Blead Breaks CPAN: CFAERBER/Net-IDN-Encode-2.500.tar.gz #16850
Comments
From @eserteThis is a bug report for perl from slaven@rezic.de, Starting with perl 5.29.8 the test suite of Net-IDN-Encode-2.500 Test reports will appear in an hour or so at CPAN Testers. Flags: This perlbug was built using Perl 5.20.2 - Mon Sep 18 18:13:32 UTC 2017 Site configuration information for perl 5.29.8: Configured by eserte at Thu Feb 21 18:36:21 CET 2019. Summary of my perl5 (revision 5 version 29 subversion 8) configuration: @INC for perl 5.29.8: Environment for perl 5.29.8: |
From @eserteDana Thu, 21 Feb 2019 12:53:01 -0800, slaven@rezic.de reče:
Two sample fail reports:
|
From @jkeenanOn Thu, 21 Feb 2019 20:53:01 GMT, slaven@rezic.de wrote:
Bisection points to this commit: ##### Move \p{user-defined} to core from utf8_heavy.pl This large commit moves the handling of user-defined properties to C code. This should speed it up, but the main reason to do this is to stop using swashes in this case, leaving only tr/// using them. Once that too is converted, all swash handling can be ripped out of perl. Doing this in perl has caused some nasty interactions that will now be fixed automatically. The change is not entirely transparent, however (besides speed and the possibility of removing these interactions). perldelta in this commit details these. Thank you very much. -- |
|
The RT System itself - Status changed from 'new' to 'open' |
From @jkeenanOn Thu, 21 Feb 2019 22:54:00 GMT, jkeenan wrote:
commit 73b9584
-- |
From @khwilliamsonOn 2/21/19 3:54 PM, James E Keenan via RT wrote:
I have found the problem, but don't know what to do about it. The module turns out to be relying on a bug in the earlier core. In this module, the function IsIgnored() is somehow being called from a The reason it used to work is that a bug in the core caused the first What can be done with this module is to remove the BEGIN block call, or But how much other code out there also relies on this former bug? I |
From @jkeenanOn Thu, 21 Feb 2019 22:58:10 GMT, jkeenan wrote:
I also encountered this problem on FreeBSD-11.1 while doing monthly "CPAN River 3000" testing. Attaching output. -- |
From @khwilliamsonOn Fri, 22 Feb 2019 12:09:22 -0800, public@khwilliamson.com wrote:
I have created a PR for this module. If you would like to comment on it, see |
From @andkAlso affected: BDFOY/Roman-Unicode-1.031.tar.gz |
From @jkeenanOn Sat, 30 Mar 2019 21:15:56 GMT, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
-- |
From [Unknown Contact. See original ticket]On Sat, 30 Mar 2019 21:15:56 GMT, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
-- |
From @khwilliamsonOn Sat, 30 Mar 2019 17:56:15 -0700, jkeenan wrote:
Because the actual trajectory of these two issues with the same commit are so divergent, I created a new ticket to track the Roman::Unicode problem. https://rt-archive.perl.org/perl5/Ticket/Display.html?id=133970 |
From @jkeenanOn Mon, 01 Apr 2019 04:56:35 GMT, khw wrote:
Does that mean that *this* RT (133860 -- dealing with Net-IDN-Encode) is closable? Thank you very much. -- |
From @khwilliamsonOn 4/1/19 4:54 PM, James E Keenan via RT wrote:
I submitted a PR, that means it isn't a blocker. But if a bunch of |
From @xsawyerxOn 4/2/19 3:04 AM, Karl Williamson wrote:
Any ideas of what we could do, other than reintroducing the original bug? |
From @khwilliamsonOn 4/6/2019 11:49 PM, Sawyer X wrote:
If a sub returned nothing, we could consider it still undefined, and |
From @eserteThis is a bug report for perl from slaven@rezic.de, The test suite of ANNO/Unicode-CharWidth-1.05.tar.gz fails Flags: Site configuration information for perl 5.29.9: Configured by eserte at Thu Mar 21 21:10:55 UTC 2019. Summary of my perl5 (revision 5 version 29 subversion 9) configuration: @INC for perl 5.29.9: Environment for perl 5.29.9: |
From @jkeenanOn Sun, 07 Apr 2019 19:31:16 GMT, slaven@rezic.de wrote:
With this bisection command: ##### ... I got this bisection result: ##### ... which is the same commit where breakage was reported in https://rt.perl.org/Public/Bug/Display.html?id=133860. So I recommend that this ticket be merged into that one.
-- |
|
The RT System itself - Status changed from 'new' to 'open' |
From @xsawyerxOn 4/7/19 6:18 PM, Karl Williamson wrote:
Similar to introducing the buggy behavior, but in a single function? Sounds like a temporary fix, though. This module will still need to be |
From @xsawyerxOn 4/8/19 2:46 AM, James E Keenan via RT wrote:
I was going to suggest the same. |
From @khwilliamsonOn 4/7/19 11:25 PM, Sawyer X wrote:
Not if we did what I said we could do. I'm not advocating it, and if |
From @khwilliamsonOn 4/7/19 11:29 PM, Sawyer X wrote:
There are at least two causes for this breakage, so merging is premature |
From @jkeenanOn 4/8/19 11:57 AM, Karl Williamson wrote:
Too late; I acted on sawyer's +1 about 4 hours ago. Suggestions? jimk |
From @khwilliamsonOn 4/8/2019 11:10 AM, James E Keenan wrote:
I don't think it's possible to unmerge. It will cause extra work for me |
From @khwilliamsonOn 4/8/19 12:02 PM, Karl Williamson wrote:
It turns out the root cause of this problem is completely different from And I'm not sure how best to fix this. I thought that I could use |
From @jkeenanOn Tue, 09 Apr 2019 07:45:35 GMT, public@khwilliamson.com wrote:
Well, what if I created a new ticket with most of the same content as in the original RT 134004 (that content now found at https://rt-archive.perl.org/perl5/Ticket/Display.html?id=133860#txn-1624393)? If I did so I, would like to include a note about why we're deviating from our normal practice of merging all BBC tickets associated with a single commit into the first such ticket created. If you were to provide me with a couple sentences for that reasoning, I would include that in the RT's first post.
Thank you very much. -- |
From @khwilliamsonOn 4/9/19 5:09 AM, James E Keenan via RT wrote:
This large commit has multiple issues. [perl #133860] is to track modules that were relying on previous [perl #133970] is because the commit needed to do save and restore the This ticket is because the commit wrongly assumed that a fully qualified The fixes for these are completely independent, and doing any one will
|
From @dur-randirOn Tue, 09 Apr 2019 00:45:35 -0700, public@khwilliamson.com wrote:
CV* are not moved, but it's address may be reused, if the sub is freed. |
From @xsawyerxOn 4/9/19 10:45 AM, Karl Williamson wrote:
Sorry about that. I figured "same commit -- same breakage."
As Jim noted, creating a new ticket that points to the original way, |
From @dur-randirOn Tue, 09 Apr 2019 08:49:44 -0700, randir wrote:
As an afterthought, thread creation changes all pointers in a new thread, if it matters in this case. |
From @khwilliamsonThe #134004 portion of this merged ticket is fixed by commit ef80af0 * PATCH: [perl #134004] BBC breaks Unicode::CharWidth |
From @khwilliamsonOn Tue, 09 Apr 2019 07:16:02 -0700, public@khwilliamson.com wrote: I've removed this from the blockers list. The only distro that has showed up so far that relied on previous buggy behavior is the one in #133860, and a PR request has been issued for it. |
From @ppisarCreated by @ppisarI noticed Net-IDN-Encode tests fail with Perl 5.30-RC1 This program: #!/usr/bin/perl sub make { my @characters = ( 0x0061 ); if ('a' =~ /\p{IsProperty}/) { correctly recognizes 'a' character as \x0061 code point on Perl 5.28.2 because $ perl /tmp/test However, this not true on Perl 5.30: $ perl /tmp/test Calling make(@characters) directly works as expected. Replacing make argument -- Petr Perl Info |
From @jkeenanOn Thu, 16 May 2019 16:14:21 GMT, ppisar wrote:
The line above is relevant. If you comment it out, this test program prints "not ok" on both 5.28 and blead. Do you know why that would be so?
Thank you very much. -- |
|
The RT System itself - Status changed from 'new' to 'open' |
From @jkeenanOn Thu, 16 May 2019 17:20:07 GMT, jkeenan wrote:
Original problem confirmed on FreeBSD-11.2. -- |
From @jkeenanOn Thu, 16 May 2019 17:29:19 GMT, jkeenan wrote:
FWIW, bisection pointed to this commit: ##### Move \p{user-defined} to core from utf8_heavy.pl Thank you very much. -- |
From @khwilliamsonNow fixed by PATCH: [perl #133860] 5.30 regression |
|
@khwilliamson - Status changed from 'open' to 'resolved' |
From @khwilliamsonOn 5/17/19 8:31 AM, Karl Williamson wrote:
It is now in blead, and very soon RC2 will be released. Thank you for pursuing this problem. This ticket is a duplicate of
|
From @ppisarOn Thu, May 16, 2019 at 03:57:43PM -0600, Karl Williamson wrote:
You are right. It also does not work with for(). I probably made a mistake.
I confirm the patch works for me. -- Petr |
From @khwilliamsonI'm thinking this should be in RC2. Earlier when this bug surfaced, as 133860, that appeared to be the only The branch smoke-me/khw-petr has a patch to fix this, now including a On 5/17/19 2:05 AM, Petr Pisar wrote:
|
p5pRT
added
BBC
Migrated from rt.perl.org#133860 (status was 'resolved')
Searchable as RT133860$
The text was updated successfully, but these errors were encountered: