Skip Menu |

This queue is for tickets about the Test-Smoke CPAN distribution.

Report information
The Basics
Id: 126010
Status: new
Priority: 0/
Queue: Test-Smoke

People
Owner: Nobody in particular
Requestors: jkeenan [...] cpan.org
Cc: public [...] khwilliamson.com
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)



CC: public [...] khwilliamson.com
Subject: ./logs/smokecurrent/logXXX.log files are mysteriously truncated
As Test::Smoke run, it logs output in 'smokecurrent.log'. ##### $ ls -l smokecurrent.log -rw-r--r-- 1 jkeenan jkeenan 603104 Aug 4 21:56 smokecurrent.log ##### At the conclusion of a run, it creates 4 files in the logs/smokecurrent subdirectory, the names of which files are composed from the SHA of the commit being tested. ##### $ ls -ltr logs/smokecurrent/ | tail -4 -rw-rw-rw- 1 jkeenan jkeenan 1454 Aug 4 21:56 rpte94c3f6a8d515b39961010524037ca044f029524.rpt -rw-rw-rw- 1 jkeenan jkeenan 2371 Aug 4 21:56 oute94c3f6a8d515b39961010524037ca044f029524.out -rw-rw-rw- 1 jkeenan jkeenan 595939 Aug 4 21:56 loge94c3f6a8d515b39961010524037ca044f029524.log -rw-rw-rw- 1 jkeenan jkeenan 6471 Aug 4 21:56 jsne94c3f6a8d515b39961010524037ca044f029524.jsn ##### "Naively" -- that is, as a typical user who has not peered deeply into the Test::Smoke code base or documentation -- I would expect that ./smokecurrent.log and ./logs/smokecurrent/log***.log to be either (a) exactly the same file or (b) to be almost exactly the same but differ in some clear, well thought-out way. By observation over repeated Test::Smoke runs, I infer that (b) is usally the case and that some data is deliberately and thoughtfully removed from ./smokecurrent.log before ./logs/smokecurrent/log***.log is written. What I am finding instead is that the data in ./logs/smokecurrent/log***.log is simply incomplete and defective. It appears that the data from the top log file breaks off in mid-line as its being written to the lower log file! I am attaching 4 gzipped files which demonstrate the problem: 2 from a Test::Smoke run on Linux; 2 from a run on FreeBSD. The problem appears on both platforms. ##### $ ls -l total 16 -rw-r--r-- 1 jkeenan jkeenan 2540 Aug 6 13:37 freebsd.smokecurrent.log.tail.gz -rw-r--r-- 1 jkeenan jkeenan 1837 Aug 6 13:37 freebsd.logs.smokecurrent.loge94c3f.log.tail.gz -rw-r--r-- 1 jkeenan jkeenan 2323 Aug 6 13:33 linode.smokecurrent.log.tail.gz -rw-r--r-- 1 jkeenan jkeenan 1671 Aug 6 13:33 linode.logs.smokecurrent.loge94c3f.log.tail.gz ##### For each of the 4 files provided I took the original file and deleted everything before the line containing "Finished smoking". On each platform the file whose name ends in "smokecurrent.log.tail.gz" is an extract from ./smokecurrent.log (the "top-level" log file); the file whose name ends in logs.smokecurrent.loge94c3f.log.tail.gz is an extract from ./logs/smokecurrent/loge94c3f6a8d515b39961010524037ca044f029524.log (the "lower-level" log file). In each platform-specific pair, you will observe that: * The top-level log file is complete, i.e., it displays all of the .json data, the URL of the report posting at perl5.test-smoke.org, and the archiver messages. * The lower-level report breaks off mid-line in the middle of the .json data. The URL of the report posting at perl5.test-smoke.org is never reached, nor are the archiver messages. This is a serious defect. Each test run overwrites ./smokecurrent.log, so the only data available for archiving for future study is that found in ./logs/smokecurrent/log***.log. If that data is incomplete, the human smoke-tester is not equipped to properly answer questions about particular Test::Smoke runs. Thank you very much. Jim Keenan
Subject: freebsd.logs.smokecurrent.loge94c3f.log.tail.gz

Message body not shown because it is not plain text.

Subject: freebsd.smokecurrent.log.tail.gz

Message body not shown because it is not plain text.

Subject: linode.logs.smokecurrent.loge94c3f.log.tail.gz

Message body not shown because it is not plain text.

Subject: linode.smokecurrent.log.tail.gz

Message body not shown because it is not plain text.

RT-Send-CC: public [...] khwilliamson.com
On Mon Aug 06 13:55:43 2018, JKEENAN wrote: Show quoted text
> As Test::Smoke run, it logs output in 'smokecurrent.log'. > > ##### > $ ls -l smokecurrent.log > -rw-r--r-- 1 jkeenan jkeenan 603104 Aug 4 21:56 smokecurrent.log > ##### > > At the conclusion of a run, it creates 4 files in the > logs/smokecurrent subdirectory, the names of which files are composed > from the SHA of the commit being tested. > > ##### > $ ls -ltr logs/smokecurrent/ | tail -4 > -rw-rw-rw- 1 jkeenan jkeenan 1454 Aug 4 21:56 > rpte94c3f6a8d515b39961010524037ca044f029524.rpt > -rw-rw-rw- 1 jkeenan jkeenan 2371 Aug 4 21:56 > oute94c3f6a8d515b39961010524037ca044f029524.out > -rw-rw-rw- 1 jkeenan jkeenan 595939 Aug 4 21:56 > loge94c3f6a8d515b39961010524037ca044f029524.log > -rw-rw-rw- 1 jkeenan jkeenan 6471 Aug 4 21:56 > jsne94c3f6a8d515b39961010524037ca044f029524.jsn > ##### > > "Naively" -- that is, as a typical user who has not peered deeply into > the Test::Smoke code base or documentation -- I would expect that > ./smokecurrent.log and ./logs/smokecurrent/log***.log to be either (a) > exactly the same file or (b) to be almost exactly the same but differ > in some clear, well thought-out way. By observation over repeated > Test::Smoke runs, I infer that (b) is usally the case and that some > data is deliberately and thoughtfully removed from ./smokecurrent.log > before ./logs/smokecurrent/log***.log is written. > > What I am finding instead is that the data in > ./logs/smokecurrent/log***.log is simply incomplete and defective. It > appears that the data from the top log file breaks off in mid-line as > its being written to the lower log file! > > I am attaching 4 gzipped files which demonstrate the problem: 2 from > a Test::Smoke run on Linux; 2 from a run on FreeBSD. The problem > appears on both platforms. > > ##### > $ ls -l > total 16 > -rw-r--r-- 1 jkeenan jkeenan 2540 Aug 6 13:37 > freebsd.smokecurrent.log.tail.gz > -rw-r--r-- 1 jkeenan jkeenan 1837 Aug 6 13:37 > freebsd.logs.smokecurrent.loge94c3f.log.tail.gz > -rw-r--r-- 1 jkeenan jkeenan 2323 Aug 6 13:33 > linode.smokecurrent.log.tail.gz > -rw-r--r-- 1 jkeenan jkeenan 1671 Aug 6 13:33 > linode.logs.smokecurrent.loge94c3f.log.tail.gz > ##### > > For each of the 4 files provided I took the original file and deleted > everything before the line containing "Finished smoking". On each > platform the file whose name ends in "smokecurrent.log.tail.gz" is an > extract from ./smokecurrent.log (the "top-level" log file); the file > whose name ends in logs.smokecurrent.loge94c3f.log.tail.gz is an > extract from > ./logs/smokecurrent/loge94c3f6a8d515b39961010524037ca044f029524.log > (the "lower-level" log file). > > In each platform-specific pair, you will observe that: > > * The top-level log file is complete, i.e., it displays all of the > .json data, the URL of the report posting at perl5.test-smoke.org, and > the archiver messages. > > * The lower-level report breaks off mid-line in the middle of the > .json data. The URL of the report posting at perl5.test-smoke.org is > never reached, nor are the archiver messages. > > This is a serious defect. Each test run overwrites > ./smokecurrent.log, so the only data available for archiving for > future study is that found in ./logs/smokecurrent/log***.log. If that > data is incomplete, the human smoke-tester is not equipped to properly > answer questions about particular Test::Smoke runs. > > Thank you very much. > Jim Keenan
On further inspection, I notice that a typical "top-level" ./smokecurrent.log file ends with text like this: ##### [2018-08-06 16:18:36-0400] http://perl5.test-smoke.org/report/68640 [2018-08-06 16:18:36-0400] ==> Starting archiver [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl-current/mktest.rpt, /home/jkeenan/p5smoke/logs/smokecurrent/rptd342516410da8af9d1824e50179d7d6ff3c7ba95.rpt): ok [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl-current/mktest.out, /home/jkeenan/p5smoke/logs/smokecurrent/outd342516410da8af9d1824e50179d7d6ff3c7ba95.out): ok [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl-current/mktest.jsn, /home/jkeenan/p5smoke/logs/smokecurrent/jsnd342516410da8af9d1824e50179d7d6ff3c7ba95.jsn): ok [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/smokecurrent.log, /home/jkeenan/p5smoke/logs/smokecurrent/logd342516410da8af9d1824e50179d7d6ff3c7ba95.log): ok ##### The last line is a message wherein we log the successful copy of smokecurrent.log to logs/smokecurrent/log***.log. But that's premature, as we're still writing to the file we want to copy. What is probably happening is that the a certain portion of the file is being copied at the point that the command to copy is issued, but the additional log messages created thereafter are not part of the copied text. archive_log() (lib/Test/Smoke/Archiver.pm line 194) is called by archive_files() (same file line 85), which in turn is called by run() (lib/Test/Smoke/App/Archiver.pm line 44), which is probably called by run() (lib/Test/Smoke/App/SmokePerl.pm line 114), which in turn is probably called by bin/tssmokeperl.pl. Perhaps archive_log() should not be called as part of Test::Smoke::Archiver::archive_files(). We should log a message like "Preparing to copy log file", close the handle writing to ./smokecurrent.log, copy the file and then *append* a line to ./logs/smokecurrent/log***.log indicating that we have successfully copied the log file. Granted, that would make a 1-line difference between the top-level and lower-level log files -- but that would be a minor flaw compared to what we currently have. Thank you very much. Jim Keenan
RT-Send-CC: public [...] khwilliamson.com
On Mon Aug 06 17:47:29 2018, JKEENAN wrote: Show quoted text
> On Mon Aug 06 13:55:43 2018, JKEENAN wrote:
> > As Test::Smoke run, it logs output in 'smokecurrent.log'. > > > > ##### > > $ ls -l smokecurrent.log > > -rw-r--r-- 1 jkeenan jkeenan 603104 Aug 4 21:56 smokecurrent.log > > ##### > > > > At the conclusion of a run, it creates 4 files in the > > logs/smokecurrent subdirectory, the names of which files are composed > > from the SHA of the commit being tested. > > > > ##### > > $ ls -ltr logs/smokecurrent/ | tail -4 > > -rw-rw-rw- 1 jkeenan jkeenan 1454 Aug 4 21:56 > > rpte94c3f6a8d515b39961010524037ca044f029524.rpt > > -rw-rw-rw- 1 jkeenan jkeenan 2371 Aug 4 21:56 > > oute94c3f6a8d515b39961010524037ca044f029524.out > > -rw-rw-rw- 1 jkeenan jkeenan 595939 Aug 4 21:56 > > loge94c3f6a8d515b39961010524037ca044f029524.log > > -rw-rw-rw- 1 jkeenan jkeenan 6471 Aug 4 21:56 > > jsne94c3f6a8d515b39961010524037ca044f029524.jsn > > ##### > > > > "Naively" -- that is, as a typical user who has not peered deeply > > into > > the Test::Smoke code base or documentation -- I would expect that > > ./smokecurrent.log and ./logs/smokecurrent/log***.log to be either > > (a) > > exactly the same file or (b) to be almost exactly the same but differ > > in some clear, well thought-out way. By observation over repeated > > Test::Smoke runs, I infer that (b) is usally the case and that some > > data is deliberately and thoughtfully removed from ./smokecurrent.log > > before ./logs/smokecurrent/log***.log is written. > > > > What I am finding instead is that the data in > > ./logs/smokecurrent/log***.log is simply incomplete and defective. > > It > > appears that the data from the top log file breaks off in mid-line as > > its being written to the lower log file! > > > > I am attaching 4 gzipped files which demonstrate the problem: 2 from > > a Test::Smoke run on Linux; 2 from a run on FreeBSD. The problem > > appears on both platforms. > > > > ##### > > $ ls -l > > total 16 > > -rw-r--r-- 1 jkeenan jkeenan 2540 Aug 6 13:37 > > freebsd.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1837 Aug 6 13:37 > > freebsd.logs.smokecurrent.loge94c3f.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 2323 Aug 6 13:33 > > linode.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1671 Aug 6 13:33 > > linode.logs.smokecurrent.loge94c3f.log.tail.gz > > ##### > > > > For each of the 4 files provided I took the original file and deleted > > everything before the line containing "Finished smoking". On each > > platform the file whose name ends in "smokecurrent.log.tail.gz" is an > > extract from ./smokecurrent.log (the "top-level" log file); the file > > whose name ends in logs.smokecurrent.loge94c3f.log.tail.gz is an > > extract from > > ./logs/smokecurrent/loge94c3f6a8d515b39961010524037ca044f029524.log > > (the "lower-level" log file). > > > > In each platform-specific pair, you will observe that: > > > > * The top-level log file is complete, i.e., it displays all of the > > .json data, the URL of the report posting at perl5.test-smoke.org, > > and > > the archiver messages. > > > > * The lower-level report breaks off mid-line in the middle of the > > .json data. The URL of the report posting at perl5.test-smoke.org is > > never reached, nor are the archiver messages. > > > > This is a serious defect. Each test run overwrites > > ./smokecurrent.log, so the only data available for archiving for > > future study is that found in ./logs/smokecurrent/log***.log. If > > that > > data is incomplete, the human smoke-tester is not equipped to > > properly > > answer questions about particular Test::Smoke runs. > > > > Thank you very much. > > Jim Keenan
> > On further inspection, I notice that a typical "top-level" > ./smokecurrent.log file ends with text like this: > > ##### > [2018-08-06 16:18:36-0400] http://perl5.test-smoke.org/report/68640 > [2018-08-06 16:18:36-0400] ==> Starting archiver > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.rpt, > /home/jkeenan/p5smoke/logs/smokecurrent/rptd342516410da8af9d1824e50179d7d6ff3c7ba95.rpt): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.out, > /home/jkeenan/p5smoke/logs/smokecurrent/outd342516410da8af9d1824e50179d7d6ff3c7ba95.out): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.jsn, > /home/jkeenan/p5smoke/logs/smokecurrent/jsnd342516410da8af9d1824e50179d7d6ff3c7ba95.jsn): > ok > [2018-08-06 16:18:36-0400] > Copy(/home/jkeenan/p5smoke/smokecurrent.log, > /home/jkeenan/p5smoke/logs/smokecurrent/logd342516410da8af9d1824e50179d7d6ff3c7ba95.log): > ok > ##### > > The last line is a message wherein we log the successful copy of > smokecurrent.log to logs/smokecurrent/log***.log. But that's > premature, as we're still writing to the file we want to copy. What > is probably happening is that the a certain portion of the file is > being copied at the point that the command to copy is issued, but the > additional log messages created thereafter are not part of the copied > text. > > archive_log() (lib/Test/Smoke/Archiver.pm line 194) is called by > archive_files() (same file line 85), which in turn is called by run() > (lib/Test/Smoke/App/Archiver.pm line 44), which is probably called by > run() (lib/Test/Smoke/App/SmokePerl.pm line 114), which in turn is > probably called by bin/tssmokeperl.pl. > > Perhaps archive_log() should not be called as part of > Test::Smoke::Archiver::archive_files(). We should log a message like > "Preparing to copy log file", close the handle writing to > ./smokecurrent.log, copy the file and then *append* a line to > ./logs/smokecurrent/log***.log indicating that we have successfully > copied the log file. Granted, that would make a 1-line difference > between the top-level and lower-level log files -- but that would be a > minor flaw compared to what we currently have. > > Thank you very much. > Jim Keenan
This problem should be solved with this pull request: https://github.com/abeltje/Test-Smoke/pull/43 I'm attaching a slightly reformatted version of the commit message to explain the cause of the problem and the solution. Please review. Am very open to suggestions for improvement but am also hoping that we can get a CPAN release soon so that this can be of benefit to smoke-testers and Perl 5 Porters during the current development cycle. Thank you very much. Jim Keenan
Subject: commit-message-20180817.txt
Ensure that entire content of smokecurrent.log is written to logs/smokecurrent/ In smoke-test configurations that only post reports via HTTP::Tiny (but perhaps in some or all other configurations as well), there were repeated observations that the content of smokecurrent.log was not entirely copied to logs/smokecurrent/log${SHA}.log. The content would be truncated near the end of the file. (When verbosity level 2 was requested, this truncation would often appear in the middle of a line.) smokecurrent.log is overwritten with each subsequent smoke-test run, so if it is not archived correctly we lack an accurate permanent record of what happened in a given smoke-test run. Analysis eventually showed that the source of the problem was that the copying of smokecurrent.log was being done while smokecurrent.log was still being written to by tssmokeperl.pl (as called from within smokecurrent.sh). So it was likely that some of the STDOUT and/or STDERR being redirected as tssmokeperl.pl was running was still buffered in memory and not yet written to disk. Hence, the copying of smokecurrent.log for archiving purposes was being done prematurely. The general outline of the solution is as follows: 1. Modify Test::Smoke::Archiver so it is no longer responsible for archiving smokecurrent.log. 2. When invoked by smokecurrent.sh, continue to have all of tssmokeperl.pl's STDOUT and STDERR redirected to smokecurrent.log. 3. Once tssmokeper.pl has completed, have smokecurrent.sh invoke a new program, tsarchivelog.pl, which composes a basename for logs/smokecurrent/log${SHA}.log and then copies smokecurrent.sh to that file. tsarchivelog.pl then appends one record to logs/smokecurrent/log${SHA}.log indicating that smokecurrent.log has been copied successfully. This commit implements the solution above. The source code files significantly modified include Makefile.PL, bin/configsmoke.pl, bin/tsarchivelog.pl (new), and lib/Test/Smoke/Archiver.pm. This commit does not attempt to implement any Windows-specific functionality, nor has it been tested on Windows, nor does it attempt to investigate whether there are any problems in the bin/*.pl programs other than those mentioned above. In the course of work on this issue, one new test file, t/app/101-archiver.t, has been added, along with dummy copy files used by that test. Certain housekeeping corrections were made to .gitignore and MANIFEST.SKIP and 'make manifest' regenerated the MANIFEST. Because the author had to install Test-Smoke in the course of working on this problem, $Test::Smoke::VERSION was implemented, though the maintainers may take a different approach.
On Mon Aug 06 17:47:29 2018, JKEENAN wrote: Show quoted text
> On Mon Aug 06 13:55:43 2018, JKEENAN wrote:
> > As Test::Smoke run, it logs output in 'smokecurrent.log'. > > > > ##### > > $ ls -l smokecurrent.log > > -rw-r--r-- 1 jkeenan jkeenan 603104 Aug 4 21:56 smokecurrent.log > > ##### > > > > At the conclusion of a run, it creates 4 files in the > > logs/smokecurrent subdirectory, the names of which files are composed > > from the SHA of the commit being tested. > > > > ##### > > $ ls -ltr logs/smokecurrent/ | tail -4 > > -rw-rw-rw- 1 jkeenan jkeenan 1454 Aug 4 21:56 > > rpte94c3f6a8d515b39961010524037ca044f029524.rpt > > -rw-rw-rw- 1 jkeenan jkeenan 2371 Aug 4 21:56 > > oute94c3f6a8d515b39961010524037ca044f029524.out > > -rw-rw-rw- 1 jkeenan jkeenan 595939 Aug 4 21:56 > > loge94c3f6a8d515b39961010524037ca044f029524.log > > -rw-rw-rw- 1 jkeenan jkeenan 6471 Aug 4 21:56 > > jsne94c3f6a8d515b39961010524037ca044f029524.jsn > > ##### > > > > "Naively" -- that is, as a typical user who has not peered deeply > > into > > the Test::Smoke code base or documentation -- I would expect that > > ./smokecurrent.log and ./logs/smokecurrent/log***.log to be either > > (a) > > exactly the same file or (b) to be almost exactly the same but differ > > in some clear, well thought-out way. By observation over repeated > > Test::Smoke runs, I infer that (b) is usally the case and that some > > data is deliberately and thoughtfully removed from ./smokecurrent.log > > before ./logs/smokecurrent/log***.log is written. > > > > What I am finding instead is that the data in > > ./logs/smokecurrent/log***.log is simply incomplete and defective. > > It > > appears that the data from the top log file breaks off in mid-line as > > its being written to the lower log file! > > > > I am attaching 4 gzipped files which demonstrate the problem: 2 from > > a Test::Smoke run on Linux; 2 from a run on FreeBSD. The problem > > appears on both platforms. > > > > ##### > > $ ls -l > > total 16 > > -rw-r--r-- 1 jkeenan jkeenan 2540 Aug 6 13:37 > > freebsd.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1837 Aug 6 13:37 > > freebsd.logs.smokecurrent.loge94c3f.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 2323 Aug 6 13:33 > > linode.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1671 Aug 6 13:33 > > linode.logs.smokecurrent.loge94c3f.log.tail.gz > > ##### > > > > For each of the 4 files provided I took the original file and deleted > > everything before the line containing "Finished smoking". On each > > platform the file whose name ends in "smokecurrent.log.tail.gz" is an > > extract from ./smokecurrent.log (the "top-level" log file); the file > > whose name ends in logs.smokecurrent.loge94c3f.log.tail.gz is an > > extract from > > ./logs/smokecurrent/loge94c3f6a8d515b39961010524037ca044f029524.log > > (the "lower-level" log file). > > > > In each platform-specific pair, you will observe that: > > > > * The top-level log file is complete, i.e., it displays all of the > > .json data, the URL of the report posting at perl5.test-smoke.org, > > and > > the archiver messages. > > > > * The lower-level report breaks off mid-line in the middle of the > > .json data. The URL of the report posting at perl5.test-smoke.org is > > never reached, nor are the archiver messages. > > > > This is a serious defect. Each test run overwrites > > ./smokecurrent.log, so the only data available for archiving for > > future study is that found in ./logs/smokecurrent/log***.log. If > > that > > data is incomplete, the human smoke-tester is not equipped to > > properly > > answer questions about particular Test::Smoke runs. > > > > Thank you very much. > > Jim Keenan
> > On further inspection, I notice that a typical "top-level" > ./smokecurrent.log file ends with text like this: > > ##### > [2018-08-06 16:18:36-0400] http://perl5.test-smoke.org/report/68640 > [2018-08-06 16:18:36-0400] ==> Starting archiver > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.rpt, > /home/jkeenan/p5smoke/logs/smokecurrent/rptd342516410da8af9d1824e50179d7d6ff3c7ba95.rpt): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.out, > /home/jkeenan/p5smoke/logs/smokecurrent/outd342516410da8af9d1824e50179d7d6ff3c7ba95.out): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.jsn, > /home/jkeenan/p5smoke/logs/smokecurrent/jsnd342516410da8af9d1824e50179d7d6ff3c7ba95.jsn): > ok > [2018-08-06 16:18:36-0400] > Copy(/home/jkeenan/p5smoke/smokecurrent.log, > /home/jkeenan/p5smoke/logs/smokecurrent/logd342516410da8af9d1824e50179d7d6ff3c7ba95.log): > ok > ##### > > The last line is a message wherein we log the successful copy of > smokecurrent.log to logs/smokecurrent/log***.log. But that's > premature, as we're still writing to the file we want to copy. What > is probably happening is that the a certain portion of the file is > being copied at the point that the command to copy is issued, but the > additional log messages created thereafter are not part of the copied > text. > > archive_log() (lib/Test/Smoke/Archiver.pm line 194) is called by > archive_files() (same file line 85), which in turn is called by run() > (lib/Test/Smoke/App/Archiver.pm line 44), which is probably called by > run() (lib/Test/Smoke/App/SmokePerl.pm line 114), which in turn is > probably called by bin/tssmokeperl.pl. > > Perhaps archive_log() should not be called as part of > Test::Smoke::Archiver::archive_files(). We should log a message like > "Preparing to copy log file", close the handle writing to > ./smokecurrent.log, copy the file and then *append* a line to > ./logs/smokecurrent/log***.log indicating that we have successfully > copied the log file. Granted, that would make a 1-line difference > between the top-level and lower-level log files -- but that would be a > minor flaw compared to what we currently have. > > Thank you very much. > Jim Keenan
Abe: Can you indicate (here in this ticket) what steps, if any, you have taken in Test-Smoke-1.72_06 to address this problem? I'm examining that beta and want to know what I should be looking for? Thank you very much. Jim Keenan
On Mon Aug 06 17:47:29 2018, JKEENAN wrote: Show quoted text
> On Mon Aug 06 13:55:43 2018, JKEENAN wrote:
> > As Test::Smoke run, it logs output in 'smokecurrent.log'. > > > > ##### > > $ ls -l smokecurrent.log > > -rw-r--r-- 1 jkeenan jkeenan 603104 Aug 4 21:56 smokecurrent.log > > ##### > > > > At the conclusion of a run, it creates 4 files in the > > logs/smokecurrent subdirectory, the names of which files are composed > > from the SHA of the commit being tested. > > > > ##### > > $ ls -ltr logs/smokecurrent/ | tail -4 > > -rw-rw-rw- 1 jkeenan jkeenan 1454 Aug 4 21:56 > > rpte94c3f6a8d515b39961010524037ca044f029524.rpt > > -rw-rw-rw- 1 jkeenan jkeenan 2371 Aug 4 21:56 > > oute94c3f6a8d515b39961010524037ca044f029524.out > > -rw-rw-rw- 1 jkeenan jkeenan 595939 Aug 4 21:56 > > loge94c3f6a8d515b39961010524037ca044f029524.log > > -rw-rw-rw- 1 jkeenan jkeenan 6471 Aug 4 21:56 > > jsne94c3f6a8d515b39961010524037ca044f029524.jsn > > ##### > > > > "Naively" -- that is, as a typical user who has not peered deeply > > into > > the Test::Smoke code base or documentation -- I would expect that > > ./smokecurrent.log and ./logs/smokecurrent/log***.log to be either > > (a) > > exactly the same file or (b) to be almost exactly the same but differ > > in some clear, well thought-out way. By observation over repeated > > Test::Smoke runs, I infer that (b) is usally the case and that some > > data is deliberately and thoughtfully removed from ./smokecurrent.log > > before ./logs/smokecurrent/log***.log is written. > > > > What I am finding instead is that the data in > > ./logs/smokecurrent/log***.log is simply incomplete and defective. > > It > > appears that the data from the top log file breaks off in mid-line as > > its being written to the lower log file! > > > > I am attaching 4 gzipped files which demonstrate the problem: 2 from > > a Test::Smoke run on Linux; 2 from a run on FreeBSD. The problem > > appears on both platforms. > > > > ##### > > $ ls -l > > total 16 > > -rw-r--r-- 1 jkeenan jkeenan 2540 Aug 6 13:37 > > freebsd.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1837 Aug 6 13:37 > > freebsd.logs.smokecurrent.loge94c3f.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 2323 Aug 6 13:33 > > linode.smokecurrent.log.tail.gz > > -rw-r--r-- 1 jkeenan jkeenan 1671 Aug 6 13:33 > > linode.logs.smokecurrent.loge94c3f.log.tail.gz > > ##### > > > > For each of the 4 files provided I took the original file and deleted > > everything before the line containing "Finished smoking". On each > > platform the file whose name ends in "smokecurrent.log.tail.gz" is an > > extract from ./smokecurrent.log (the "top-level" log file); the file > > whose name ends in logs.smokecurrent.loge94c3f.log.tail.gz is an > > extract from > > ./logs/smokecurrent/loge94c3f6a8d515b39961010524037ca044f029524.log > > (the "lower-level" log file). > > > > In each platform-specific pair, you will observe that: > > > > * The top-level log file is complete, i.e., it displays all of the > > .json data, the URL of the report posting at perl5.test-smoke.org, > > and > > the archiver messages. > > > > * The lower-level report breaks off mid-line in the middle of the > > .json data. The URL of the report posting at perl5.test-smoke.org is > > never reached, nor are the archiver messages. > > > > This is a serious defect. Each test run overwrites > > ./smokecurrent.log, so the only data available for archiving for > > future study is that found in ./logs/smokecurrent/log***.log. If > > that > > data is incomplete, the human smoke-tester is not equipped to > > properly > > answer questions about particular Test::Smoke runs. > > > > Thank you very much. > > Jim Keenan
> > On further inspection, I notice that a typical "top-level" > ./smokecurrent.log file ends with text like this: > > ##### > [2018-08-06 16:18:36-0400] http://perl5.test-smoke.org/report/68640 > [2018-08-06 16:18:36-0400] ==> Starting archiver > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.rpt, > /home/jkeenan/p5smoke/logs/smokecurrent/rptd342516410da8af9d1824e50179d7d6ff3c7ba95.rpt): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.out, > /home/jkeenan/p5smoke/logs/smokecurrent/outd342516410da8af9d1824e50179d7d6ff3c7ba95.out): > ok > [2018-08-06 16:18:36-0400] Copy(/home/jkeenan/p5smoke/perl- > current/mktest.jsn, > /home/jkeenan/p5smoke/logs/smokecurrent/jsnd342516410da8af9d1824e50179d7d6ff3c7ba95.jsn): > ok > [2018-08-06 16:18:36-0400] > Copy(/home/jkeenan/p5smoke/smokecurrent.log, > /home/jkeenan/p5smoke/logs/smokecurrent/logd342516410da8af9d1824e50179d7d6ff3c7ba95.log): > ok > ##### > > The last line is a message wherein we log the successful copy of > smokecurrent.log to logs/smokecurrent/log***.log. But that's > premature, as we're still writing to the file we want to copy. What > is probably happening is that the a certain portion of the file is > being copied at the point that the command to copy is issued, but the > additional log messages created thereafter are not part of the copied > text. > > archive_log() (lib/Test/Smoke/Archiver.pm line 194) is called by > archive_files() (same file line 85), which in turn is called by run() > (lib/Test/Smoke/App/Archiver.pm line 44), which is probably called by > run() (lib/Test/Smoke/App/SmokePerl.pm line 114), which in turn is > probably called by bin/tssmokeperl.pl. > > Perhaps archive_log() should not be called as part of > Test::Smoke::Archiver::archive_files(). We should log a message like > "Preparing to copy log file", close the handle writing to > ./smokecurrent.log, copy the file and then *append* a line to > ./logs/smokecurrent/log***.log indicating that we have successfully > copied the log file. Granted, that would make a 1-line difference > between the top-level and lower-level log files -- but that would be a > minor flaw compared to what we currently have. >
Today I have been reviewing Test-Smoke-1.72_06. ##### [install] $ ack '^\$VERSION' ../Test-Smoke-1.72_06/lib/Test/Smoke.pm $VERSION = '1.72_06'; ##### I believe that it has not corrected the problem of a truncated logs/smokecurrent/log*.log file reported here in RT 126010. I conducted 4 smoke test runs using 1.72_06. If you go to http://perl5.test-smoke.org/search and enter 'freebsd' for OS and 'perlmonger.nycbug.org' for host, the 4 runs executed on 2018-11-14 are the runs in question. ##### Arch Host OS Git-id Branch Smoke date Status URL amd64 perlmonger.nycbug.org freebsd - 11.2-STABLE v5.28.0-47-g94027dc85b maint-5.28 2018-11-14 21:47:08+00 PASS http://perl5.test-smoke.org/report/74039 amd64 perlmonger.nycbug.org freebsd - 11.2-STABLE v5.26.2-85-gadf1b5d28a maint-5.26 2018-11-14 21:11:06+00 PASS http://perl5.test-smoke.org/report/74035 amd64 perlmonger.nycbug.org freebsd - 11.2-STABLE v5.29.4-42-g61edc94a36 blead 2018-11-14 17:33:51+00 FAIL(F) http://perl5.test-smoke.org/report/74026 amd64 perlmonger.nycbug.org freebsd - 11.2-STABLE v5.29.1-31-gd3bb03e084 undefined 2018-11-14 15:26:08+00 FAIL(F) http://perl5.test-smoke.org/report/74020 ##### (Also in .csv attachment.) (The "Status" grades are not relevant, as there were test failures in blead over the course of the day.) (The "undefined" value for Branch in v5.29.1-31-gd3bb03e084 is not very relevant, as that was a commit I checked out to test https://rt.cpan.org/Ticket/Display.html?id=125932.) I am attaching the tails of the 'smokecurrent.log' files for each of these 4 smoke-test runs. For convenience, I'm also attaching a file which cats those 4 tails into a single file, as well as the perl -V output for the last of the 4 smoke-test runs (maint-5.28). Note that in each of these tails there is no line logging the "Copy" of smokecurrent.log as "logs/smokecurrent/log*.log". So the archived version of the log is therefore incomplete. The pre-archived smokecurrent.log file does, however, have a "Copy" line in it. Example: ##### [install] $ diff smokecurrent.log logs/smokecurrent/log94027dc85b896cfea4bd017b273f22e472882d68.log |less 1483d1482 < [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/smokecurrent.log, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/log94027dc85b896cfea4bd017b273f22e472882d68.log): ok ##### Thank you very much. Jim Keenan
Subject: 4_log_tails.txt
61edc94a.log.tail.txt [2018-11-14 18:03:55+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 18:03:55+0000] ==> Starting sendreport [2018-11-14 18:03:55+0000] ==> Skipping mailer [2018-11-14 18:03:55+0000] ==> Starting poster [2018-11-14 18:03:55+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 18:03:57+0000] https://perl5.test-smoke.org/report/74026 [2018-11-14 18:03:57+0000] ==> Starting archiver [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rpt61edc94a361bb37ba6c7541ee4bf61130c98c1d7.rpt): ok [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/out61edc94a361bb37ba6c7541ee4bf61130c98c1d7.out): ok [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsn61edc94a361bb37ba6c7541ee4bf61130c98c1d7.jsn): ok 94027dc8.log.tail.txt [2018-11-14 22:16:15+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 22:16:15+0000] ==> Starting sendreport [2018-11-14 22:16:15+0000] ==> Skipping mailer [2018-11-14 22:16:15+0000] ==> Starting poster [2018-11-14 22:16:15+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 22:16:16+0000] https://perl5.test-smoke.org/report/74039 [2018-11-14 22:16:16+0000] ==> Starting archiver [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rpt94027dc85b896cfea4bd017b273f22e472882d68.rpt): ok [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/out94027dc85b896cfea4bd017b273f22e472882d68.out): ok [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsn94027dc85b896cfea4bd017b273f22e472882d68.jsn): ok adf1b5d2.log.tail.txt [2018-11-14 21:39:23+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 21:39:23+0000] ==> Starting sendreport [2018-11-14 21:39:23+0000] ==> Skipping mailer [2018-11-14 21:39:23+0000] ==> Starting poster [2018-11-14 21:39:23+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 21:39:24+0000] https://perl5.test-smoke.org/report/74035 [2018-11-14 21:39:24+0000] ==> Starting archiver [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rptadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.rpt): ok [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/outadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.out): ok [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsnadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.jsn): ok d3bb03e0.log.tail.txt [2018-11-14 15:55:46+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 15:55:46+0000] ==> Starting sendreport [2018-11-14 15:55:46+0000] ==> Skipping mailer [2018-11-14 15:55:46+0000] ==> Starting poster [2018-11-14 15:55:46+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 15:55:47+0000] https://perl5.test-smoke.org/report/74020 [2018-11-14 15:55:47+0000] ==> Starting archiver [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rptd3bb03e0842cd3214922ab134a48314ee3fe2077.rpt): ok [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/outd3bb03e0842cd3214922ab134a48314ee3fe2077.out): ok [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsnd3bb03e0842cd3214922ab134a48314ee3fe2077.jsn): ok
Subject: 61edc94a.log.tail.txt
[2018-11-14 18:03:55+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 18:03:55+0000] ==> Starting sendreport [2018-11-14 18:03:55+0000] ==> Skipping mailer [2018-11-14 18:03:55+0000] ==> Starting poster [2018-11-14 18:03:55+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 18:03:57+0000] https://perl5.test-smoke.org/report/74026 [2018-11-14 18:03:57+0000] ==> Starting archiver [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rpt61edc94a361bb37ba6c7541ee4bf61130c98c1d7.rpt): ok [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/out61edc94a361bb37ba6c7541ee4bf61130c98c1d7.out): ok [2018-11-14 18:03:57+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsn61edc94a361bb37ba6c7541ee4bf61130c98c1d7.jsn): ok
Subject: 94027dc8.log.tail.txt
[2018-11-14 22:16:15+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 22:16:15+0000] ==> Starting sendreport [2018-11-14 22:16:15+0000] ==> Skipping mailer [2018-11-14 22:16:15+0000] ==> Starting poster [2018-11-14 22:16:15+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 22:16:16+0000] https://perl5.test-smoke.org/report/74039 [2018-11-14 22:16:16+0000] ==> Starting archiver [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rpt94027dc85b896cfea4bd017b273f22e472882d68.rpt): ok [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/out94027dc85b896cfea4bd017b273f22e472882d68.out): ok [2018-11-14 22:16:16+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsn94027dc85b896cfea4bd017b273f22e472882d68.jsn): ok
Subject: adf1b5d2.log.tail.txt
[2018-11-14 21:39:23+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 21:39:23+0000] ==> Starting sendreport [2018-11-14 21:39:23+0000] ==> Skipping mailer [2018-11-14 21:39:23+0000] ==> Starting poster [2018-11-14 21:39:23+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 21:39:24+0000] https://perl5.test-smoke.org/report/74035 [2018-11-14 21:39:24+0000] ==> Starting archiver [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rptadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.rpt): ok [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/outadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.out): ok [2018-11-14 21:39:24+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsnadf1b5d28a4fd0e20bbbb72cab062fdd520cb065.jsn): ok
Subject: d3bb03e0.log.tail.txt
[2018-11-14 15:55:46+0000] Write to '/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn': ok [2018-11-14 15:55:46+0000] ==> Starting sendreport [2018-11-14 15:55:46+0000] ==> Skipping mailer [2018-11-14 15:55:46+0000] ==> Starting poster [2018-11-14 15:55:46+0000] Posting to https://perl5.test-smoke.org/report via HTTP::Tiny. [2018-11-14 15:55:47+0000] https://perl5.test-smoke.org/report/74020 [2018-11-14 15:55:47+0000] ==> Starting archiver [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.rpt, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/rptd3bb03e0842cd3214922ab134a48314ee3fe2077.rpt): ok [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.out, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/outd3bb03e0842cd3214922ab134a48314ee3fe2077.out): ok [2018-11-14 15:55:47+0000] Copy(/usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/perl-current/mktest.jsn, /usr/home/jkeenan/learn/perl/test-smoke/rtc-125932/install/logs/smokecurrent/jsnd3bb03e0842cd3214922ab134a48314ee3fe2077.jsn): ok
Subject: perl_V.maint-5.28.txt
Summary of my perl5 (revision 5 version 28 subversion 1) configuration: Snapshot of: 94027dc85b896cfea4bd017b273f22e472882d68 Platform: osname=freebsd osvers=11.2-stable archname=amd64-freebsd-thread-multi uname='freebsd perlmonger.nycbug.org 11.2-stable freebsd 11.2-stable #0 r339445: sat oct 20 00:08:11 utc 2018 root@perlmonger.nycbug.org:usrobjusrsrcsysgeneric amd64 ' config_args='-des -Dusedevel -Duseithreads -Doptimize=-O2 -pipe -fstack-protector -fno-strict-aliasing -Dcc=clang -Qunused-arguments' hint=recommended useposix=true d_sigaction=define useithreads=define usemultiplicity=define use64bitint=define use64bitall=define uselongdouble=undef usemymalloc=n default_inc_excludes_dot=define bincompat5005=undef Compiler: cc='clang -Qunused-arguments' ccflags ='-DHAS_FPSETMASK -DHAS_FLOATINGPOINT_H -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include -D_FORTIFY_SOURCE=2' optimize='-O2 -pipe -fstack-protector -fno-strict-aliasing' cppflags='-DHAS_FPSETMASK -DHAS_FLOATINGPOINT_H -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include' ccversion='' gccversion='4.2.1 Compatible FreeBSD Clang 6.0.1 (tags/RELEASE_601/final 335540)' gccosandvers='' intsize=4 longsize=8 ptrsize=8 doublesize=8 byteorder=12345678 doublekind=3 d_longlong=define longlongsize=8 d_longdbl=define longdblsize=16 longdblkind=3 ivtype='long' ivsize=8 nvtype='double' nvsize=8 Off_t='off_t' lseeksize=8 alignbytes=8 prototype=define Linker and Libraries: ld='clang -Qunused-arguments' ldflags ='-pthread -Wl,-E -fstack-protector-strong -L/usr/local/lib' libpth=/usr/lib /usr/local/lib /usr/lib/clang/6.0.1/lib /usr/lib libs=-lpthread -lgdbm -ldl -lm -lcrypt -lutil perllibs=-lpthread -ldl -lm -lcrypt -lutil libc= so=so useshrplib=false libperl=libperl.a gnulibc_version='' Dynamic Linking: dlsrc=dl_dlopen.xs dlext=so d_dlsymun=undef ccdlflags=' ' cccdlflags='-DPIC -fPIC' lddlflags='-shared -L/usr/local/lib -fstack-protector-strong' Characteristics of this binary (from libperl): Compile-time options: HAS_TIMES MULTIPLICITY PERLIO_LAYERS PERL_COPY_ON_WRITE PERL_DONT_CREATE_GVSV PERL_IMPLICIT_CONTEXT PERL_MALLOC_WRAP PERL_OP_PARENT PERL_PRESERVE_IVUV PERL_USE_DEVEL USE_64_BIT_ALL USE_64_BIT_INT USE_ITHREADS USE_LARGE_FILES USE_LOCALE USE_LOCALE_COLLATE USE_LOCALE_CTYPE USE_LOCALE_NUMERIC USE_LOCALE_TIME USE_PERLIO USE_PERL_ATOF USE_REENTRANT_API Locally applied patches: SMOKE94027dc85b896cfea4bd017b273f22e472882d68 Built under freebsd Compiled at Nov 14 2018 22:05:05 %ENV: PERL2DIR="/home/jkeenan/gitwork/perl2" PERL_WORKDIR="/home/jkeenan/gitwork/perl" @INC: lib /usr/local/lib/perl5/site_perl/5.28.1/amd64-freebsd-thread-multi /usr/local/lib/perl5/site_perl/5.28.1 /usr/local/lib/perl5/5.28.1/amd64-freebsd-thread-multi /usr/local/lib/perl5/5.28.1
Subject: smoke-tests-with-1.72_06.csv
"Arch","Host","OS","Git-id","Branch","Smoke date","Status","URL" "amd64","perlmonger.nycbug.org","freebsd - 11.2-STABLE","v5.28.0-47-g94027dc85b","maint-5.28","2018-11-14 21:47:08+00","PASS","http://perl5.test-smoke.org/report/74039" "amd64","perlmonger.nycbug.org","freebsd - 11.2-STABLE","v5.26.2-85-gadf1b5d28a","maint-5.26","2018-11-14 21:11:06+00","PASS","http://perl5.test-smoke.org/report/74035" "amd64","perlmonger.nycbug.org","freebsd - 11.2-STABLE","v5.29.4-42-g61edc94a36","blead","2018-11-14 17:33:51+00","FAIL(F)","http://perl5.test-smoke.org/report/74026" "amd64","perlmonger.nycbug.org","freebsd - 11.2-STABLE","v5.29.1-31-gd3bb03e084","undefined","2018-11-14 15:26:08+00","FAIL(F)","http://perl5.test-smoke.org/report/74020"