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
5.11.3: benchmark tests fail on netbsd 5.0.1 #10031
Comments
From @obra-- |
From @obraMessage RFC822: Automated smoke report for 5.11.3 patch 16576cb v5.11.2-235-g16576cb Summary: FAIL(F) O = OK F = Failure(s), extended report at the bottom v5.11.2-235-g16576cb Configuration (common) none F F F F Locally applied patches: Failures: (common-args) none [perlio] [perlio] -Duseithreads -Duse64bitint Compiler messages(gcc): -- |
From @obraAutomated smoke report for 5.11.3 patch 16576cb v5.11.2-235-g16576cb Summary: FAIL(F) O = OK F = Failure(s), extended report at the bottom v5.11.2-235-g16576cb Configuration (common) none Locally applied patches: Failures: (common-args) none [perlio] [perlio] -Duseithreads -Duse64bitint Compiler messages(gcc): -- |
From @toddrI just re-produced this on suse 9.1 64bit. So it's not netbsd only. I am building from blead. In my case, I failed test 140,159, but it's the same test from a different iteration. Here is the harness output: # Failed test 'fastslow should be > 0' # Failed test 'fastslow should be > 0' ... ../lib/Benchmark.t (Wstat: 512 Tests: 194 Failed: 2) |
The RT System itself - Status changed from 'new' to 'open' |
From @toddrPart of the problem is that this test file for good or for evil has tests that fail 2-3 subs deep @@ -333,12 +333,14 @@ sub check_graph_consistency { |
From @toddrtypo on the patch (cmp_ok line) @@ -333,12 +333,14 @@ sub check_graph_consistency { |
From @smpetersOn Mon Dec 21 06:43:29 2009, jesse wrote: Benchmark has a tendency of failing on old, slow PCs. I irregularly had Steve |
From @smpetersOn Wed Feb 10 18:52:14 2010, stmpeters wrote:
I cannot replicate this with a straight build of bleadperl. IIRC, smoke Steve |
From @smpetersOn Wed Feb 10 20:35:49 2010, stmpeters wrote:
I couldn't replicate the problems seen above. That said I still had one Steve Automated smoke report for 5.11.4 patch Summary: FAIL(X) O = OK F = Failure(s), extended report at the bottom v5.11.4-76-g482d21b Configuration (common) none Locally applied patches: Failures: (common-args) none Compiler messages(gcc): |
From @toddrOn Thu Feb 11 11:18:13 2010, stmpeters wrote:
No offense is meant, but after digging in Benchmark.t for the past couple of days, I'm not As my feel that this is more a performance/and/or logic thing with the test, I agree this |
From @obra
Thanks very much for the triage. I'm really unhappy about tests that I'm going to unlink this as a 5.12 blocker, but I'd be thrilled to see someone |
From @toddrI just reproduced on Suse 10 64bit no threads for tests 126/127 I had added in the patch above to try to track stack information. Interestingly I notice some ../lib/AnyDBM_File.t .............................................. ok # Failed test 'slowfast should be less than or equal to zero, and >= -100' # Failed test 'fastslow should be > 0' |
From @toddrSummary of what's failing: You have 2 code snippets: sub fib { We're supposed to calculate the speed at which one or the other runs. In this case, I assume How to possibly fix: 2. The other option would be to address how slightly different the runs are designed to be. in Possible patch: NOTE: I wanted to get people's opinion on this so I haven't actually tried the patch as of this |
From [Unknown Contact. See original ticket]Summary of what's failing: You have 2 code snippets: sub fib { We're supposed to calculate the speed at which one or the other runs. In this case, I assume How to possibly fix: 2. The other option would be to address how slightly different the runs are designed to be. in Possible patch: NOTE: I wanted to get people's opinion on this so I haven't actually tried the patch as of this |
From @toddrThe suggested patch is this +60 in the prev patch turns out to be really slow Inline Patchdiff --git a/lib/Benchmark.t b/lib/Benchmark.t
index 9b8e2af..29c2bb0 100644
--- a/lib/Benchmark.t
+++ b/lib/Benchmark.t
@@ -259,8 +259,8 @@ like ($got, $Default_Pattern, 'should find default format somewhere');
}
-my $code_to_test = { Foo => sub {$foo+=fib($ballast)},
- Bar => sub {$bar+=fib($ballast)}};
+my $code_to_test = { Foo => sub {$foo+=fib($ballast+10)},
+ Bar => sub {$bar+=fib($ballast+60)}};
# Keep these for later.
my $results;
{ |
From @toddrAnother approach might be to simply use time() in fib to make sure X time goes by. This #fib right now is: #Change it to this: |
From [Unknown Contact. See original ticket]Another approach might be to simply use time() in fib to make sure X time goes by. This #fib right now is: #Change it to this: |
From @toddrGiven the age of this case, I'd suggest closing it. We have not seen this build failure on our servers in years. |
Migrated from rt.perl.org#71512 (status was 'open')
Searchable as RT71512$
The text was updated successfully, but these errors were encountered: