Re: Pantherbites
[EMAIL PROTECTED] writes: What about cvs update -ko? Something with -ko was what I did to get perl into CVS at work.
[OT] backslash when working with Japanese (Re: Pantherbites)
* input method is a mess! I like the old one ("disintegrated") better. Yeah, I think the IM is going backwards as much as forwards, myself. * took 15 minutes to find how to enter '\' (backslash) instead of 」、 (yen) ONCE Kotoeri is enabled. Once Kotoeri is enabled, '\' key refuses to backslash and yens yens yens! Solution: Add "US Extended" via [Input Menu] tab of [International] Preference Panel and use it. I was able to get the backslash in the Project Builder editor in Jaguar using the option key on the yen key or the forward slash key, I don't remember which and I didn't bring my iBook to work today. Do any of those key combinations work? (I'll try to remember to check that tonight.) -- Joel Rees, programmer, Kansai Systems Group Altech Corporation (Alpsgiken), Osaka, Japan http://www.alpsgiken.co.jp
Re: [OT] backslash when working with Japanese (Re: Pantherbites)
On Friday, July 25, 2003, at 03:44 PM, Joel Rees wrote: I was able to get the backslash in the Project Builder editor in Jaguar using the option key on the yen key or the forward slash key I'll vouch for the [option] + [yen] combo. FWIW [ option] + [ / ] yields a mathematical division sign Robin
Re: Pantherbites
On Thursday, July 24, 2003, at 11:30 PM, Sherm Pendley wrote: On Thursday, July 24, 2003, at 3:07 PM, Edward Moy wrote: A 5.8.1 version of Perl is in Panther, but whether *the* 5.8.1 final release will make Panther is another question. I hope, at least, that the final Panther will include a newer Perl than the one in the preview. I spent most of today trying to track down a SIG11 bug in CamelBones. Out of desperation, I installed the newer 5.8.1-RC2 release of Perl under /opt. It's compiled with GCC 3.3, but otherwise configured identically to the 5.8.1 snapshot that's included with the Panther preview. With the 5.8.1-RC2 release, the SIG11 problems vanished and all the test CB cases worked perfectly. Having cleared this response with my boss, I can say while I can't promise on product releases, we have every intention of shipping the latest version of Perl we can for Panther, that currently being RC2 (which we have tested in-house). I'll grant you that CB embeds Perl and makes heavy use of the libperl API; it probably tickles bugs in Perl that most people will never see. Still, the experience hasn't left me with much confidence in the Perl version that's included in the Panther preview. The 5.8.1-RC2 release, on the other paw, is the first release I can recall that passed every single self-test on Mac OS X - even the long-standing libdb issue seems to have been fixed. This is not actually due to Perl, but an updated Berkeley DB implementation (1.85) in Panther that passes the test, but still remain file-format compatible. --- Edward Moy Apple
5.8.1 differences (was Re: Pantherbites)
[EMAIL PROTECTED] wrote: Sorry for the probably silly question, but just what are the differences between 5.6.1 and 5.8.1? I am 100% ig'nant, but I assume that most/all 5.6.1 scripts will port fairly hands-off? What are some gotchas that us noobs should watch for to make sure our stuff works when we try to run it on Panther? Two things that occur to me right away to watch out for (I'm sure other people will have other comments): 1) There are a large number of perl modules which ship with 5.8.1 but did not ship with 5.6.1. For the most part, this is a good thing, but you'll need to watch the version numbers of the modules if the version number is important to your code. 2) Any XS modules which were compiled against perl 5.6.1 will break. This also applies to any executable or library which you may have which links against the main perl library /System/Library/Perl/darwin/CORE/libperl.dylib HTH, Dave
Re: Pantherbites
On Thursday, July 24, 2003, at 11:55 AM, Tim Jenness wrote: On Wed, 23 Jul 2003 [EMAIL PROTECTED] wrote: On Wednesday, July 23, 2003, at 7:42 PM, Dan Kogai wrote: Apology accepted. Nevertheless, consider this problem fixed for final Panther. I am 80% proud and 20% scared to see my codes become a part of the most popular *NIX available today. Keep going! Dan the (Encode Maintainer|Panther-user-to-be) Let me welcome you. Apologies if people know this already, but when does 5.8.1 have to be out for real in order for it to be integrated into Panther? Is it too late already? A 5.8.1 version of Perl is in Panther, but whether *the* 5.8.1 final release will make Panther is another question. All I can say is that Python is in a similar boat, and they are shooting for early August. --- Edward Moy Apple
Re: Pantherbites
On Wed, 23 Jul 2003 [EMAIL PROTECTED] wrote: On Wednesday, July 23, 2003, at 7:42 PM, Dan Kogai wrote: Apology accepted. Nevertheless, consider this problem fixed for final Panther. I am 80% proud and 20% scared to see my codes become a part of the most popular *NIX available today. Keep going! Dan the (Encode Maintainer|Panther-user-to-be) Let me welcome you. Apologies if people know this already, but when does 5.8.1 have to be out for real in order for it to be integrated into Panther? Is it too late already? -- Tim Jenness JAC software http://www.jach.hawaii.edu/~timj
Re: Pantherbites
Hmm Mail 1.3 seems to need some more work (back in Jaguar) On Wednesday, Jul 23, 2003, at 21:32 Asia/Tokyo, Dan Kogai wrote: I have finally gotten an access to Panther so here is my preliminary report on it as a Perl5 Porter. Let me begin with pros * It is good looking * And fast * I love Expos~{(~}. I meant Expos or qq(Expos\x{00e9}). * took 15 minutes to find how to enter '\' (backslash) instead of ~{#$~} (yen) ONCE Kotoeri is And (FULLWIDTH YEN: \x{ffe5}. Content-Type: text/plain; charset=HZ-GB-2312; format=flowed Content-Transfer-Encoding: 7bit Mail 1.3 was obviously out of mind! Dan the Perl5 Porter
Re: Pantherbites
Arthur Bergman [EMAIL PROTECTED] wrote: On Wednesday, July 23, 2003, at 01:32 pm, Dan Kogai wrote: Aha! Now I see. emoy has checked in MAINT19524 source and RCS retagged $Revision$. I'm sorry emoy but this is NOT THE RIGHT THING (TM). Perl module infrastructure depends heavily upon version numbers and auto-versioning via RCS/CVS as above is a common technique in perl. I surely hope this will be fixed in real Panther... Yes, this is a nightmare for anyone tracks perl releases using cvs vendor function. The solution is to artificially increase the version number, but it is a pain. What about cvs update -ko? Regards, Slaven
Re: Pantherbites
On Wednesday, Jul 23, 2003, at 05:32 US/Pacific, Dan Kogai wrote: Yes. As announced earlier Panther Preview comes w/ Perl 5.8.1-tobe (MAINT19524 to be exact) so I hope we will make Perl 5.8.1 a reality before Panther is. Sorry for the probably silly question, but just what are the differences between 5.6.1 and 5.8.1? I am 100% ig'nant, but I assume that most/all 5.6.1 scripts will port fairly hands-off? What are some gotchas that us noobs should watch for to make sure our stuff works when we try to run it on Panther? I'm sure that there are ample docs on this matter, but a pointer in their direction would be most helpful... - Daniel C. Stillwaggon ([EMAIL PROTECTED])
Re: Pantherbites
On Wednesday, July 23, 2003, at 01:32 pm, Dan Kogai wrote: Aha! Now I see. emoy has checked in MAINT19524 source and RCS retagged $Revision$. I'm sorry emoy but this is NOT THE RIGHT THING (TM). Perl module infrastructure depends heavily upon version numbers and auto-versioning via RCS/CVS as above is a common technique in perl. I surely hope this will be fixed in real Panther... Yes, this is a nightmare for anyone tracks perl releases using cvs vendor function. The solution is to artificially increase the version number, but it is a pain. Arthur
Re: Pantherbites
AB == Arthur Bergman [EMAIL PROTECTED] writes: AB On Wednesday, July 23, 2003, at 01:32 pm, Dan Kogai wrote: Aha! Now I see. emoy has checked in MAINT19524 source and RCS retagged $Revision$. I'm sorry emoy but this is NOT THE RIGHT THING (TM). Perl module infrastructure depends heavily upon version numbers and auto-versioning via RCS/CVS as above is a common technique in perl. I surely hope this will be fixed in real Panther... AB Yes, this is a nightmare for anyone tracks perl releases using cvs AB vendor function. AB The solution is to artificially increase the version number, but it is AB a pain. Better solution is to use FreeBSD's patches for CVS to support custom revision tags. This solves this problem nicely by allowing several maintainers to have their own revision tags in same source code. For examples snippet from /usr/src/crypto/openssh/ssh-agent.c RCSID($OpenBSD: ssh-agent.c,v 1.97 2002/06/24 14:55:38 markus Exp $); RCSID($FreeBSD: src/crypto/openssh/ssh-agent.c,v 1.2.2.8 2002/07/03 22:11:43 des Exp $); -- Ilya Martynov, [EMAIL PROTECTED] CTO IPonWEB (UK) Ltd Quality Perl Programming and Unix Support UK managed @ offshore prices - http://www.iponweb.net Personal website - http://martynov.org
Re: Pantherbites
On Wednesday, July 23, 2003, at 5:32AM, Dan Kogai wrote: # # $Id: Encode.pm,v 1.4 2003/05/20 22:49:15 emoy Exp $ # package Encode; use strict; our $VERSION = do { my @r = (q$Revision: 1.4 $ =~ /\d+/g); sprintf %d..%02d x $#r, @r }; Aha! Now I see. emoy has checked in MAINT19524 source and RCS retagged $Revision$. I'm sorry emoy but this is NOT THE RIGHT THING (TM). Perl module infrastructure depends heavily upon version numbers and auto-versioning via RCS/CVS as above is a common technique in perl. I surely hope this will be fixed in real Panther... Did you file a bug report? Don't worry, I already did (#3340036). Do you think I did this on purpose? I hope not. Do you think this has never happened before? No, Perl 5.6.0 on Jaguar and previous had the same problem with CVS (no -ko option) and only now was it ever noticed. And in the several months since I inherited Perl (actually asked for it), I would have fixed it if I had known. Sorry, I don't usually get personal, but tone of this message seemed inappropriate to me. Nevertheless, consider this problem fixed for final Panther. Edward Moy Apple
Re: Pantherbites
In article [EMAIL PROTECTED], [EMAIL PROTECTED] (Edward Moy) wrote: Do you think this has never happened before? No, Perl 5.6.0 on Jaguar and previous had the same problem with CVS (no -ko option) and only now was it ever noticed. I noticed! :-) Shortly after Fred Sanchez first released his perl 5.6 stuff for Mac OS X, I commented on it. Don't remember to whom, though. I had the same problem with MacPerl, which I keep in a separate CVS repository. Sorry, I don't usually get personal, but tone of this message seemed inappropriate to me. I don't think it was intended to be personal or mean-spirited in any way, for what it's worth. Nevertheless, consider this problem fixed for final Panther. Thanks! BTW, while I've got your eyes, should it be possible to compile perl and its extensions with prebinding? I know .bundles cannot be prebound, but I have successfully built bundles as .dylibs (I think that's what I did, anyway, it's been a few months), and I wonder if this is something that might be useful. My main problem is Mac::Carbon takes a LONG time to load (a few seconds), and almost all of that is in dyld. Panther's perl with gcc 3.3 has the same problem. I am going to try a bunch of things, but figure you might have some insight before I go and try things that won't work. Thanks, -- Chris Nandor [EMAIL PROTECTED]http://pudge.net/ Open Source Development Network[EMAIL PROTECTED] http://osdn.com/
Re: Pantherbites
On Wednesday, July 23, 2003, at 07:32 AM, Dan Kogai wrote: Anyway, As an Encode Maintainer I naturally typed perl -MEncode -le 'print Encode-VERSION' and here is what Panther said 1.04 Ohmygod! Even Perl 5.8.0 comes with Encode ver. 1.75. WHAT HAS HAPPEND!? So I 'perldoc -m Encode' and found this. # # $Id: Encode.pm,v 1.4 2003/05/20 22:49:15 emoy Exp $ # package Encode; use strict; our $VERSION = do { my @r = (q$Revision: 1.4 $ =~~ /\d+/g); sprintf %d..%02d x $#r, @r }; Really, it's a bad idea to use a CVS revision as $VERSION. For one thing, it can cause the problem you found when other people integrate the sources into other projects. For another thing, it means you can't make frequent commits to CVS unless you want to skip several version numbers with each release. This is why the CVS manuals recommend against using %Revision% to generate release versions: http://www.loria.fr/~molli/cvs/doc/cvs_4.html I used to do things the same way you do, but life's gotten much better now that I just increment $VERSION manually. =) -Ken
Re: Pantherbites
On Thursday, Jul 24, 2003, at 01:18 Asia/Tokyo, Edward Moy wrote: Did you file a bug report? Don't worry, I already did (#3340036). I was about to do but I am a Perl5 Porter before OS X user so @perl.org had precedence. Plus it is in the middle of the change of the season (rainy - summer) and I always get cold during then. Cold won before I file the report. Thanks for moving so quickly. Do you think I did this on purpose? I hope not. Not at all. If I were you I'd have fallen in the same pit. And I am rather glad we have found this problem before Panther matures to adulthood. Do you think this has never happened before? No, Perl 5.6.0 on Jaguar and previous had the same problem with CVS (no -ko option) and only now was it ever noticed. And in the several months since I inherited Perl (actually asked for it), I would have fixed it if I had known. That was the conclusion I have reached; Perl 5.6.0 has far less modules and most modules hardcoded $VERSION so it went unnoticed. I think it is a good timing to introduce what Ilya has mentioned. On Wednesday, Jul 23, 2003, at 23:34 Asia/Tokyo, Ilya Martynov wrote: Better solution is to use FreeBSD's patches for CVS to support custom revision tags. This solves this problem nicely by allowing several maintainers to have their own revision tags in same source code. For examples snippet from /usr/src/crypto/openssh/ssh-agent.c RCSID($OpenBSD: ssh-agent.c,v 1.97 2002/06/24 14:55:38 markus Exp $); RCSID($FreeBSD: src/crypto/openssh/ssh-agent.c,v 1.2.2.8 2002/07/03 22:11:43 des Exp $); I would love to see the $Darwin$ tags! Back to emoy Sorry, I don't usually get personal, but tone of this message seemed inappropriate to me. And here is my apology for the tone of my voice. Well, I have cold to blame (my nasal cavity is so stuffy that I should have literally stood in a way of frontal lobe of my cerebrum). Nevertheless, consider this problem fixed for final Panther. I am 80% proud and 20% scared to see my codes become a part of the most popular *NIX available today. Keep going! Dan the (Encode Maintainer|Panther-user-to-be)
Re: Pantherbites
On Wednesday, July 23, 2003, at 7:42 PM, Dan Kogai wrote: On Thursday, Jul 24, 2003, at 01:18 Asia/Tokyo, Edward Moy wrote: Sorry, I don't usually get personal, but tone of this message seemed inappropriate to me. And here is my apology for the tone of my voice. Well, I have cold to blame (my nasal cavity is so stuffy that I should have literally stood in a way of frontal lobe of my cerebrum). Apology accepted. Nevertheless, consider this problem fixed for final Panther. I am 80% proud and 20% scared to see my codes become a part of the most popular *NIX available today. Keep going! Dan the (Encode Maintainer|Panther-user-to-be) Let me welcome you. -- Edward Moy Apple