New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bleadperl v5.25.3-22-g8901dde breaks FLORA/Class-C3-XS-0.13.tar.gz #15484
Comments
From @andkbisect commit 8901dde dist/: remove . from @INC when loading optional modules diagnostics 5.25.3 http://www.cpantesters.org/cpan/report/74f15440-5329-11e6-9a64-21b9ed09dfea affected FLORA/Class-C3-XS-0.13.tar.gz perl -V Summary of my perl5 (revision 5 version 25 subversion 4) configuration: Characteristics of this binary (from libperl): |
From @cpansproutOn Thu Jul 28 21:21:06 2016, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
I see something that really does need to be fixed before those maintenance releases are made. The error message from base.pm about making the module available in @INC lists @INC *with* . on the end, and mentions nothing about . having been stripped. So we have an error message that effectively says that the module cannot be found in ./ which is a lie. At least the diagnostics can be made more informative. Currently the message is: Base class package "t::lib::C" is empty. Maybe we need to add: (Note that . was stripped from the end for security Awful. Any better suggestions? Should base.pm instead be made smarter and allow . if the calling package itself was loaded from the current directory? (That would fix the Class::C3::XS failure.) -- Father Chrysostomos |
The RT System itself - Status changed from 'new' to 'open' |
From @cpansproutOn Thu Jul 28 21:21:06 2016, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
Those three are failing because of base.pm.
That one depends on File::HomeDir, so I cannot test it.
Those three also fail because of base.pm. -- Father Chrysostomos |
From @karenetheridgeOn Thu Jul 28 22:26:27 2016, sprout wrote:
I'm going to do a release of Class::C3::XS this week, so don't worry about that one. |
From @cpansproutOn Thu Jul 28 21:41:09 2016, sprout wrote:
In order to get things moving, I have pushed this commit: commit bca5527 [perl #128769] Improve base.pm @INC '.' handling I am hoping that it--or something like it--can be backported to the maint branches. -- Father Chrysostomos |
From zefram@fysh.orgFather Chrysostomos via RT wrote:
There's obvious appeal in doing this, but on the whole I think it's -zefram |
From @xsawyerxOn 08/01/2016 08:41 AM, Zefram wrote:
We should determine this point in order to decide whether to include (I had to send to vendors the versioning fix of PathTools, for example.) |
From @cpansproutOn Mon Aug 01 08:02:24 2016, xsawyerx@gmail.com wrote:
Yes, but it would be nice if you would express an opinion one way on another. :-) I think what Zefram says makes sense. -- Father Chrysostomos |
From @cpansproutThen again, what Chris Travers writes in #15490 also makes sense. I think this is the sort of situation in which the pumpking needs to decide, because nobody will ever agree. -- Father Chrysostomos |
From @cpansproutOn Mon Aug 01 08:50:17 2016, sprout wrote:
And so I have made the change in commit 362f3f7. -- Father Chrysostomos |
From @eserteOn Thu Jul 28 21:21:06 2016, andreas.koenig.7os6VVqR@franz.ak.mind.de wrote:
Also posibly affected: Tangence-0.21 (CPAN RT ticket 116579) |
From zefram@fysh.orgFather Chrysostomos via RT wrote:
What Chris Travers writes is a bunch of good reasons to not perform any Although base.pm incurs the ugliness especially much, it actually -zefram |
From @xsawyerxOn 08/01/2016 05:50 PM, Father Chrysostomos via RT wrote:
My mail client decided to just not send my emails. Great. I wrote that I usually wait with comments until I hear other considerations. However, my position is that it should be consistent - the error and the |
From chris.travers@gmail.comOn Thu Jul 28 21:41:09 2016, sprout wrote:
Recommending use lib '.' to get old behavior back is not what I would recommend other than as a last resort. Makes sense mostly when you have another module chdir'ing during module load or import (yikes!). "Note that base.pm now ignores the current working directory when requiring if this was not explicitly selected. If your software worked on previous versions of Perl, the best solution is to use FindBin to detect the path properly. As a last resort, you can re-enable looking in the current working directory by adding use lib '.' to your code."
|
From [Unknown Contact. See original ticket]On Thu Jul 28 21:41:09 2016, sprout wrote:
Recommending use lib '.' to get old behavior back is not what I would recommend other than as a last resort. Makes sense mostly when you have another module chdir'ing during module load or import (yikes!). "Note that base.pm now ignores the current working directory when requiring if this was not explicitly selected. If your software worked on previous versions of Perl, the best solution is to use FindBin to detect the path properly. As a last resort, you can re-enable looking in the current working directory by adding use lib '.' to your code."
|
From @cpansproutOn Wed Aug 03 03:58:11 2016, chris.travers@gmail.com wrote:
Thank you for taking the time to write a much better message than I could come up with. :-) What I actually ended up committing was this: If you mean to load $fn from the current directory, you may But your message is clearer and more informative. Nevertheless, I think it could do with some tweaks, specifically to mention that it is @INC we are talking about. (Also, I find the use of ‘when requiring’ a little arcane.) What do you think of this? The file t/lib/Foo.pm exists in the current directory. Note that -- Father Chrysostomos |
From chris.travers@gmail.comOn Thu Aug 04 22:15:37 2016, sprout wrote:
Even better. I like it. |
From @cpansproutOn Thu Aug 04 22:25:28 2016, chris.travers@gmail.com wrote:
Thank you. I have applied it as 458470f. -- Father Chrysostomos |
From @cpansproutHow should we resolve this ticket now (in addition to patching the CPAN modules)? Do my changes to base.pm need to be +1’ed by two other people so that we can backport them to the maint branches and release new RCs? The patches in question are: bca5527 -- Father Chrysostomos |
From @xsawyerxOn 08/07/2016 07:07 AM, Father Chrysostomos via RT wrote:
+1 |
From @steve-m-hayOn 7 August 2016 at 11:03, Sawyer X <xsawyerx@gmail.com> wrote:
+1. I'll go ahead and make new RCs with backports of these five patches |
From @andkAlso affected RPETTETT/ClearPress-v473.0.5.tar.gz (only with DBD::SQLite installed, -- |
From @jkeenanOn Tue, 09 Aug 2016 20:37:52 GMT, shay wrote:
Has the work which P5P had to do for this RT been completed? If so, do we have a plan for the BBC issues? Thank you very much. -- |
From @jkeenanOn Tue, 06 Dec 2016 22:30:13 GMT, jkeenan wrote:
Repeating the question: Has the work which P5P had to do for this RT been completed? If so, do we have a plan for the BBC issues? -- |
From @xsawyerxOn 12/26/2016 03:11 AM, James E Keenan via RT wrote:
(Apparently adding "shift" just removes the email I was about to send. Yay.) These were all merged but I think we should remove them in favor of the |
From @iabynOn Mon, Dec 26, 2016 at 06:07:55PM +0100, Sawyer X wrote:
Are we still planning to do this for 5.26.0? If not, can this ticket -- |
From @jkeenanOn Mon, 20 Mar 2017 12:53:42 GMT, davem wrote:
My vote would be to remove it from the blockers list. Even if the work in the ap/baseincguard were to be completed, there's not sufficient time to test it against CPAN before full code freeze. But this is a pumpking's call. Thank you very much. |
From @haargOn Mon, Mar 20, 2017 at 4:37 PM, James E Keenan via RT
Can't we just revert the base.pm changes without putting anything |
From @xsawyerxOn 03/21/2017 12:51 PM, Graham Knop wrote:
This makes sense. I spoke to Graham more about this on #p5p. There's no need to have an So we should undo the changes to base.pm in blead. Post 5.26 we could |
From @iabynOn Tue, Mar 21, 2017 at 09:17:17PM +0100, Sawyer X wrote:
I've just pushed the following 3 commits: commit 6ee05a9 Revert base.pm's dot-in-INC changes. commit 2250cd0 bump base.pm $VERSION and un-CUSTOMISE commit 939e7f2 INSTALL: add entry for -Ddefault_inc_excludes_dot -- |
From @iabynOn Fri, Mar 24, 2017 at 09:00:25AM +0000, Dave Mitchell wrote:
.. and am removing this ticket from the 5.26.0 blockers list. I'm not sure whether the ticket can be closed. -- |
Given the age of this ticket I'd say we can close it? If there are further |
Yes, we can close it. I reviewed all the CPAN distros listed in this ticket. All but three have passing tests on CPANtesters. The three which do not appear to be actively maintained.
Newly found breakage due to Thank you very much. |
Migrated from rt.perl.org#128769 (status was 'open')
Searchable as RT128769$
The text was updated successfully, but these errors were encountered: