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
perl blead: Update to Test-Simple [patch] #14027
Comments
From @exodistHere is a patch against perl blead to update to the latest Test-Simple Please see attached patch. Things changed: Thanks! -Chad |
From @jkeenanOn Thu Aug 14 12:22:51 2014, exodist7@gmail.com wrote:
Chad, I applied your patch locally in a branch. While applying it with 'git am', I got error output stating that there were 25 whitespace errors. I then ran 'make test'. I got this failure: ##### ../cpan/Test-Simple/t/Modern/encoding_test.t (Wstat: 512 Tests: 9 Failed: 2) ok 1 - filename is not in utf8 already # Failed test 'sanity check' # Failed test 'The utf8 diag has a different output' Test Summary Report ../cpan/Test-Simple/t/Modern/encoding_test.t (Wstat: 512 Tests: 9 Failed: 2) Can you investigate? Thank you very much. |
The RT System itself - Status changed from 'new' to 'open' |
From @exodistHere is an updated patch, whitespace errors fixed. This patch passes all tests for me, hopefully it will do the same for you. Note there are some files that use utf8, but also one file that uses non-utf8 special characters. |
From @jkeenanOn Sat Aug 16 10:02:21 2014, exodist7@gmail.com wrote:
Committed to blead in commit 6bdb887. Thanks for your persistent work on this. I will monitor smoke reports and close the ticket within 7 days unless there are significant complaints. I do have one question about the output from one test file: t/versions.t. ##### That "You should load Test::Tester ..." message is one of the longest -- perhaps *the* longest -- messages that get printed during perl5's 'make test'. I found it ... surprising. So is there any way we could bring it into accord with the "principle of least surprise"? Thank you very much. |
From @jkeenanOn Sat Aug 16 14:31:21 2014, jkeenan wrote:
Chad: We have one test failure reported during smoke testing which is probably tied to this revision. Please see: http://perl.develop-help.com/raw/?id=165720 Failure appears to be Win32-related. Excerpt from log (http://m-l.org/~perl/smoke/perl/win32/blead/log6bdb88770f849c47b5c09e733ac460ce3e9dbc97.log.gz) ##### Can you investigate? Thank you very much. |
From @exodistOn Sun Aug 17 03:57:20 2014, jkeenan wrote:
Looking into it. What would be the recommended way to check if fork is available? if fork is not available the test should simply be skipped. Usually fork is implemented on windows perls, I was not aware there were some without it. |
From @exodistI was able to figure it out by looking at what other forking tests did, I will supply a new patch shortly. |
From @exodistThis patch updates to a newer alpha, which should fix the previously listed problems on this ticket. |
From @iabynOn Sun, Aug 17, 2014 at 03:57:20AM -0700, James E Keenan via RT wrote:
There are two other serious issues. First, perl's overall test suite (I mean all 700,000 tests, not just the Files=2320, Tests=713707, 122 wallclock secs (80.00 usr 6.02 sys + 430.26 cusr 23.44 csys = 539.72 CPU) Also, /cpan/Test-Simple/t/Modern/More.t is failing on Merijn's linux Message-ID: <PRCR55389531@procura.nl> Failures: (common-args) none -- |
From @exodistOn Mon Aug 18 06:18:29 2014, davem wrote:
Is there anywhere I can see the verbose output of the failure? As for the CPU stuff, I will do some profiling and see if I can tighten things up... |
From @exodistThis includes: * Lots of performance fixes, won't be quite as fast as the old Test::Builder, but will be leaps and bounds faster than what is currently in blead In my testing this patch takes half as much time to pass all the tests compared to current blead with a previous alpha. |
From @jkeenanOn Mon Aug 18 15:47:30 2014, exodist7@gmail.com wrote:
To further the process of evaluating these revisions, I committed your most recent patch, 0001-Update-blead-to-Test-Builder-alpha-40.patch, to blead in commit 59c96ae. Here is some comparative data. The following 3 runs were all made on dromedary with 'sh ./Configure -des -Dusedevel && TEST_JOBS=${TEST_JOBS} make -j${TEST_JOBS} test_harness' where $TEST_JOBS is 8. ##### All tests successful. II. Just before applying second Exodist patch (commit b756670) All tests successful. III. Just after applying second Exodist patch (commit 59c96ae) All tests successful. So the second patch runs dramatically better than the first one -- but still measurably slower than what preceded the first patch. Thank you very much. |
From @exodistThe new Test::Builder alphas introduce a lot of new features, debugging tools, and a few new ways of getting things done. Ultimately the goal is more predictable testing tools, and better tools for diagnosing failures. It also fixes several bugs, some intentionally, some by virtue of doing things better (or perhaps just different) ways. It is probably not possible for it to be as fast as the old Test::Builder, unless someone wants to dive in an re-implement some parts in XS.... I will keep performance in mind moving forward, and see if I can squeeze anymore speed out of it, but I have already nabbed the low hanging fruit, as well as some of the fruit from the middle... Is this performance change too drastic to keep? At this point I would be very reluctant to remove the key feature that causes the performance change, the benefit in diagnostics and debugging are just too high to drop. |
From @doughera88On Mon, Aug 18, 2014 at 02:17:56PM +0100, Dave Mitchell wrote:
Yes, I see this too. Most seriously, on my aging Solaris/SPARC system, -- |
From PeterCMartini@GMail.com
5.21.3 is scheduled for tomorrow, and this seems like it will be the biggest change for that point release. Is there something we can do about this, at the least the test failure portion, before then? Thanks, |
From @cpansproutOn Tue Aug 19 06:10:34 2014, doughera wrote:
It also makes it extremely slow for those who test from time to time -- Father Chrysostomos |
From @jhiI am working at getting some more concrete timing data. On Tuesday, August 19, 2014, Father Chrysostomos via RT <
-- |
From @exodistOn Tue Aug 19 06:24:44 2014, pcm wrote:
Test failure portion? MY latest patch, which has been applied, fixes all the test failures I know about, the one exception is Test::Exception which fails to install with the latest T:B alpha, I am working on that (I am also now maintainer of Test::Exception) |
From @doughera88On Tue, Aug 19, 2014 at 09:24:00AM -0400, Peter Martini wrote:
I'm afraid I probably can't do anything useful in a timely fashion. -- |
From @TuxOn Tue, 19 Aug 2014 09:09:52 -0400, Andrew Dougherty
Only 12 hours? 1 configuration to finish, estimated completion in 1 day 1 hour 12 minutes -- |
From @iabynOn Tue, Aug 19, 2014 at 09:24:00AM -0400, Peter Martini wrote:
Is there any reason why the new Test::Simple can't be reverted for now? -- |
From @exodistOn Tue Aug 19 09:02:56 2014, davem wrote:
I have no objections, it is still in alpha, but simply never upgrading Test::Builder is probably not a valid long-term solution. I am working on more performance enhancements, trying to make it as fast as possible. I will submit a new patch later. |
From @jkeenanOn Tue Aug 19 09:02:56 2014, davem wrote:
No, it can be reverted. My original instinct was to simply put Chad's patches into a smoke-me branch. I did that, but on IRC people pointed out that we have ample precedents for including alpha releases of cpan/ modules in blead. In fact, the more important the distribution, the more we want to maximize both smoke tests and performance evaluations. That's better accomplished in blead than in a smoke-me branch. We now have evidence that these patches slow down the Perl 5 core test suite substantially. On that basis, I would vote to revert. It's early in the yearly production cycle -- though late in the monthly devel release cycle -- so we have time to ponder how we go forward. The pumpking and this month's release manager are meeting in onionsketch later this evening. I suggest they make the call. Thank you very much. |
From PeterCMartini@GMail.comOn Tue, Aug 19, 2014 at 6:41 PM, James E Keenan via RT
Discussed in the onionsketch and reverted by commit |
From @greergaOn Tue, 19 Aug 2014, H.Merijn Brand wrote:
August 18th: August 11th: That's my "most thorough" blead smoker. -- |
From @exodistWe can close this ticket. I submitted a new patch with a re-architected Test-Simple that will not have the performance issues of this one. I am not sure what ticket number was assigned to my latest submission, may not have one yet, sent the email a short time ago. |
From @exodistOn Wed Oct 22 09:53:55 2014, exodist7@gmail.com wrote:
Actually, since I cannot find a *new* ticket from my email, I figure it might be best just to provide my new patch here, so here it is. |
From @tonycozOn Wed Oct 22 09:56:57 2014, exodist7@gmail.com wrote:
The performance is ok, blead: All tests successful. vs patched: Files=2394, Tests=720189, 213 wallclock secs (71.58 usr 12.05 sys + 470.17 cusr 39.18 csys = 592.98 CPU) which is an 8% increase in CPU usage. Run with TEST_JOBS=6. Unfortunately there's a change of behaviour WRT $TODO. Before the patch, setting $TODO to the empty string (and possibly other false values) would not treat the test as TODO, but with the patch such cases are treated as TODO, so several tests are treated TODO passed: ../ext/XS-APItest/t/call_checker.t (Wstat: 0 Tests: 76 Failed: 0) I could update perl's tests to supply undef instead, but this change in behaviour may be a problem for CPAN distributions too. git still complains about whitespace errors. Tony |
From @jkeenanOn 10/23/2014 12:59 AM, Tony Cook via RT wrote:
I got a similar ratio: blead, before applying patch real 8m13.607s 122538-test-simple branch, after applying patch real 8m49.408s Branch takes 1.07 times longer than blead. |
From @jkeenanOn 10/23/2014 12:59 AM, Tony Cook via RT wrote:
I've created a smoke test branch for this but, as we did the last time |
From @exodist
Very interesting, and good catch, I had not noticed that in my own testing of cpan modules, but it is easy to miss since it is something that by definition will not result in failures. I will look into changing it on my end, don't modify perl's tests for it. Details: In Test::Builder/More there are 2 ways to set TODO, $TODO and Test::Builder->start_todo, the latter DOES count as todo for any defined value, and is tested to do so. $TODO on the other hand, which is older, simply had a boolean check. I unified a lot of logic, including these 2, and in that unification I went with the defined check. I should *probably* add some extra conditionals to make it do what it did before.
Can you share the whitespace errors with me? I will also try merging my own patch to see if I can reproduce them, but would be nice to have your list to check against. Also of note, there is only 1 cpan module I am aware of that is broken by this alpha, Test::Moose::More, and the author of it is already working on fixing it on his end, it is related to the subtest indentation changes from David Golden. |
From @exodistOn Thu Oct 23 07:29:29 2014, exodist7@gmail.com wrote:
I have fixed the fixable whitespace errors (I think I got them all anyway). The remaining whitespace errors are places where lines really do need to have trailing whitespace, or tabs and spaces mixed, such as unit tests that check output to ensure it respects back-compat. Here is a new patch against current blead (do not apply previous patch first). It fixes the $TODO issue, and includes a new test to validate the fix. |
From @exodistChecking on the status here. Is this going to go to blead, or is it waiting on something? |
From @jkeenanOn Sun Oct 26 07:04:18 2014, exodist7@gmail.com wrote:
Applied to blead in commit 07308ed Import Test-More 1.301001 alpha 63 Let's get this reviewed and benchmarked. -- |
From @cpansproutOn Sun Oct 26 09:31:36 2014, jkeenan wrote:
It introduces an incompatibility in how skip() is handled. Before: $ ./perl -Ilib -MTest::More -e 'SKIP: {skip "aa"}' It warns, but it still works. After: $ ./perl -Ilib -MTest::More -e 'SKIP: {skip "aa"}' No skip. This is causing smoke failures on Windows. (Personally I would like to see the warning obliterated. 1 is a perfectly sane default.) -- Father Chrysostomos |
From @jkeenanOn Sun Oct 26 09:31:36 2014, jkeenan wrote:
I built and installed two versions of perl, one at the commit just before that in which I applied the alpha 63 patch, and one at the alpha 36 commit. I then constructed a benchmark program (attached) in which I ran both versions over a selection of tests from the Parrot VM test suite. I got the results in the second attached file (10_iterations_benchmark.txt). (I made runs as high as 50 iterations; the results were consistent.) It appears the new Test-More runs at about 82% of the current Test-More. Thank you very much. -- |
From @jkeenan$ perl benchmark-122538.pl 10 before: after: Benchmark: timing 10 iterations of 000_b176455, 001_07308ed... Test::More::VERSION: 1.001008 Test::More::VERSION: 1.301001063 |
From @exodistOn Sun Oct 26 14:32:01 2014, jkeenan wrote:
Just to be clear, is this sufficient? I am not sure it is possible to have the new features without some slow down. Also any test suite with lots of short tests will have worse results then longer tests. Right now my benchmarking shows my updates actually being faster during run-time, but slower in build-time. I suspect the only remaining option for speedup is to remove some of the compile-time performance by pre-building the accessors that are currently generated. In my benchmarking compile time of simply loading Test::More and calling ok() and done_testing() the old version took ~30ms less time than the new one (the new one taking ~50ms) I don't have the numbers at hand, but they were close to that, if anything I made the alpha look worse than it is with those numbers. I am not convinced improvingg that compile-time is worth the effort and complexity it adds to the dist release process. |
From @exodistOn Sun Oct 26 12:11:57 2014, sprout wrote:
I will fix this and bring a new patch against blead. I will try to get to this tonight, but it may be tomorrow night.
My initial reaction is to not make the change, in case removing the warnings breaks any tests downstream. However since I have checked a lot of cpan, and never even found this bug, let alone a warnings related one, I think it is safe to remove the warning and make a sane default of 1. so I will remove the warning. If I recall from the code this was actually a conditional behavior, when there is no plan or a plan of 'no_plan' skip would be fine, but warn when not given a number. However when there is a numeric plan it would need to know the number (which is sane, but 1 still sounds like a good default). |
From @jkeenanOn Sun Oct 26 14:54:46 2014, exodist7@gmail.com wrote:
Ask 10 people, you'll get at least 11 answers! I doubt that we have any pre-established standards for how fast Test-More should be. We aren't very good at measuring the performance of perl itself (though see the thread initiated by Dave Mitchell at http://perl.markmail.org/search/?q=%22performance+testing%22+list%3Aorg.perl.perl5-porters#query:%22performance%20testing%22%20list%3Aorg.perl.perl5-porters+page:1+mid:dawtdydqpu3v4r7g+state:results). So evaluating the performance of modules distributed with core will inevitably entail some "making it up as we go along" -- and I concede that that's unsettling when you've put a lot of work into some effort. So for now I'm simply reporting the results for one set of tests that I am personally familiar with. I suspect that in the coming days other contributors will post their results and express their opinions -- just as they did with the first round of patches.
I suspect, unfortunately, that people are primarily going to evaluate the new Test-More on the basis of how it handles their *existing* test suites -- a situation where the number of test files or their individual length is not subject to change. For example, that is certainly the case on my day job. Thank you very much. -- |
From @exodistHere is a patch against current blead (which has my old one already applied) that fixes the TODO issue as well as a couple other minor things (doc updates mainly). |
From @exodist0001-Update-to-Test-More-alpha-064.patchFrom c275c31fd6515e44e3438a0b8c075f7623b6e58e Mon Sep 17 00:00:00 2001
From: Chad Granum <chad.granum@dreamhost.com>
Date: Sun, 26 Oct 2014 21:32:27 -0700
Subject: [PATCH] Update to Test::More alpha 064
* Fixes some docs
* Fixes TODO behavior change
---
cpan/Test-Simple/lib/Test/Builder.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Module.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Tester.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm | 2 +-
cpan/Test-Simple/lib/Test/More.pm | 4 ++--
cpan/Test-Simple/lib/Test/Simple.pm | 2 +-
cpan/Test-Simple/lib/Test/Stream.pm | 2 +-
cpan/Test-Simple/lib/Test/Stream/Architecture.pod | 20 ++++++++++----------
cpan/Test-Simple/lib/Test/Tester.pm | 2 +-
cpan/Test-Simple/lib/Test/use/ok.pm | 2 +-
cpan/Test-Simple/lib/ok.pm | 2 +-
cpan/Test-Simple/t/Legacy/skip.t | 4 +---
12 files changed, 22 insertions(+), 24 deletions(-)
diff --git a/cpan/Test-Simple/lib/Test/Builder.pm b/cpan/Test-Simple/lib/Test/Builder.pm
index c307e85..e7b5a99 100644
--- a/cpan/Test-Simple/lib/Test/Builder.pm
+++ b/cpan/Test-Simple/lib/Test/Builder.pm
@@ -4,7 +4,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::More::Tools;
diff --git a/cpan/Test-Simple/lib/Test/Builder/Module.pm b/cpan/Test-Simple/lib/Test/Builder/Module.pm
index 32b57b0..53a166a 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Module.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Module.pm
@@ -7,7 +7,7 @@ use Test::Builder 0.99;
require Exporter;
our @ISA = qw(Exporter);
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
diff --git a/cpan/Test-Simple/lib/Test/Builder/Tester.pm b/cpan/Test-Simple/lib/Test/Builder/Tester.pm
index 6e42c4c..9219b22 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Tester.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Tester.pm
@@ -1,7 +1,7 @@
package Test::Builder::Tester;
use strict;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Builder 1.301001;
diff --git a/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm b/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
index 774e16a..6338200 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
@@ -1,7 +1,7 @@
package Test::Builder::Tester::Color;
use strict;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
require Test::Builder::Tester;
diff --git a/cpan/Test-Simple/lib/Test/More.pm b/cpan/Test-Simple/lib/Test/More.pm
index 7a2135c..7f8afaf 100644
--- a/cpan/Test-Simple/lib/Test/More.pm
+++ b/cpan/Test-Simple/lib/Test/More.pm
@@ -4,7 +4,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream::Carp qw/croak carp/;
@@ -257,7 +257,7 @@ sub _skip {
my $plan = $ctx->stream->plan;
# If there is no plan we do not need to worry about counts
- my $need_count = $plan ? !($plan->directive && $plan->directive eq 'NO PLAN') : 1;
+ my $need_count = $plan ? !($plan->directive && $plan->directive eq 'NO PLAN') : 0;
$ctx->alert("$func() needs to know \$how_many tests are in the block")
if $need_count && !defined $how_many;
diff --git a/cpan/Test-Simple/lib/Test/Simple.pm b/cpan/Test-Simple/lib/Test/Simple.pm
index 90a53fe..2b24453 100644
--- a/cpan/Test-Simple/lib/Test/Simple.pm
+++ b/cpan/Test-Simple/lib/Test/Simple.pm
@@ -5,7 +5,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream '-internal';
diff --git a/cpan/Test-Simple/lib/Test/Stream.pm b/cpan/Test-Simple/lib/Test/Stream.pm
index ec470b4..816576b 100644
--- a/cpan/Test-Simple/lib/Test/Stream.pm
+++ b/cpan/Test-Simple/lib/Test/Stream.pm
@@ -2,7 +2,7 @@ package Test::Stream;
use strict;
use warnings;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream::Threads;
diff --git a/cpan/Test-Simple/lib/Test/Stream/Architecture.pod b/cpan/Test-Simple/lib/Test/Stream/Architecture.pod
index 620adcf..b98ce50 100644
--- a/cpan/Test-Simple/lib/Test/Stream/Architecture.pod
+++ b/cpan/Test-Simple/lib/Test/Stream/Architecture.pod
@@ -7,7 +7,7 @@ Test::Stream::Architecture - Overview of how the Test-More dist works.
This is the document that explains the architecture of Test::More and all the
stuff driving it under the hood.
-=head1 KEY COMPONTENTS
+=head1 KEY COMPONENTS
This is the list of primary components and their brief description, The most
critical ones will have more details in later sections.
@@ -90,9 +90,9 @@ proper object. All events must use Test::Stream::Event as a base.
This is the L<Moose> of Test::Stream. It is responsible for generating
accessors and similar work. Unlike moose and others it uses an arrayref as the
-underlying object. This design descision was made to improve performance.
+underlying object. This design decision was made to improve performance.
Performance was a real problem in some early alphas, the gains from the
-descision are huge.
+decision are huge.
=item Test::Stream::Tester
@@ -118,7 +118,7 @@ number, and shove them through the correct output.
The singleton model proved to be a major headache.
-Intercepting events typically meant replacing the singleton perminantly
+Intercepting events typically meant replacing the singleton permanently
(Test::Tester) or for a limited scope. Another option people took
(Test::Builder::Tester) was to simply replace the IO handles Test::Builder was
tracking.
@@ -160,7 +160,7 @@ in the order they are added.
There is not currently any way to remove a munger.
-B<Note:> each munger is called in a loop in the C<send()> method, so keep it is
+B<Note:> each munger is called in a loop in the C<send()> method, so keep it as
fast and small as possible.
=item Forward all events to listeners (including TAP output)
@@ -224,7 +224,7 @@ If there is a context already in play, that instance will be returned.
Otherwise a new context will be returned. The context assumes that the stack
level just above your call is where errors should be reported.
-You can optionally provide an integer as the only argument, in whcih case that
+You can optionally provide an integer as the only argument, in which case that
number will be added to the C<caller()> call to find the correct frame for
reporting. This will be completely ignored if there is already an active
context.
@@ -245,7 +245,7 @@ In the example above c<my_ok()> generates a new context, then it calls C<ok()>,
in this case both function will have the same context object, the one generated
by my_ok. The result is that C<ok> will report errors to the correct place.
-=head3 IMPLEMENTEATION
+=head3 IMPLEMENTATION
There is a variable C<$CURRENT> in C<Test::Stream::Context>, it is a lexical,
so you can not touch it directly. When the C<context()> function is called, it
@@ -258,7 +258,7 @@ scope, or is otherwise removed, C<$CURRENT> will vanish on its own. This means
that so long as you hold on to your context object, anything you call will find
it.
-B<The caveat> here is that if you descide to hold on to your context beyond
+B<The caveat> here is that if you decide to hold on to your context beyond
your scope, you could sabatoge any future test functions. If you need to hold
on to a context you need to call C<< $context->snapshot >>, and store the
cloned object it returns. In general you should not need to do this, event
@@ -266,9 +266,9 @@ objects all store the context, but do so using a snapshot.
B<Note> I am open to changing this to remove the weak-reference magic and
instead require someone to call C<< $context->release >> or similar when they
-are done with a context, but that seams more likely to result in rougue
+are done with a context, but that seems more likely to result in rougue
contexts... This method would also require its own form of reference counting..
-This descision will need to be made before we go stable.
+This decision will need to be made before we go stable.
=head2 GENERATING EVENTS
diff --git a/cpan/Test-Simple/lib/Test/Tester.pm b/cpan/Test-Simple/lib/Test/Tester.pm
index 25a62a9..1e52c1a 100644
--- a/cpan/Test-Simple/lib/Test/Tester.pm
+++ b/cpan/Test-Simple/lib/Test/Tester.pm
@@ -15,7 +15,7 @@ require Exporter;
use vars qw( @ISA @EXPORT $VERSION );
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
@EXPORT = qw( run_tests check_tests check_test cmp_results show_space );
diff --git a/cpan/Test-Simple/lib/Test/use/ok.pm b/cpan/Test-Simple/lib/Test/use/ok.pm
index ccbd5ad..68d6307 100644
--- a/cpan/Test-Simple/lib/Test/use/ok.pm
+++ b/cpan/Test-Simple/lib/Test/use/ok.pm
@@ -3,7 +3,7 @@ use strict;
use warnings;
use 5.005;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
1;
diff --git a/cpan/Test-Simple/lib/ok.pm b/cpan/Test-Simple/lib/ok.pm
index f20230e..909b7c6 100644
--- a/cpan/Test-Simple/lib/ok.pm
+++ b/cpan/Test-Simple/lib/ok.pm
@@ -5,7 +5,7 @@ use warnings;
use Test::More 1.301001 ();
use Test::Stream::Carp qw/croak/;
-our $VERSION = '1.301001_063';
+our $VERSION = '1.301001_064';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
sub import {
diff --git a/cpan/Test-Simple/t/Legacy/skip.t b/cpan/Test-Simple/t/Legacy/skip.t
index df4d6e1..c67e680 100644
--- a/cpan/Test-Simple/t/Legacy/skip.t
+++ b/cpan/Test-Simple/t/Legacy/skip.t
@@ -7,7 +7,7 @@ BEGIN {
}
}
-use Test::More;
+use Test::More tests => 15;
BEGIN {
require warnings;
@@ -103,5 +103,3 @@ SKIP: {
like $warning, qr/^skip\(\) was passed a non-numeric number of tests/;
}
-
-done_testing;
--
1.9.1
|
From @cpansproutOn Sun Oct 26 21:53:18 2014, exodist7@gmail.com wrote:
Thank you. I have applied it as 6213b78. However, it still does not fix the skip incompatibility, so I patch Test::More in place in commit df8d573. While we usually try to avoid modifying modules under cpan/ directly, in this case we have been getting smoke failures from Windows for 30 hours already, which potentially allows other problems to get lost in the noise. So I have chosen the lesser of two weevils. I hope you can make a new Test-Simple alpha release with a similar fix soon. -- Father Chrysostomos |
From @exodistOn Mon Oct 27 15:48:47 2014, sprout wrote:
Ah, thank you. I thought I had it fixed, but I missed that one critical part. I will probably copy your patch as-is into my repo. I don't have a windows system handy that I could use to check this myself. I also would not know the first thing about getting a blead perl running on windows (cygwin maybe?). You mentioned smokers were having issues with this, is that data public/published somewhere that I can check it regularly? |
From @cpansproutOn Mon Oct 27 22:03:15 2014, exodist7@gmail.com wrote:
I usually just watch the smoke messages coming to p5p. See for instance <20141027193832.184BE4E023E@zwei> and <225808580.160.1414426906096.JavaMail.jenkins@jenkins-001.ams6.corp.booking.com>, which show dist/PathTools/t/cwd.t failing. Those two particular messages can be viewed at <http://www.nntp.perl.org/group/perl.perl5.porters/2014/10/msg221733.html> and <http://www.nntp.perl.org/group/perl.perl5.porters/2014/10/msg221728.html>. -- Father Chrysostomos |
From @exodistHere is a patch to alpha 065 which includes the TODO fix you already applied, as well as a couple other things. |
From @exodist0001-Update-Test-Simple-to-alpha-065.patchFrom fbb29cbf1aac8c5bfebbb03e82104c66a813cecb Mon Sep 17 00:00:00 2001
From: Chad Granum <chad.granum@dreamhost.com>
Date: Tue, 28 Oct 2014 21:35:07 -0700
Subject: [PATCH] Update Test-Simple to alpha 065
This includes a patch from Father Chrysostomos which was already in
blead. An additional fix to some windows configs. And a version bump.
---
cpan/Test-Simple/lib/Test/Builder.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Module.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Tester.pm | 2 +-
cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm | 2 +-
cpan/Test-Simple/lib/Test/More.pm | 3 ++-
cpan/Test-Simple/lib/Test/Simple.pm | 2 +-
cpan/Test-Simple/lib/Test/Stream.pm | 2 +-
cpan/Test-Simple/lib/Test/Tester.pm | 2 +-
cpan/Test-Simple/lib/Test/use/ok.pm | 2 +-
cpan/Test-Simple/lib/ok.pm | 2 +-
cpan/Test-Simple/t/Legacy/exit.t | 4 ++--
11 files changed, 13 insertions(+), 12 deletions(-)
diff --git a/cpan/Test-Simple/lib/Test/Builder.pm b/cpan/Test-Simple/lib/Test/Builder.pm
index e7b5a99..54e392d 100644
--- a/cpan/Test-Simple/lib/Test/Builder.pm
+++ b/cpan/Test-Simple/lib/Test/Builder.pm
@@ -4,7 +4,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::More::Tools;
diff --git a/cpan/Test-Simple/lib/Test/Builder/Module.pm b/cpan/Test-Simple/lib/Test/Builder/Module.pm
index 53a166a..d7936da 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Module.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Module.pm
@@ -7,7 +7,7 @@ use Test::Builder 0.99;
require Exporter;
our @ISA = qw(Exporter);
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
diff --git a/cpan/Test-Simple/lib/Test/Builder/Tester.pm b/cpan/Test-Simple/lib/Test/Builder/Tester.pm
index 9219b22..ae98639 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Tester.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Tester.pm
@@ -1,7 +1,7 @@
package Test::Builder::Tester;
use strict;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Builder 1.301001;
diff --git a/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm b/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
index 6338200..c205551 100644
--- a/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
+++ b/cpan/Test-Simple/lib/Test/Builder/Tester/Color.pm
@@ -1,7 +1,7 @@
package Test::Builder::Tester::Color;
use strict;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
require Test::Builder::Tester;
diff --git a/cpan/Test-Simple/lib/Test/More.pm b/cpan/Test-Simple/lib/Test/More.pm
index 1b12ac0..5aea2de 100644
--- a/cpan/Test-Simple/lib/Test/More.pm
+++ b/cpan/Test-Simple/lib/Test/More.pm
@@ -4,7 +4,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream::Carp qw/croak carp/;
@@ -103,6 +103,7 @@ sub ok ($;$) {
}
sub plan {
+ return unless @_;
my ($directive, $arg) = @_;
my $ctx = context();
diff --git a/cpan/Test-Simple/lib/Test/Simple.pm b/cpan/Test-Simple/lib/Test/Simple.pm
index 2b24453..59c7a28 100644
--- a/cpan/Test-Simple/lib/Test/Simple.pm
+++ b/cpan/Test-Simple/lib/Test/Simple.pm
@@ -5,7 +5,7 @@ use 5.008001;
use strict;
use warnings;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream '-internal';
diff --git a/cpan/Test-Simple/lib/Test/Stream.pm b/cpan/Test-Simple/lib/Test/Stream.pm
index 816576b..87aa059 100644
--- a/cpan/Test-Simple/lib/Test/Stream.pm
+++ b/cpan/Test-Simple/lib/Test/Stream.pm
@@ -2,7 +2,7 @@ package Test::Stream;
use strict;
use warnings;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
use Test::Stream::Threads;
diff --git a/cpan/Test-Simple/lib/Test/Tester.pm b/cpan/Test-Simple/lib/Test/Tester.pm
index 1e52c1a..62a9001 100644
--- a/cpan/Test-Simple/lib/Test/Tester.pm
+++ b/cpan/Test-Simple/lib/Test/Tester.pm
@@ -15,7 +15,7 @@ require Exporter;
use vars qw( @ISA @EXPORT $VERSION );
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
@EXPORT = qw( run_tests check_tests check_test cmp_results show_space );
diff --git a/cpan/Test-Simple/lib/Test/use/ok.pm b/cpan/Test-Simple/lib/Test/use/ok.pm
index 68d6307..26417bb 100644
--- a/cpan/Test-Simple/lib/Test/use/ok.pm
+++ b/cpan/Test-Simple/lib/Test/use/ok.pm
@@ -3,7 +3,7 @@ use strict;
use warnings;
use 5.005;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
1;
diff --git a/cpan/Test-Simple/lib/ok.pm b/cpan/Test-Simple/lib/ok.pm
index 909b7c6..a8c8dc4 100644
--- a/cpan/Test-Simple/lib/ok.pm
+++ b/cpan/Test-Simple/lib/ok.pm
@@ -5,7 +5,7 @@ use warnings;
use Test::More 1.301001 ();
use Test::Stream::Carp qw/croak/;
-our $VERSION = '1.301001_064';
+our $VERSION = '1.301001_065';
$VERSION = eval $VERSION; ## no critic (BuiltinFunctions::ProhibitStringyEval)
sub import {
diff --git a/cpan/Test-Simple/t/Legacy/exit.t b/cpan/Test-Simple/t/Legacy/exit.t
index 8c9e16d..ba29394 100644
--- a/cpan/Test-Simple/t/Legacy/exit.t
+++ b/cpan/Test-Simple/t/Legacy/exit.t
@@ -61,7 +61,7 @@ END { 1 while unlink "exit_map_test" }
for my $exit (0..255) {
# This correctly emulates Test::Builder's behavior.
- my $out = qx[$Perl exit_map_test $exit];
+ my $out = qx["$Perl" exit_map_test $exit];
$TB->like( $out, qr/^exit $exit\n/, "exit map test for $exit" );
$Exit_Map{$exit} = exitstatus($?);
}
@@ -92,7 +92,7 @@ chdir 't';
my $lib = File::Spec->catdir(qw(lib Test Simple sample_tests));
while( my($test_name, $exit_code) = each %Tests ) {
my $file = File::Spec->catfile($lib, $test_name);
- my $wait_stat = system(qq{$Perl -"I../blib/lib" -"I../lib" -"I../t/lib" $file});
+ my $wait_stat = system(qq{"$Perl" -"I../blib/lib" -"I../lib" -"I../t/lib" $file});
my $actual_exit = exitstatus($wait_stat);
if( $exit_code eq 'not zero' ) {
--
1.9.1
|
From @cpansproutOn Tue Oct 28 21:37:25 2014, exodist7@gmail.com wrote:
Thank you. Applied as 64f5fd4. -- Father Chrysostomos |
@cpansprout - Status changed from 'open' to 'resolved' |
Migrated from rt.perl.org#122538 (status was 'resolved')
Searchable as RT122538$
The text was updated successfully, but these errors were encountered: