Re: Test::Simple API incompatibility - cpan-testers FAIL results
Nicholas Clark wrote: On Fri, Feb 16, 2007 at 09:17:00AM +1100, Adam Kennedy wrote: Because the other alternative would be Doing a Schwern and I I'm sure you'd prefer to preserve that term for something positive :) Such as herding MakeMaker for years with no-one ever saying thanks. I offer up a prayer to Schwern every time something like Module::Build reminds me just how much worse it could be. -- David Cantrell
Re: Test::Simple API incompatibility - cpan-testers FAIL results
On 15 Feb 2007, at 12:35, David Cantrell wrote: [snip] The version of Test::Simple distributed with perl 5.6.2, which is presumably being used by one of your dependencies, is 0.47, so one result was testing with the earlier version, the other with the current version. There was an API change in 0.48. [snip] Ta. I updated the T::B::T dependency when the API changed, but never thought somebody would be running an old T::S with the new T::B::T. A simple fix would be to add Test::Simple = 0.48 as a dependency in your PREREQ_PM. That, however, might break other stuff in 5.6 installations that depends on the old behaviour. Yuck. I've added a T::S dependency now - which I'm happy with, since it was really implicit in the T::B::T dependency for most folk. 0.25 hitting CPAN soon. Cheers, Adrian
Re: Test::Simple API incompatibility - cpan-testers FAIL results
David Cantrell wrote: On Thu, Feb 15, 2007 at 11:48:57AM +, Adrian Howard wrote: Probably being stupid... but if you can point me towards the difference between http://www.nntp.perl.org/group/perl.cpan.testers/2007/02/msg415380.html and http://www.nntp.perl.org/group/perl.cpan.testers/2007/02/msg415375.html I'd appreciate it. Looks like the output of different T::B versions, but they seem to be running on identical ones. Me confused. It's this: PERL5LIB = vs this PERL5LIB = /blah/blah/Test-Simple-0.67/blah The version of Test::Simple distributed with perl 5.6.2, which is presumably being used by one of your dependencies, is 0.47, so one result was testing with the earlier version, the other with the current version. There was an API change in 0.48. This is not the result of the 0.48 API change which had to do with threading. This comes from the OMGYOUBROKECPAN! diagnostics formatting change in 0.60_01. Let me make this very clear once again, just so everyone knows: *--THERE WAS NO API CHANGE BECAUSE THERE IS NO API TO CHANGE--* This is not to be pedantic but a very real warning not to rely on the formatting of Test::More's diagnostic output. It changes. The only safe way to do it is through Test::Builder::Tester. This sucks but I'm not going to freeze my formatting and its a lame way to scrape extra information out of the tests anyway. If you want to do something about it help with this: http://perl-qa.yi.org/index.php/TAP_diagnostic_syntax The change in 0.60_01 broke TBT which then caused a cascade failure through all the rest of the testing modules. TBT is now shipped with Test-Simple to avoid their falling out of sync. Looks like they managed it anyway. Here's how the diagnostics have changed over time. 0 windhund ~/releases/Test-Simple-0.60$ perl -Ilib -wle 'use Test::More tests = 1; fail' 1..1 not ok 1 # Failed test (-e at line 1) # Looks like you failed 1 test of 1. 0 windhund ~/releases/Test-Simple-0.62$ perl -Ilib -wle 'use Test::More tests = 1; fail' 1..1 not ok 1 # Failed test in -e at line 1. # Looks like you failed 1 test of 1. And then again in 0.64_01. $ perl -wle 'use Test::More tests = 1; fail' 1..1 not ok 1 # Failed test at -e line 1. # Looks like you failed 1 test of 1. The purpose of that last one was so it reports the file/line like Perl does. $ perl -wle 'die' Died at -e line 1. While everyone seems to just upgrade Test::Simple and to hell with the consequences, I can't help but feel some concern about this practice. You'll note the sky remained firmly overhead during that last formatting change.
Re: Test::Simple API incompatibility - cpan-testers FAIL results
--- Michael G Schwern [EMAIL PROTECTED] wrote: This comes from the OMGYOUBROKECPAN! diagnostics formatting change in 0.60_01. Let me make this very clear once again, just so everyone knows: *--THERE WAS NO API CHANGE BECAUSE THERE IS NO API TO CHANGE--* In the Test::More docs, for diagnostics, it reads: NOTE The exact formatting of the diagnostic output is still changing, but it is guaranteed that whatever you throw at it it won't interfere with the test. Given that this issue has bitten people more than once, can I recommend that this warning be spelled out even clearer, maybe even in big, bold letters or in its own section of the docs? Cheers, Ovid -- Buy the book -- http://www.oreilly.com/catalog/perlhks/ Perl and CGI -- http://users.easystreet.com/ovid/cgi_course/
Re: Test::Simple API incompatibility - cpan-testers FAIL results
Ovid wrote: In the Test::More docs, for diagnostics, it reads: NOTE The exact formatting of the diagnostic output is still changing, but it is guaranteed that whatever you throw at it it won't interfere with the test. Given that this issue has bitten people more than once, can I recommend that this warning be spelled out even clearer, maybe even in big, bold letters or in its own section of the docs? rt.cpan.org eagerly awaits your ticket and patch.
Re: Test::Simple API incompatibility - cpan-testers FAIL results
This comes from the OMGYOUBROKECPAN! diagnostics formatting change in 0.60_01. Can we use that as the official name for this sort of error? Because the other alternative would be Doing a Schwern and I I'm sure you'd prefer to preserve that term for something positive :) Adam K
Re: Test::Simple API incompatibility - cpan-testers FAIL results
On Fri, Feb 16, 2007 at 09:17:00AM +1100, Adam Kennedy wrote: This comes from the OMGYOUBROKECPAN! diagnostics formatting change in 0.60_01. Can we use that as the official name for this sort of error? Because the other alternative would be Doing a Schwern and I I'm sure you'd prefer to preserve that term for something positive :) Such as herding MakeMaker for years with no-one ever saying thanks. Thanks, Schwern. (although I'm feeling a bit false as I'm not sure how much of that is simply because I'm glad that it wasn't me who had to do it) Nicholas Clark
Re: Test::Simple API incompatibility - cpan-testers FAIL results
Nicholas Clark wrote: On Fri, Feb 16, 2007 at 09:17:00AM +1100, Adam Kennedy wrote: This comes from the OMGYOUBROKECPAN! diagnostics formatting change in 0.60_01. Can we use that as the official name for this sort of error? Because the other alternative would be Doing a Schwern and I I'm sure you'd prefer to preserve that term for something positive :) I don't know, having the power to ruin the day of thousands of programmers around the world IN TWO DIFFERENT WAYS is pretty cool. :P Such as herding MakeMaker for years with no-one ever saying thanks. Thanks, Schwern. (although I'm feeling a bit false as I'm not sure how much of that is simply because I'm glad that it wasn't me who had to do it) Yeah, well--pssh. Its not like you've got anything better to do! :P
