, falling back to no constants at Makefile.PL line 30. Strange. At least I never noticed such an error ... Regards, Stefan _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 14 ==========================** Date: Sun, 23 Feb 2003 18:03:50 -0700 (MST) From: "James Moe" Subject: Build munges filesystem -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, I ran ux2_bootstrap.cmd and it built a unixos2 environment, and successfully built perl v5.8.0. It has not built anything since. A typical log file is listed below (it's short) for autoconf v2.50. At the end is a message "build.sh[153]: cannot fork - try again" which signals the end of the filesystem's usefulness. Yes, the filesystem. Very fortunately this is not a permanent condition; all is functional again after a (required!) reboot. One of the side effects of munging the filesystem is that errors messages from build show up as files ("a non-recoverable error has occurred..."), and subdirectories do not exist. Also new os/2 windows do not open. Line 153 of build.sh: cp -p $BLD_HOME/lib/aux/mkinstalldirs . i:/unixos2/lib/aux/mkinstalldirs is not updated. System: amd ahtlon 1800 512 MB ram c: scsi Quantum viking 4.5GB d:, e: scsi Fujitsu 18GB os/2 server v4.5 fp xr-e003 jfs: Oct 16 2002 c: hpfs386 d: jfs e: jfs i: compressed drive on E: using Zipstream (Yes, I did try to install it directly on E:. It made no difference.) - ----[ autoconf build log (abbreviated) ]---- DIR URL CFLAGS LDFLAGS CFGPARMS MAKEPARM SRC ftp://ftp.gnu.org/pub/gnu/autoconf/autoconf-2.50.tar.gz --prefix=${UXRT}/usr/local/TMP . URL ftp://ftp.gnu.org/pub/gnu/autoconf/autoconf-2.50.tar.gz CFLAGS LDFLAGS CFGPARMS --prefix=${UXRT}/usr/local/TMP MAKEPARM SRC . REPOSITORY/ARCHIVEFILE = i:/unixos2/archives/source/source/source/autoconf-2.50.tar.gz autoconf-2.50.tar.gz already retrieved i:/unixos2/workdir i:/unixos2/workdir/autoconf-2.50 patching file `acgeneral.m4' Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 4 out of 4 hunks ignored -- saving rejects to acgeneral.m4# ...[ several other ignored patches ]... ./configure --prefix=${UXRT}/usr/local/TMP CONFIGURE.: loading site script i:/unixos2/lib/config.site CONFIGURE.: creating cache /dev/null checking for a BSD compatible install... i:/usr/bin/install.exe checking whether build environment is sane... yes checking whether make sets ${MAKE}... yes checking for working aclocal... missing checking for working autoconf... missing checking for working automake... cp -p i:/unixos2/lib/aux/mkinstalldirs build.sh[153]: cannot fork - try again - -- jimoe at sohnen-moe dot com -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 5.0 OS/2 for non-commercial use Comment: PGP 5.0 for OS/2 Charset: cp850 wj8DBQE+WW92sxxMki0foKoRAlNbAKDxwOVe5HFjM2UZQeHMSmfSxVCgyACgyiA9 uBywcDeodZuZ5uX39WC2Ncw= =fXNV -----END PGP SIGNATURE----- _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 15 ==========================** Date: Sun, 23 Feb 2003 18:34:37 +0100 From: Stefan.Neis at t-online.de Subject: Re: Perl & Posix/2 Hi, > Perl scripts never execute C functions directly; it's calling the tmpnam > on line 2722 of POSIX.c, which is generated from 1454 of POSIX.xs. Thanks for explaining ... > This > particular call is further complicated by the fact that tmpnam is mapped > to my_tmpnam in os2ish.h; my_tmpnam is implemented in os2.c. So that's > very likely what it's really calling. Now my problem is that it apparently does not end up calling that function.. I should see something like "pltmp" in the filename quoted in the error message, if it is going through that routine, shouldn't I? Very confusing... Regards, Stefan _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 16 ==========================** Date: Sun, 23 Feb 2003 21:05:40 -0700 From: brpms at earthlink.net Subject: perl build issues My computer froze during the perl build of ux2_bootstrap [too many mozilla windows probably]. After the reboot, I deleted the perl part of workdir and did build perl. This second build seems to give better statistics at the end: Failed 4/726 test scripts, 99.45% okay. 6/68652 subtests failed, 99.99% okay. Failed Test Stat Wstat Total Fail Failed List of Failed ------------------------------------------------------------------------------- ../lib/ExtUtils/t/basic.t 1 256 17 1 5.88% 14 ../lib/Net/t/hostname.t 2 1 50.00% 1 lib/os2_process.t 2 512 227 2 0.88% 174 209 lib/os2_process_kid.t 227 2 0.88% 174 209 63 tests and 557 subtests skipped. Are the skipped tasks significant? Are they supposed to be skipped? Here is an example that I don't understand t/io/pipe............................Error reading "/no_such_process": No such file or directory at io/pipe.t line 186. g:/usr/bin/sh.exe: /no_such_process: not found ok the file pipe.t is in that directory g:/usr/bin/sh.exe exists why can't the install find them? Thanks for the help. If I learn enough I may be able to contribute. Paul Schwartz -- ----------------------------------------------------------- brpms at earthlink.net ----------------------------------------------------------- _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 17 ==========================** Date: Sun, 23 Feb 2003 21:37:40 +0000 From: John Poltorak Subject: Re: Perl & Posix/2 On Sun, Feb 23, 2003 at 10:04:58PM +0100, Stefan.Neis at t-online.de wrote: > Hi, > > > > I'm using p2-gcc.exe and have Posix/2 on INCLUDE and LIBRARY paths. > > > > ... while I'm still using standard gcc and adding "-lcExt -lstcExt" explicitly > > in hints/os2.sh (as I started this text adventure prior to p2-gcc's existence). > > This results in a difference in linking order (you'll end up with p2-gcc doing > > "-lsocket -lm -lbsd -lcExt -lstcExt" while I am using "-lcExt -lstcExt -lsocket > > -lm -lbsd") - this might make a difference (in that case, cleaning the > > libs in hints/os2.sh should help, socket library is linked implicitly anyway by > > p2-gcc and "m" and "bsd" shouldn't be needed any more with Posix/2. > > I'll give p2-gcc a try and see myself what happens. > > No visible difference over here (with p2-gcc and without explicitly adding > -lcExt -lstcExt, which is very reassuring), except for a variation in the > numbers of errors in os2_process(_kid).t, but those failures seem to be > probabilistic. :-( In two successive runs, I never manage to get the exactly > same result. Sometimes it's only "90 174 209" failing, then it's "90 106-227". > Sometimes the errors for os2_process.t and os2_process_kid.t are the same, > sometimes they are not. :-( I'm currently stumped. I started again from scratch with a reboot. I'm building everything on a RAMFS drive so there is no chance of picking up stuff from leftover from a previous build. What happens is the build appears to go fine. There are no major error msgs. perl.exe and perlB12E.dll get built OK but the build script just ends prematurely without running 'make test', and explantion for the early exit. Very strange... I'm running 'make test' now and that seems to be running OK. > Regards, > Stefan -- John _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs **= Email 18 ==========================** Date: Sun, 23 Feb 2003 22:04:58 +0100 From: Stefan.Neis at t-online.de Subject: Re: Perl & Posix/2 Hi, > > I'm using p2-gcc.exe and have Posix/2 on INCLUDE and LIBRARY paths. > > ... while I'm still using standard gcc and adding "-lcExt -lstcExt" explicitly > in hints/os2.sh (as I started this text adventure prior to p2-gcc's existence). > This results in a difference in linking order (you'll end up with p2-gcc doing > "-lsocket -lm -lbsd -lcExt -lstcExt" while I am using "-lcExt -lstcExt -lsocket > -lm -lbsd") - this might make a difference (in that case, cleaning the > libs in hints/os2.sh should help, socket library is linked implicitly anyway by > p2-gcc and "m" and "bsd" shouldn't be needed any more with Posix/2. > I'll give p2-gcc a try and see myself what happens. No visible difference over here (with p2-gcc and without explicitly adding -lcExt -lstcExt, which is very reassuring), except for a variation in the numbers of errors in os2_process(_kid).t, but those failures seem to be probabilistic. :-( In two successive runs, I never manage to get the exactly same result. Sometimes it's only "90 174 209" failing, then it's "90 106-227". Sometimes the errors for os2_process.t and os2_process_kid.t are the same, sometimes they are not. :-( Regards, Stefan _______________________________________________ UX2BS mailing list UX2BS at powerusersbbs.net http://powerusersbbs.net/mailman/listinfo/ux2bs