eserte / cpan-testers-matrix Goto Github PK

Can you tell what is the reason that 5.20 and 5.20-RC1 are thrown into the same bucket? Could this be changed? Would it be Barbie's end or this end?

I think in general one would prefer to have them separated. WDYT?

Please check perlancar/perl-Perinci-Sub-Complete#1

The current Perinci::Sub::Complete (0.943) requires Complete::Getopt::Long 0.479, which doesn't exists (latest current version is 0.478).
See https://metacpan.org/source/PERLANCAR/Perinci-Sub-Complete-0.943/Makefile.PL and https://metacpan.org/release/Complete-Getopt-Long.
As a result, the Perinci::Sub::Complete can't be installed in fresh environment (for example, new Perl install via plenv).
However, testers matrix is almost green: http://matrix.cpantesters.org/?dist=Perinci-Sub-Complete+0.943

And according to logs, the strict requirement of 0.479 was ignored and 0.478 was used, for example:
http://www.cpantesters.org/cpan/report/b87d2228-b808-11ea-a319-193bcccccfd6
http://www.cpantesters.org/cpan/report/8a7d103e-b805-11ea-9066-e374b0ba08e8

I'm not sure if this bug in cpan-testers-matrix or somewhere else, will be glad if you confirm or point me to the right direction.

Today http://matrix.cpantesters.org/?dist=File-AptFetch shows the headline

CPAN Testers Matrix: File-AptFetch v0.1.5

and at the bottom of the table below "Other Versions" in red:

NOTE: no report for latest version 0.1.5

I seem to remember that this was already fixed a long time ago, so maybe this is just a new manifestation.

I've released Net::Cisco::FMC::v1 to CPAN and checked CPANTesters but the URL http://matrix.cpantesters.org/?dist=Net-Cisco-FMC-v1 gives an error page which reads:

An error was encountered:
Distribution results for <Net-Cisco-FMC> at http://www.cpantesters.org/show/Net-Cisco-FMC.json not found.

Passing the version too works: http://matrix.cpantesters.org/?dist=Net-Cisco-FMC-v1%200.001001

It looks like if the -v1 is cut off.

The current version of the CPAN Matrix leaves modules that have been removed from CPAN on the author page. While I do agree the test results should not be thrown away, I think they're only distraction on that page. Hence I would like them to be removed from there.

I've long had this error:

As have anyone who has a large number of distributions.

Its really annoying and basically renders the matrix significantly less useful.

Some people have in excess of 60Mb JSON files.

Right now those three use the same color by default, but in my development of OpenGL.pm, a module that has a lot of fragility in the makefile, being able to use signaling beyond FAIL/PASS is important. Could we change the default settings to differentiate the three by using different colors?

A temporary fix i'm using right now is to use UserCSS to apply the following:

.action_UNKNOWN { background:coral !important; } .action_INVALID { background:tomato !important; }

With these results:

This happened when I clicked on the CPAN Testers Author Matrix link for Jens Rehsack (http://search.cpan.org/~rehsack/).

Changing the stylesheet preference seems to be broken due to the style tag on cpantestersmatrix.pl being commented out. (

Example:

  <style type="text/css"><!--
  .action_PASS    { background:#00ff00; }
  .action_NA      { background:#0000c0; }
  .action_UNKNOWN { background:#0000c0; }
  .action_FAIL    { background:#800000; }
  .action_INVALID { background:#0000c0; }

  .maxver_PASSNEW { background:green;      }
  .maxver_PASSANY { background:lightgreen; }
  .maxver_NONE    { background:red;        }  

  .fgaction_PASS    { color:green;  }
  .fgaction_NA      { color:orange; }
  .fgaction_UNKNOWN { color:orange; }
  .fgaction_FAIL    { color:red;    }
  .fgaction_INVALID { color:orange; }

  table		  { border-collapse:collapse; }
  th,td           { border:1px solid black; }
  th                { padding: 0 }
  body		  { font-family:sans-serif; }
  form            { display: inline-block; margin: 0 0 1em 0; }

  .search_container { margin: 0 2em; }
  .bt th,td	  { border:none; height:2.2ex; }

  .reports th	  { border:2px solid black; padding-left:3px; padding-right:3px; }
  .reports td	  { border:1px solid black; padding-left:3px; padding-right:3px; }

  .warn           { color:red; font-weight:bold; }
  .warn a:link    { color:red; font-weight:bold; }
  .warn a:visited { color:red; font-weight:bold; }
  .warn a:hover   { color:red; font-weight:bold; }
  .warn a:active  { color:red; font-weight:bold; }
  .sml            { font-size: x-small; }
  .unimpt         { font-size: smaller; }

  th>a            { text-decoration: none; padding: 1px 5px; display: inline-block; min-width: 70px; margin: 0; }
  th>a:hover      { background-color: blue; color: white; }

  td.action_PASS:hover  { background-color: #050; }

  h1>a            { color:black; text-decoration: none; }

  div.downtime_teaser { float:right; font-size:x-small; background-color:#fffff0; color:#000000; border: 1px solid black; -moz-border-radius:10px; border-radius:10px; padding:10px; }
  --></style>

While the number of passed tests under Stats is great, it would also be useful (to me at least) to have the number of different test environments summarized in an extra line. That would be the number of non empty cells in the matrix.

Thanks, hexcoder

Don't know if this is the best place, but the site doesn't seem to be working eg: http://matrix.cpantesters.org/?author=IVANWILLS

Starting at Net::DNS test result matrix

http://matrix.cpantesters.org/?dist=Net-DNS+0.75

Follow link to "Max version with a PASS"

Observe: Leftmost column shows a pass for 5.18.2, but the OS name is not shown at head of column.

Observe: Bottom row shows Perl rev 0, and at least one test failure in the openbsd column.

I'm using stylesheet gradient which I find very useful, and when I close the browser the testers matrix forgets this preference, and I have to set it again.
Can you add an expire date so the cookie survives browser closing?

The matrix page for BioPerl-1.6.923 sports the following note in red under Other Versions: "NOTE: no report for latest version 1.006923". It appears as if the conditional generating that note is comparing a non-numified list of report versions to the latest numified version.

Here's a screenshot:

Looking at test report matrix for Net::DNS 1.18, I was astonished to see the last report shown 7 hours into the future.

First report: 2018-09-21 15:52 BST (31 days ago)
Last report: 2018-10-23 01:26 BST
PASS: 1211
Number of tested configurations: 215

Current time is: Mon 22 Oct 18:18:38 BST 2018

The fault also appears in the table under the "ALL" link:

AFAICT this is not the only example.

Would be nice to be able to use a badge for the github, like from travis. With test matrix packed to the image. Like cell->pixel or smth. With the link to the report page.

Today I found on http://matrix.cpantesters.org/?dist=PDL the version 2.005 and the red note:

NOTE: no report for latest version v2.4.11

In the list of found versions on that page one finds the link to the newest version:

http://matrix.cpantesters.org/?dist=PDL%202.4.11

On that page I also see the red note "no report for latest version v2.4.11."

Neither on the downloaded files PDL.yaml and PDL.json vontain the "v" that gets in the way here.

On page: http://matrix.cpantesters.org/?dist=Math-Disarrange-List%2020170828

(a) Clicking on: by Slaven Rezić produces "problems loading page"

(b) Please consider making the title at the top of the page clickable to the corresponding version of the module as this would allow me to send this page to people to demonstrate the module quality (first) and then the actual module content (second).

Thanks!