Failing Ecosystem Modules List about ecosystem

[zoffixznet]

FWIW, I've built ecosystem toaster we'll be using to test for any regression in pre-release commits. https://toast.perl6.party/ (still need to add a way to view STDERR/OUT of install runs; though data is already in the SQLite DB you can download).

Could be used to see which modules fail.

[Altai-man]

I have removed some modules out:

• DateTime::Math - fixed.
• DateTime::Utils - is not broken.
• if - fixed.
• Tardis - was removed.
• File::Ignore - was fixed.
• Druid - was fixed.
• Devel::Trace - was removed.

[pierre-vigier]

I Fixed my module Math::Matrix, a change in behaviour of Seq/List broke it. It's fixed now.
I added the license in the META file as well

[bbkr]

GeoIP::City is deprecated, new module is https://github.com/bbkr/GeoIP2.

The reason is that GeoIP::City supports old binary database formats that will be phased out in early 2019, but they are already not accessible for download from Maxmind site. And that's why tests fail.

I want to keep this module in ecosystem till phase out so I'll try to fix tests to use fixed database versions. Thanks for pointing this out.

[zoffixznet]

sena_kun buggable, eco Tardis
sena_kun, Nothing found
Is it even here? I cannot find it, though I remember such a module.

Tardis was removed from ecosystem: 74a4b9d

[teodozjan]

Pod::Coverage is failing because rakudo has changed and I don't know gow to align it to new behaviour

[zoffixznet]

Pod::Coverage is failing because rakudo has changed and I don't know gow to align it to new behaviour

Sent a small PR fixing one problem. This diff fixes half of the other problems, and currently I'm at "crashes in Pod::Coverage::Full.check's require because lib is not set" stage. Somewhere there it's missing use lib <lib> or -Ilib. Maybe you'd know where, since you know codebase better.

zoffix@leliana:/tmp/tmp.uDxsRRfX5p$ gd
diff --git a/lib/Pod/Coverage/Anypod.pm6 b/lib/Pod/Coverage/Anypod.pm6
index 47724ce..d4687b8 100644
--- a/lib/Pod/Coverage/Anypod.pm6
+++ b/lib/Pod/Coverage/Anypod.pm6
@@ -25,7 +25,7 @@ method !file-haspod {
 
 sub has-pod($path) returns Bool {
     #dd qqx/$*EXECUTABLE-NAME --doc $path/.lines.elems;
-    $path.IO ~~ :f and qqx/$*EXECUTABLE-NAME --doc $path/.lines.elems.Bool;
+    $path.IO ~~ :f and qqx/$*EXECUTABLE-NAME --Ilib --doc $path/.lines.elems.Bool;
 }
 
 
diff --git a/lib/Pod/Coverage/Full.pm6 b/lib/Pod/Coverage/Full.pm6
index 45f4bcf..bf3e87e 100644
--- a/lib/Pod/Coverage/Full.pm6
+++ b/lib/Pod/Coverage/Full.pm6
@@ -137,9 +137,9 @@ method correct-pod($filename) {
 }
 
 sub read_pod($filename){
-    say "$*EXECUTABLE-NAME --doc=Keywords $filename";
-    dd qqx/$*EXECUTABLE-NAME --doc=Keywords $filename/;
-    return qqx/$*EXECUTABLE-NAME --doc=Keywords $filename/.lines;
+ #   say "$*EXECUTABLE-NAME --doc=Keywords $filename";
+#    dd qqx/$*EXECUTABLE-NAME --doc=Keywords $filename/;
+    return qqx/$*EXECUTABLE-NAME --Ilib --doc=Keywords $filename/.lines;
 }
 
 sub routine-result($what){
diff --git a/t/selftest.t b/t/selftest.t
index cb67ab4..18d3a53 100644
--- a/t/selftest.t
+++ b/t/selftest.t
@@ -5,5 +5,5 @@ use Test::Coverage;
 plan 3;
 
 ok 1, "At least loads";
-subtest_anypod_ok('META.info');
-subtest_coverage_ok('META.info');
+subtest_anypod_ok('META6.json');
+subtest_coverage_ok('META6.json');
zoffix@leliana:/tmp/tmp.uDxsRRfX5p$ 

[samcv]

Sent an email to @ingydotnet regarding XXX and the lack of having a license field causing problems due to not being able to be worked on further without them. Will let you know if I hear back.

[jonathanstowe]

• Manifesto I know about - there is one test that is flaky on some platforms which I can't replicate locally so am struggling to fix. May just TODO the test for the time being.
• ~~Audio::Silan - it probably isn't checking the external dependency properly, just making a clean environment to see about that ~~ FIXED
• Audio::PortMIDI - I think this is probably a permissions problem I have seen on Linux machines which have never run any audio software before, it may be just check upfront and skip all the tests with a loud warning in this case. FIXED
• Net::AMQP - could be one of a number things as I haven't been testing it very frequently, but will look into it. It appears that it was just skipping fewer tests than planned when there was no AMQP broker. FIXED

BTW I am smoke testing all of my modules for every rakudo change locally and Manifesto is the only one that I ever see any fails.

[titsuki]

I have already fixed these:

• Algorithm::Kruskal
• Algorithm::LibSVM

[JJ]

Could we use this for the bitrot squashathon?

[zoffixznet]

We have the recent data on failing modules in toaster; I see no reason to use statics that are more than a year out of date. This issue should probably be closed.

[cosimo]

Please ignore Facter, it was just an experiment, not worth including it.
String::CRC32 seems fine to me, plus the "log" entry points to an invalid link?

[zoffixznet]

String::CRC32 seems fine to me, plus the "log" entry points to an invalid link

Right. This Issue is way too out of date. Closing.

[AlexDaniel]

Right. This Issue is way too out of date. Closing.

FWIW See https://github.com/perl6/ecosystem-unbitrot for a slightly newer attempt to fix the ecosystem.