From: UnixOS2 Archive To: "UnixOS2 Archive" Date: Mon, 28 Jan 2002 04:10:08 EST-10EDT,10,-1,0,7200,3,-1,0,7200,3600 Subject: [UnixOS2_Archive] No. 118 ************************************************** Sunday 27 January 2002 Number 118 ************************************************** Subjects for today 1 Re: Building Perl 5.6.1 from source : Lyn St George" 2 Re: Building Perl 5.6.1 from source : csaba.raduly at sophos.com 3 Re: Perl enhancements : Henry Sobotka 4 Re: Building Perl 5.6.1 from source : John Poltorak 5 Re: Building Perl 5.6.1 from source : John Poltorak 6 Re: OT: stdout/stderror : John Poltorak 7 Re: Perl enhancements : Henry Sobotka 8 OT: stdout/stderror : Adrian Gschwend" 9 Re: Perl enhancements : Henry Sobotka 10 Re: OT: stdout/stderror : Michael Warmuth 11 Re: OT: stdout/stderror : Adrian Gschwend" 12 Re: OT: stdout/stderror : Achim Hasenmueller" 13 Re: Make 3.79.1 : Henry Sobotka 14 An archive for list? : email at eracc.hypermart.net 15 Perl enhancements : John Poltorak 16 Perl enhancements : John Poltorak 17 Perl enhancements : John Poltorak 18 Perl enhancements : John Poltorak 19 Perl enhancements : John Poltorak 20 Re: Building Perl 5.6.1 from source : John Poltorak 21 Re: Make 3.79.1 : Henry Sobotka 22 Re: Building Perl 5.6.1 from source : Stefan Neis 23 Re: Building Perl 5.6.1 from source : Stefan Neis 24 Re: Building Perl 5.6.1 from source : Stefan Neis 25 Re: Building Perl 5.6.1 from source : Stefan Neis 26 Re: Building Perl 5.6.1 from source : Stefan Neis 27 Re: Perl enhancements : Lyn St George" 28 Re: Building Perl 5.6.1 from source : Andreas Buening 29 Re: Make 3.79.1 : Andreas Buening 30 Re: Make 3.79.1 : John Poltorak 31 Re: Make 3.79.1 : Andreas Buening 32 Re: An archive for list? : John Poltorak 33 Re: Make 3.79.1 : Andreas Buening 34 Re: An archive for list? : Holger Veit **= Email 1 ==========================** Date: Mon, 28 Jan 2002 02:06:24 +0000 From: "Lyn St George" Subject: Re: Building Perl 5.6.1 from source On Sun, 27 Jan 2002 23:29:34 +0000, John Poltorak wrote: > >About a year ago I managed to build Perl 5.6.0 from the latest stable >tarball and was quite surprised how relatively straightforward it was. >Now I'm attemepting to do the same with 5.6.1 although there has been no >updates for a long time, but this time I'm hoping to create a UnixOS/2 >package from the build. > >Unfortunately when I run the same script as before, 'Make install' says >`install' is up to date. Same thing here, but there is something called 'installperl' which seems to do the job (at least I /think/ that's what I used) >In the process of testing the build, 5 scripts failed out of 258 - that's >98.06% pass. Is that as good as I can expect? I got 3 scripts failed, though so far none of them seem too important. I've put some notes up on os2docs.org if you want to compare notes. It will build into whatever directory structure you like, so as far as creating a "UnixOS/2" package is concerned you just need to give it a suitable prefix, eg /usr, and it will do the rest. >I'm prepared to go through the fails just to see if they are to be >expected. Also, quite a lot of tests have been skipped and I'd like to >know if they should have been skipped. >-- >John Cheers Lyn St George +--------------------------------------------------------------------------------- + http://www.zolotek.net .. eCommerce hosting, consulting + http://www.os2docs.org .. some 'How To' stuff ... +---------------------------------------------------------------------------------- **= Email 2 ==========================** Date: Mon, 28 Jan 2002 09:58:03 +0000 From: csaba.raduly at sophos.com Subject: Re: Building Perl 5.6.1 from source On 27/01/2002 23:29:34 owner-os2-unix wrote: >About a year ago I managed to build Perl 5.6.0 from the latest stable >tarball and was quite surprised how relatively straightforward it was. >Now I'm attempting to do the same with 5.6.1 although there has been no >updates for a long time, but this time I'm hoping to create a UnixOS/2 >package from the build. > >Unfortunately when I run the same script as before, 'Make install' says >`install' is up to date. > That's case blindness for you :-( INSTALL (note the uppercase) is a text file with explanations. install is meant as a pseudotarget. Unfortunately case-insensitive make or (HPFS) sees it as a file to be updated. What I've done: rename INSTALL to INSTALL1, and edit Manifest to match it (otherwise "Checking to see if your kit is complete" will fail). > >In the process of testing the build, 5 scripts failed out of 258 - that's >98.06% pass. Is that as good as I can expect? > I think so, give or take couple of tests. >I'm prepared to go through the fails just to see if they are to be >expected. Also, quite a lot of tests have been skipped and I'd like to >know if they should have been skipped. > IlyaZ could tell you more, but tests are not skipped by accident. (although it's the easy way to reduce failed tests :-) -- Csaba Ráduly, Software Engineer Sophos Anti-Virus email: csaba.raduly at sophos.com http://www.sophos.com US Support: +1 888 SOPHOS 9 UK Support: +44 1235 559933 **= Email 3 ==========================** Date: Mon, 28 Jan 2002 10:02:23 -0500 From: Henry Sobotka Subject: Re: Perl enhancements John Poltorak wrote: > > IMV we should include the following as standard in os2ish.h before > building Perl:- > > #define HAS_PASSWD > #define HAS_GETPWENT > #define HAS_CRYPT > > Is there any reason not to do so? Crypt works, but as long as we only have fake password support in EMX, I think it's wiser to leave the other two turned off. One of the main uses of Perl is with CGI and Apache, and pretending to provide security features that aren't really there is not a risk worth exposing users to. h~ **= Email 4 ==========================** Date: Mon, 28 Jan 2002 10:24:35 +0000 From: John Poltorak Subject: Re: Building Perl 5.6.1 from source On Mon, Jan 28, 2002 at 09:58:03AM +0000, csaba.raduly at sophos.com wrote: > > On 27/01/2002 23:29:34 owner-os2-unix wrote: > > >Unfortunately when I run the same script as before, 'Make install' says > >`install' is up to date. > > > > That's case blindness for you :-( > INSTALL (note the uppercase) is a text file with explanations. > install is meant as a pseudotarget. > Unfortunately case-insensitive make or (HPFS) sees it as a file > to be updated. Aahh I see... Thanks for the explanantion. > What I've done: rename INSTALL to INSTALL1, and edit Manifest > to match it (otherwise "Checking to see if your kit is complete" > will fail). OK this works, although I did not need to do anything like this with 5.6.0. Wonder what has changed to cause this problem... Do you know if this has been reported? It's a pain having to mess about like this. > > > >In the process of testing the build, 5 scripts failed out of 258 - that's > >98.06% pass. Is that as good as I can expect? > > > > I think so, give or take couple of tests. I'd really like to get this build as good as possible. In a way it also provides a good test of my development environment. > -- > Csaba Ráduly, Software Engineer Sophos Anti-Virus > email: csaba.raduly at sophos.com http://www.sophos.com > US Support: +1 888 SOPHOS 9 UK Support: +44 1235 559933 > -- John **= Email 5 ==========================** Date: Mon, 28 Jan 2002 10:59:56 +0000 From: John Poltorak Subject: Re: Building Perl 5.6.1 from source On Mon, Jan 28, 2002 at 02:06:24AM +0000, Lyn St George wrote: > On Sun, 27 Jan 2002 23:29:34 +0000, John Poltorak wrote: > > > >Unfortunately when I run the same script as before, 'Make install' says > >`install' is up to date. > > Same thing here, but there is something called 'installperl' which seems > to do the job (at least I /think/ that's what I used) At least I'm getting results consistant with others... Following the post from Csaba Raduly I have now got 'make install' running. I did try 'installperl' and thought it worked, but it omits a couple of directories. I'd check the BIN and MAN dirs on your system to see if you have all the correct files. > >In the process of testing the build, 5 scripts failed out of 258 - that's > >98.06% pass. Is that as good as I can expect? > > I got 3 scripts failed, though so far none of them seem too important. Here's what I got:- lib/bigfltpm.........FAILED at test 165 lib/db-btree.........FAILED at test 0 lib/db-hash..........FAILED at test 0 lib/db-recno.........FAILED at test 0 lib/rx_cmprt.........FAILED at test 16 How do I get rid of any of those? > I've put some notes up on os2docs.org if you want to compare notes. I didn't know about this. Looks useful.. > It will build into whatever directory structure you like, so as far as > creating a "UnixOS/2" package is concerned you just need to give > it a suitable prefix, eg /usr, and it will do the rest. What prefix does Slackware use? If anyone has Slackware could you run 'Perl -V' and tell me what prefix is used? I hate having f:/emx as it causes me so many problems. I've used c:/usr/lib/perl which I think is in line with Slackware but I don't have a Unix system at hand to check. > Cheers > Lyn St George > +--------------------------------------------------------------------------------- > + http://www.zolotek.net .. eCommerce hosting, consulting > + http://www.os2docs.org .. some 'How To' stuff ... > +---------------------------------------------------------------------------------- > -- John **= Email 6 ==========================** Date: Mon, 28 Jan 2002 11:29:47 +0000 From: John Poltorak Subject: Re: OT: stdout/stderror On Mon, Jan 28, 2002 at 11:55:23AM +0100, Adrian Gschwend wrote: > I have a question regarding stdout/stderror. There is a java > application which explores networks, a lot of code relies on catching > the ping-output which is called external in the guest-OS (there are no > free ping-implementations in Java). > > The application does not work on OS/2, the stdout hook does not get any > text as it looks like. Is there a way to check if the OS/2-ping does > write to stdout? If it's really an error of the OS/2 ping I would have > to try another ping-port. Does this answer your question? :- C:\>ping 192.168.0.1 64 1 PING 192.168.0.1: 64 data bytes 72 bytes from 192.168.0.1: icmp_seq=0. time=0. ms ----192.168.0.1 PING Statistics---- 1 packets transmitted, 1 packets received, 0% packet loss round-trip (ms) min/avg/max = 0/0/0 C:\>ping 192.168.0.1 64 1 1>nul C:\> > cu > > Adrian > > > -- > Adrian Gschwend > at OS/2 Netlabs > > ICQ: 22419590 > ktk at netlabs.org > ------- > The OS/2 OpenSource Project: > http://www.netlabs.org > -- John **= Email 7 ==========================** Date: Mon, 28 Jan 2002 11:45:03 -0500 From: Henry Sobotka Subject: Re: Perl enhancements IanM wrote: > > Does the recent builds of PERL for OS/2 have the Crypt function ? The two packages I put on Hobbes do, and this simple test script: #! perl -w use strict; my at passes = qw(foo Foo fOo FOo foO FoO fOO FOO); my $key = "Q7"; foreach my $pass( at passes) { my $encrypted = crypt($pass, $key); print "$encrypted\n"; } exit 0; 1; outputs: Q7RBM4sjoXYag Q7C4XHAzQtPJ2 Q73cloVCS/z0g Q7ta.2RHZRujI Q70/DDmSJs2yQ Q7TonR707kzJg Q79p5Eq4b3ioQ Q74Xqr5XJEV36 as expected, i.e. 13-char string with the salt as the first two. h~ **= Email 8 ==========================** Date: Mon, 28 Jan 2002 11:55:23 +0100 (CET) From: "Adrian Gschwend" Subject: OT: stdout/stderror I have a question regarding stdout/stderror. There is a java application which explores networks, a lot of code relies on catching the ping-output which is called external in the guest-OS (there are no free ping-implementations in Java). The application does not work on OS/2, the stdout hook does not get any text as it looks like. Is there a way to check if the OS/2-ping does write to stdout? If it's really an error of the OS/2 ping I would have to try another ping-port. cu Adrian -- Adrian Gschwend at OS/2 Netlabs ICQ: 22419590 ktk at netlabs.org ------- The OS/2 OpenSource Project: http://www.netlabs.org **= Email 9 ==========================** Date: Mon, 28 Jan 2002 12:51:28 -0500 From: Henry Sobotka Subject: Re: Perl enhancements Lyn St George wrote: > > Socket.obj(Socket.obj) : error L2029: 'inet_aton' : unresolved external You probably have d_inetaton set to 'define' in config.sh and offspring. It's not in EMX libsocket. > And, to make it worse, Perl while building complains that it can't > find crypt() (though it finds -lcrypt). Do you have d_crypt set to 'define'? h~ **= Email 10 ==========================** Date: Mon, 28 Jan 2002 12:54:44 GMT From: Michael Warmuth Subject: Re: OT: stdout/stderror On 28.1.02, 10:55:23, "Adrian Gschwend" wrote regarding " OT: stdout/stderror": [...] > Is there a way to check if the OS/2-ping does write to stdout? [...] ping localhost 56 3 1> ping.stdout 2> ping.stderr Content of file ping.stdout: --------------------------------------------------------- PING localhost.warmuth.at: 56 data bytes 64 bytes from 127.0.0.1: icmp_seq=0. time=0. ms 64 bytes from 127.0.0.1: icmp_seq=1. time=0. ms 64 bytes from 127.0.0.1: icmp_seq=2. time=0. ms ----localhost.warmuth.at PING Statistics---- 3 packets transmitted, 3 packets received, 0% packet loss round-trip (ms) min/avg/max = 0/0/0 --------------------------------------------------------- Content of file ping.stderr: --------------------------------------------------------- (empty) --------------------------------------------------------- So ping does write to stdout. Hope this helps. Greetings Michael -- Michael Warmuth Michael Warmuth - EDV-Dienstleistungen Austria - The place in the heart of Europe www.warmuth.at where no kangaroos are hopping around **= Email 11 ==========================** Date: Mon, 28 Jan 2002 12:56:16 +0100 (CET) From: "Adrian Gschwend" Subject: Re: OT: stdout/stderror On Mon, 28 Jan 2002 11:29:47 +0000, John Poltorak wrote: >Does this answer your question? :- hmm looks good, in the meanwhile I found out that Java itself has the bug with this. Should be fixed in the latest Java 1.3.0 release but didn't work before cu Adrian -- Adrian Gschwend at OS/2 Netlabs ICQ: 22419590 ktk at netlabs.org ------- The OS/2 OpenSource Project: http://www.netlabs.org **= Email 12 ==========================** Date: Mon, 28 Jan 2002 13:20:21 +0100 From: "Achim Hasenmueller" Subject: Re: OT: stdout/stderror Hi, you can easily determine which handle is being written to: ping > stdout.txt 2> stderr.out Check in which file the output is. And as I already told you before, Java 1.1.8 has a known bug with this and so do earlier releases of Java 1.3.0 for OS/2. Kind regards / mit freundlichen Gruessen, Achim Hasenmueller InnoTek Systemberatung GmbH phon: +49 7151 9965-30 achimha at innotek.de http://www.innotek.de Germany InnoTek at the CeBIT fair (March 13-20, 2002, Hannover, Germany): Hall 4, booth A04 (IBM Partner booth), demo point A12 "Adrian Gschwend" at eyup.org on 28.01.2002 11.55.23 Please respond to os2-unix at eyup.org Sent by: owner-os2-unix at eyup.org To: "UnixOS2" cc: Subject: OT: stdout/stderror I have a question regarding stdout/stderror. There is a java application which explores networks, a lot of code relies on catching the ping-output which is called external in the guest-OS (there are no free ping-implementations in Java). The application does not work on OS/2, the stdout hook does not get any text as it looks like. Is there a way to check if the OS/2-ping does write to stdout? If it's really an error of the OS/2 ping I would have to try another ping-port. cu Adrian -- Adrian Gschwend at OS/2 Netlabs ICQ: 22419590 ktk at netlabs.org ------- The OS/2 OpenSource Project: http://www.netlabs.org **= Email 13 ==========================** Date: Mon, 28 Jan 2002 13:29:03 -0500 From: Henry Sobotka Subject: Re: Make 3.79.1 Andreas Buening wrote: > > Thanks. I've download and compiled that package but I wonder > whether it is expected to work (it doesn't). There seems to be > no info who made this package and what its status is. > Have I missed something? That's my zip, with the line-endings converted to NL for Netlabs CVS. There's also a CRLF version floating around. It ran here but still needed a fair amount of work, mainly on the SHELL game. If you look at the archives for this mailing list from around the time it got uploaded to Netlabs, you should find a to-do list. h~ **= Email 14 ==========================** Date: Mon, 28 Jan 2002 13:56:05 -0600 From: email at eracc.hypermart.net Subject: An archive for list? I was wondering if there is an archive of all the messages sent to this list? If so, where? If not then what do we need to do to get such a beast? It's important to be able to go back and check for old subjects when possible. Especially when one has missed e-mail due to an ISP error or other event. Gene -- +=========================-=>Unix & OS/2<=-=========================+ # Owner and C.E.O. - ERA Computer Consulting - Jackson, TN USA # # OS/2, UnixWare, OpenServer & Linux Business Computing Solutions # # Please visit our www pages at http://eracc.hypermart.net/ # +===================================================================+ We run IBM OS/2 v.4.00, Revision 9.036 Sysinfo: 38 Processes, 146 Threads, uptime is 1d 16h 17m 32s 185ms **= Email 15 ==========================** Date: Mon, 28 Jan 2002 14:27:41 +0000 From: John Poltorak Subject: Perl enhancements IMV we should include the following as standard in os2ish.h before building Perl:- #define HAS_PASSWD #define HAS_GETPWENT #define HAS_CRYPT Is there any reason not to do so? -- John **= Email 16 ==========================** Date: Mon, 28 Jan 2002 14:27:41 +0000 From: John Poltorak Subject: Perl enhancements IMV we should include the following as standard in os2ish.h before building Perl:- #define HAS_PASSWD #define HAS_GETPWENT #define HAS_CRYPT Is there any reason not to do so? -- John **= Email 17 ==========================** Date: Mon, 28 Jan 2002 14:27:41 +0000 From: John Poltorak Subject: Perl enhancements IMV we should include the following as standard in os2ish.h before building Perl:- #define HAS_PASSWD #define HAS_GETPWENT #define HAS_CRYPT Is there any reason not to do so? -- John **= Email 18 ==========================** Date: Mon, 28 Jan 2002 14:27:41 +0000 From: John Poltorak Subject: Perl enhancements IMV we should include the following as standard in os2ish.h before building Perl:- #define HAS_PASSWD #define HAS_GETPWENT #define HAS_CRYPT Is there any reason not to do so? -- John **= Email 19 ==========================** Date: Mon, 28 Jan 2002 14:27:41 +0000 From: John Poltorak Subject: Perl enhancements IMV we should include the following as standard in os2ish.h before building Perl:- #define HAS_PASSWD #define HAS_GETPWENT #define HAS_CRYPT Is there any reason not to do so? -- John **= Email 20 ==========================** Date: Mon, 28 Jan 2002 15:24:51 +0000 From: John Poltorak Subject: Re: Building Perl 5.6.1 from source On Mon, Jan 28, 2002 at 03:34:38PM +0100, Stefan Neis wrote: > On Mon, 28 Jan 2002, John Poltorak wrote: > > > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > > to match it (otherwise "Checking to see if your kit is complete" > > > will fail). > > > > > > OK this works, although I did not need to do anything like this with > > 5.6.0. Wonder what has changed to cause this problem... > > It's the typical gnumake-3.75 bug. If you used a different version > the last time you installed it, then this is the reason... In this case I rebuilt 5.6.0 just to see if exactly the same script works and it does, so Make is not a factor here. One thing which does differ is the install target in the Makefile... 5.6.1 has:- install: all install.perl install.man 5.6.0 has:- install: $(MAKE) install.perl install.man STRIPFLAGS=$(STRIPFLAGS) I suspect it's 'all' causing the problem... > Regards, > Stefan > -- > Micro$oft is not an answer. It is a question. The answer is 'no'. > -- John **= Email 21 ==========================** Date: Mon, 28 Jan 2002 15:25:40 -0500 From: Henry Sobotka Subject: Re: Make 3.79.1 Andreas Buening wrote: > > Did you also write the OS/2 specific code or is it more copy&paste > from older make implementations, i.e. is it possible to put that > code into the "official" GNU make? The people at gnu.org really > insist on that copyright issue. John's mentioning "last March" made it easy to find the note that went with the zip. It also answers your question. h~ --------------- Attached. Most are a carryover of the 3.76.1 patches. The relatively numerous test failures boil down to eight problems (in no particular order): 1) calls to sleep complaining about an "invalid time interval"; 2) "/c: /c: No such file or directory"; 3) "target patterns contains no '%'; 4) escaped colons so we get p\:foo instead of p:foo; 5) missing single quotes; 6) extra blanks; 7) error # (255 instead of 1); 8) $(MAKE) expanding to path (minus drive-letter) instead of just "make". I also found that _not_ defining HAVE_CASEINSENSITIVE_FS in config.h (a macro which now covers some of the EMX/DOS blocks in the old patches) eliminates the annoying "s.o. override" warnings prevalent with most current makefiles and so far hasn't produced any negative side-effects. I think we're better off pretending we have a case-sensitive filesystem, because case is at least preserved, so that .s|.S and .c|.C rules are properly differentiated. There's also a block in variable.c which we share with DOS in 3.76.1, but if we use it now, results in a failure that warrants a closer look. It's XXXX'd. Also not happy with the kludge for turning /c on and off in job.c, and that may be the origin of 2) above. **= Email 22 ==========================** Date: Mon, 28 Jan 2002 15:34:38 +0100 (CET) From: Stefan Neis Subject: Re: Building Perl 5.6.1 from source On Mon, 28 Jan 2002, John Poltorak wrote: > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > to match it (otherwise "Checking to see if your kit is complete" > > will fail). > > > OK this works, although I did not need to do anything like this with > 5.6.0. Wonder what has changed to cause this problem... It's the typical gnumake-3.75 bug. If you used a different version the last time you installed it, then this is the reason... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 23 ==========================** Date: Mon, 28 Jan 2002 15:34:38 +0100 (CET) From: Stefan Neis Subject: Re: Building Perl 5.6.1 from source On Mon, 28 Jan 2002, John Poltorak wrote: > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > to match it (otherwise "Checking to see if your kit is complete" > > will fail). > > > OK this works, although I did not need to do anything like this with > 5.6.0. Wonder what has changed to cause this problem... It's the typical gnumake-3.75 bug. If you used a different version the last time you installed it, then this is the reason... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 24 ==========================** Date: Mon, 28 Jan 2002 15:34:38 +0100 (CET) From: Stefan Neis Subject: Re: Building Perl 5.6.1 from source On Mon, 28 Jan 2002, John Poltorak wrote: > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > to match it (otherwise "Checking to see if your kit is complete" > > will fail). > > > OK this works, although I did not need to do anything like this with > 5.6.0. Wonder what has changed to cause this problem... It's the typical gnumake-3.75 bug. If you used a different version the last time you installed it, then this is the reason... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 25 ==========================** Date: Mon, 28 Jan 2002 15:34:38 +0100 (CET) From: Stefan Neis Subject: Re: Building Perl 5.6.1 from source On Mon, 28 Jan 2002, John Poltorak wrote: > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > to match it (otherwise "Checking to see if your kit is complete" > > will fail). > > > OK this works, although I did not need to do anything like this with > 5.6.0. Wonder what has changed to cause this problem... It's the typical gnumake-3.75 bug. If you used a different version the last time you installed it, then this is the reason... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 26 ==========================** Date: Mon, 28 Jan 2002 15:34:38 +0100 (CET) From: Stefan Neis Subject: Re: Building Perl 5.6.1 from source On Mon, 28 Jan 2002, John Poltorak wrote: > > What I've done: rename INSTALL to INSTALL1, and edit Manifest > > to match it (otherwise "Checking to see if your kit is complete" > > will fail). > > > OK this works, although I did not need to do anything like this with > 5.6.0. Wonder what has changed to cause this problem... It's the typical gnumake-3.75 bug. If you used a different version the last time you installed it, then this is the reason... Regards, Stefan -- Micro$oft is not an answer. It is a question. The answer is 'no'. **= Email 27 ==========================** Date: Mon, 28 Jan 2002 17:19:47 +0000 From: "Lyn St George" Subject: Re: Perl enhancements On Mon, 28 Jan 2002 10:02:23 -0500, Henry Sobotka wrote: >John Poltorak wrote: >> >> IMV we should include the following as standard in os2ish.h before >> building Perl:- >> >> #define HAS_PASSWD >> #define HAS_GETPWENT >> #define HAS_CRYPT >> >> Is there any reason not to do so? > >Crypt works, but as long as we only have fake password support in EMX, I >think it's wiser to leave the other two turned off. One of the main uses >of Perl is with CGI and Apache, and pretending to provide security >features that aren't really there is not a risk worth exposing users to. > >h~ > Hmm ... I've just tried to build Perl again so as to answer John P's previous question, and discovered that it won't build any more - I'm now getting this error: -------------- Socket.obj(Socket.obj) : error L2029: 'inet_aton' : unresolved external There was 1 error detected make[1]: *** [../../lib/auto/Socket/SocketDC.dll] Error 1 make[1]: Leaving directory `/dev/perl/perl-5.6.1/ext/Socket' make: *** [lib/auto/Socket/Socket.dll] Error 2 ---------------- I've just found an earlier set of notes that also had this same error, but no clue as to how to fix it. My later notes just say: -------- remade with extra #define's: getgrgid, getgrnam, getgrent, crypt, passwd, getpwent Added 'struct group *getgrent (void);' to /emx/include/grp.h Built crypt() from ufc-crypt, renamed as crypt.exe, crypt.h, crypt.lib, and added -lcrypt to libs flags. All built OK. --------- (I also see that I was doing this over Christmas/NY, with a very befuddled head - stupid bastard that I am) Now I've lost that build, and really can't remember what needs to be done to fix this :( And, to make it worse, Perl while building complains that it can't find crypt() (though it finds -lcrypt). Would you have any ideas at all on this ??? TIA Cheers Lyn St George +--------------------------------------------------------------------------------- + http://www.zolotek.net .. eCommerce hosting, consulting + http://www.os2docs.org .. some 'How To' stuff ... +---------------------------------------------------------------------------------- **= Email 28 ==========================** Date: Mon, 28 Jan 2002 19:01:13 +0100 From: Andreas Buening Subject: Re: Building Perl 5.6.1 from source John Poltorak wrote: > > On Mon, Jan 28, 2002 at 03:34:38PM +0100, Stefan Neis wrote: [gnumake] > > It's the typical gnumake-3.75 bug. If you used a different version > > the last time you installed it, then this is the reason... > > In this case I rebuilt 5.6.0 just to see if exactly the same script works > and it does, so Make is not a factor here. It definitely _is_ a bug of most make implementations. I must know, I removed it to get Makefiles working. ;-) That code below seems to be a workaround. [perl 5.6.0] bye, Andreas **= Email 29 ==========================** Date: Mon, 28 Jan 2002 19:11:19 +0100 From: Andreas Buening Subject: Re: Make 3.79.1 IanM wrote: > > Hi John > > http://www.os2site.com/sw/unixport/misc/make-3.79.1-emx-nl.zip Thanks. I've download and compiled that package but I wonder whether it is expected to work (it doesn't). There seems to be no info who made this package and what its status is. Have I missed something? bye, Andreas **= Email 30 ==========================** Date: Mon, 28 Jan 2002 19:41:47 +0000 From: John Poltorak Subject: Re: Make 3.79.1 On Mon, Jan 28, 2002 at 01:29:03PM -0500, Henry Sobotka wrote: > Andreas Buening wrote: > > > > Thanks. I've download and compiled that package but I wonder > > whether it is expected to work (it doesn't). There seems to be > > no info who made this package and what its status is. > > Have I missed something? > > That's my zip, with the line-endings converted to NL for Netlabs CVS. > There's also a CRLF version floating around. It ran here but still > needed a fair amount of work, mainly on the SHELL game. If you look at > the archives for this mailing list from around the time it got uploaded > to Netlabs, you should find a to-do list. Unfortunately the Web enabled archives are not uptodate - I keep meaning to get it sorted but... Anyhow I had a look through and now recall spending much of last March trying to get Make 3.79.1 working without success, but many of the things which caused problems have now been updated, so it may be safe to try again... A list of problems at the time included:- 1) calls to sleep complaining about an "invalid time interval"; 2) "/c: /c: No such file or directory"; 3) "target patterns contains no '%'; 4) escaped colons so we get p\:foo instead of p:foo; 5) missing single quotes; 6) extra blanks; 7) error # (255 instead of 1); 8) $(MAKE) expanding to path (minus drive-letter) instead of just "make". Also there were a number of problems which were isolated to some routine in job.c which didn't handle shells correctly. > h~ -- John **= Email 31 ==========================** Date: Mon, 28 Jan 2002 20:17:09 +0100 From: Andreas Buening Subject: Re: Make 3.79.1 Henry Sobotka wrote: > > Andreas Buening wrote: > > > > Thanks. I've download and compiled that package but I wonder > > whether it is expected to work (it doesn't). There seems to be > > no info who made this package and what its status is. > > Have I missed something? > > That's my zip, with the line-endings converted to NL for Netlabs CVS. > There's also a CRLF version floating around. It ran here but still > needed a fair amount of work, mainly on the SHELL game. If you look at > the archives for this mailing list from around the time it got uploaded > to Netlabs, you should find a to-do list. Did you also write the OS/2 specific code or is it more copy&paste from older make implementations, i.e. is it possible to put that code into the "official" GNU make? The people at gnu.org really insist on that copyright issue. bye, Andreas **= Email 32 ==========================** Date: Mon, 28 Jan 2002 20:29:29 +0000 From: John Poltorak Subject: Re: An archive for list? On Mon, Jan 28, 2002 at 01:56:05PM -0600, email at eracc.hypermart.net wrote: > I was wondering if there is an archive of all the messages sent to > this list? If so, where? If not then what do we need to do to get > such a beast? It's important to be able to go back and check for old > subjects when possible. Especially when one has missed e-mail due to > an ISP error or other event. There is an incomplete archive here:- http://www.eyup.org I really need to get it updated, but keep hoping to get up and running with MailMan any time soon, which, AFAIK maintains a web archive automatically. I'm using Mhonarc which formats the msgs nicely enough, but can't find anything which creates a monthly index. Without that it makes painful reading when there are 10,000 msgs to display. It would be nice to get a search engine in place too, but didn't have much success with Webglimpse when I last tried. > Gene > -- > +=========================-=>Unix & OS/2<=-=========================+ > # Owner and C.E.O. - ERA Computer Consulting - Jackson, TN USA # > # OS/2, UnixWare, OpenServer & Linux Business Computing Solutions # > # Please visit our www pages at http://eracc.hypermart.net/ # > +===================================================================+ > We run IBM OS/2 v.4.00, Revision 9.036 > Sysinfo: 38 Processes, 146 Threads, uptime is 1d 16h 17m 32s 185ms > > -- John **= Email 33 ==========================** Date: Mon, 28 Jan 2002 21:25:20 +0100 From: Andreas Buening Subject: Re: Make 3.79.1 John Poltorak wrote: [make 3.79.1] > Anyhow I had a look through and now recall spending much of last March > trying to get Make 3.79.1 working without success, but many of the things > which caused problems have now been updated, so it may be safe to try > again... > > A list of problems at the time included:- > > 1) calls to sleep complaining about an "invalid time interval"; > 2) "/c: /c: No such file or directory"; > 3) "target patterns contains no '%'; > 4) escaped colons so we get p\:foo instead of p:foo; > 5) missing single quotes; > 6) extra blanks; > 7) error # (255 instead of 1); > 8) $(MAKE) expanding to path (minus drive-letter) instead of just "make". > > Also there were a number of problems which were isolated to some routine > in job.c which didn't handle shells correctly. That's good. Then I'm not the only one. ;-) bye, Andreas **= Email 34 ==========================** Date: Mon, 28 Jan 2002 22:03:51 +0100 From: Holger Veit Subject: Re: An archive for list? On Mon, Jan 28, 2002 at 08:29:29PM +0000, John Poltorak wrote: > On Mon, Jan 28, 2002 at 01:56:05PM -0600, email at eracc.hypermart.net wrote: > > I was wondering if there is an archive of all the messages sent to > > this list? If so, where? If not then what do we need to do to get > > such a beast? It's important to be able to go back and check for old > > subjects when possible. Especially when one has missed e-mail due to > > an ISP error or other event. > > There is an incomplete archive here:- > > http://www.eyup.org You might also look at http://borneo.gmd.de/~veit/os2/unixsearch.html Holger -- Please update your tables to my new e-mail address: holger.veit$ais.fhg.de (replace the '$' with ' at ' -- spam-protection)