Permalink
Browse files

r2502@k75: k | 2007-06-07 07:40:54 +0200

 distropref for SWF::Builder


git-svn-id: https://pause.perl.org:5460/svn/cpanpm/trunk@1876 81f69c26-07c4-0310-b1c3-f486c8728884
  • Loading branch information...
1 parent 9f267fa commit c154fd552026d66ec0a68068fefe765b3766684d @andk committed Jun 7, 2007
Showing with 315 additions and 4 deletions.
  1. +1 −0 MANIFEST
  2. +148 −4 Todo
  3. +59 −0 bin/add-perl-meta.pl
  4. +101 −0 bin/dir-to-jifty.pl
  5. +6 −0 distroprefs/YSAS.SWF-Builder.yml
View
@@ -178,6 +178,7 @@ distroprefs/VKON.Tcl-Tk.yml
distroprefs/VKON.Tcl.yml
distroprefs/YARON.WWW-Monitor.yml
distroprefs/YOSHIDA.WebService-YouTube.yml
+distroprefs/YSAS.SWF-Builder.yml
distroprefs/YVES.Data-Dump-Streamer.yml
distroprefs/ZEV.Test-Dependencies.yml
inc/Test/Builder.pm
View
152 Todo
@@ -1,3 +1,145 @@
+2007-06-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
+
+ * SWF::Builder flaky test. Seemingly random results. This RT ticket is
+ now 4 weeks old. Need to patch the distro.
+
+ http://rt.cpan.org/Ticket/Display.html?id=26971
+
+ * POE 0.9989 fails a tk test under very rare unknown circumstances.
+
+ Only three time (2007-06-06T19:55, 2007-05-29T05:48, 2007-05-26T14:17)
+ in 26*6+1 times. The failure is not reproducable. The logfiles show all
+ three times exactly the same:
+
+t/30_loops/50_tk/wheel_run...................
+# Failed test 'coderef/shutdown close'
+# at t/30_loops/00_base/wheel_run.pm line 332.
+# got: 'close'
+# expected: 'stdout'
+# Looks like you planned 99 tests but only ran 97.
+# Looks like you failed 1 test of 97 run.
+dubious
+^ITest returned status 1 (wstat 256, 0x100)
+DIED. FAILED tests 53, 98-99
+^IFailed 3/99 tests, 96.97% okay
+
+ * MSISK/HTML-TableExtract
+
+ https://rt.cpan.org/Ticket/Display.html?id=20709
+
+ The output looks very similar to HTML::CalendarMonth. It's about the
+ difference between <foo attr=0> and <foo attr="0">.
+
+ I see very randomly looking test results. There must be a correlation to
+ some already installed module.
+
+$ grep -r 'eval.*use ' lib/
+lib/HTML/TableExtract.pm: eval "use HTML::TreeBuilder";
+lib/HTML/TableExtract.pm: eval "use HTML::ElementTable 1.17";
+
+ The TreeBuilder is it.
+
+ But this does not answer the HTML::CalendarMonth problem.
+
+ * why does an upgrade install Chart 2.4.1?
+
+2007-06-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
+
+ * TANIGUCHI/Unicode-RecursiveDowngrade-0.03.tar.gz is the next candidate.
+ And JUERD/encoding-split-0.01.tar.gz?
+
+ * Todo: Test if we can work with YAML::Tiny.
+ http://svn.ali.as/cpan/concept/cpan-yaml-tiny/
+
+ * WWW::Live365 has very mixed/random test results. Most recent breakage
+ was WWW::Mechanize but this seems to be fixed now, so probably
+ everything OK.
+
+2007-06-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
+
+ * Watch HTML-TreeBuilder-XPath-0.08
+
+ It failed suddenly in current build and has quite a good history of
+ green builds, although not perfect.
+
+ * Binary searches in some regions below 23023 are getting uglier and
+ uglier. I now see test failures in CPAN.pm's 10version.t (Out of
+ memory!) so CPAN.pm does not get installed, we have to use a very old
+ CPAN.pm which cannot install current CPAN::Reporter and cannot run
+ distroprefs. So it asks all the questions that we had so happily
+ eliminated.
+
+ Some time ago I had always used the very current CPAN.pm from the repo
+ for running the binary searches. Maybe I should switch back to that?
+
+ Notest is your friend! First we check if we are below patchlevel 23023.
+ Then we check if we are below 1.8802. If so, we install 1.8802 so that
+ we get at "notest". Then we install current version with the help of
+ notest. Fini. Now it flies again.
+
+ * SADAHIRO/ShiftJIS-X0213-MapUTF-0.30.tar.gz might also be a candidate,
+ or with the same history pattern:
+ SADAHIRO/ShiftJIS-CP932-MapUTF-1.01.tar.gz
+
+ Let's try with
+
+ ShiftJIS::X0213::MapUTF
+
+ This became a BBC article on 2007-06-05
+
+ * Time-HiRes-Value-0.04 has an unreproducable test failure.
+
+make[3]: Entering directory `/home/sand/.cpan/build/Time-HiRes-Value-0.04-GZ0nGl'
+PERL_DL_NONLAZY=1 /home/src/perl/repoperls/installed-perls/perl/p4tyIz4/perl-5.8.0@31330/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
+t/01test......ok
+t/02timing....
+# Failed test at t/02timing.t line 19.
+# got: ''
+# expected: '1'
+# Looks like you failed 1 test of 2.
+dubious
+^ITest returned status 1 (wstat 256, 0x100)
+DIED. FAILED test 2
+^IFailed 1/2 tests, 50.00% okay
+Failed Test Stat Wstat Total Fail List of Failed
+-------------------------------------------------------------------------------
+t/02timing.t 1 256 2 1 2
+Failed 1/2 test scripts. 1/52 subtests failed.
+
+ * LBROCARD/Devel-ebug-0.48.tar.gz has intermittent failures, should be
+ reported.
+
+ * SAVA/Encode-IMAPUTF7-1.00.tar.gz candidate for binary search? 23023
+ fails also. 20002 doesn't fail. After many hours we reach the interval
+ between 22774 and 22872 and we need to build new perls again.
+
+ Before we reach the endpoint I already see that the test output has
+ already changed somewhere, so maybe we need more binary searches. The
+ failure @22872 looks so simple:
+
+t/0-test....Modification of a read-only value attempted at /home/sand/.cpan/build/Encode-IMAPUTF7-1.00/blib/lib/Encode/IMAPUTF7.pm line 39.
+# Looks like your test died before it could output anything.
+t/0-test....dubious
+ Test returned status 255 (wstat 65280, 0xff00)
+
+----Program----
+eval q{use Encode::IMAPUTF7 1.00};
+print $@ ? "N/A" : "OK";
+print "\n";
+
+----Output of .../p1d1ohA/perl-5.8.0@22841/bin/perl----
+OK
+
+----EOF ($?='0')----
+----Output of .../p4U7uLf/perl-5.8.0@22842/bin/perl----
+N/A
+
+----EOF ($?='0')----
+
+
+ * Net-Write has a negative dependency on Pod::Coverage on bleadperl. The
+ type we saw with File::Slurp and Graph::Easy.
+
2007-06-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
* New idea how to quickly locate more candidates for binary searches:
@@ -770,6 +912,7 @@ No patch available between 22739 and 22741
Include variable names in "Use of uninitialized value" warnings
+ http://groups.google.com/group/perl.perl5.porters/browse_thread/thread/4b3dafe41b70c5a6/b056a4d7ae5a673a?lnk=gst&q=22741+bbc&rnum=1#b056a4d7ae5a673a
* MCMAHON!Devel-Command-0.08.xml consistently succeeds with maint and
fails with blead.
@@ -1479,8 +1622,9 @@ Module id = Coro::Event
Implement configure_requires? (testcase RPM::XML)
Implement a depends/requires keyword for distroprefs? (Testcase
- Spreadsheet ParseExcel or XML-Filter-XInclude which does not declare
- XML::SAX::Writer as dependency)
+ Spreadsheet-ParseExcel (http://rt.cpan.org/Ticket/Display.html?id=24693)
+ or XML-Filter-XInclude
+ (http://rt.cpan.org/Ticket/Display.html?id=27401))
Speedup the cleanup? DONE [...time passes...] But I do not like the new
version. I want to see how long it takes/will take. svk rev. 2216 seems
@@ -2077,8 +2221,8 @@ in my Todo file now and will be dealt with after 1.90.
have to pass on this one. Pity.
And we have a similar issue with Flickr::API which also depends on
- XML::Parser::Lite (but which still fails tests, even if we resolve the
- X:P:L dependency).
+ XML::Parser::Lite::Tree (but which still fails tests, even if we resolve
+ the X:P:L dependency).
I think it's best described as a pseudo dependency. It depends on a
thing that is contained in another but this other thing hides this (with
View
@@ -0,0 +1,59 @@
+#!/usr/local/bin/perl
+
+=pod
+
+ $0 --logs=megainstalldir
+
+looks at one dirent that tells us which perl we are talking about
+
+Ask that perl for its config.
+
+
+=cut
+
+
+use strict;
+use warnings;
+
+use FindBin ();
+use Getopt::Long;
+use XML::LibXML;
+our %Opt;
+
+sub Usage (){
+ "Usage: $0 [--logs=logsdirector]";
+}
+
+GetOptions(\%Opt, "logs=s"
+ ) or die Usage;
+
+my $logdir = $Opt{logs} || "$FindBin::Bin/../logs";
+opendir my $dh, $logdir or die "cannot opendir '$logdir': $!";
+
+my $p = XML::LibXML->new;
+for my $dirent (sort readdir $dh) {
+ next if $dirent =~ /^\./;
+ my $abs = "$logdir/$dirent";
+ next unless $abs =~ /(?:^|\/)megainstall\.(\d+T\d+)\.d(?:\/|$)/;
+ my $time = $1;
+ my $xfile = sprintf "%s/%s", $logdir, $dirent;
+ next unless $xfile =~ /\.xml$/;
+ next unless -e $xfile;
+ my $xml = $p->parse_file($xfile);
+ my($ok,$seq,$perl,$patchlevel,$branch);
+ $ok = $xml->findvalue("/distro/\@ok");
+ $seq = $xml->findvalue("/distro/\@seq") || 0;
+ $perl = $xml->findvalue("/distro/\@perl");
+ $perl .= "/bin/perl";
+ unless (-e $perl) {
+ die "perl[$perl] n'exists";
+ }
+ open my $fh, ">", "$logdir/perl-V.txt" or die "Could not open >$logdir/perl-V.txt: $!";
+ open my $pfh, "-|", $perl, "-V" or die "cannot fork: $!";
+ while (<$pfh>) {
+ print $fh $_;
+ }
+ close $pfh or die "perl died during -V";
+ close $fh or die "could not write: $!";
+ last;
+}
View
@@ -0,0 +1,101 @@
+#!/usr/local/perl-m-5.8.0@31162/bin/perl
+
+use strict;
+use warnings;
+use UNIVERSAL::require;
+use Jifty::ClassLoader ();
+BEGIN {
+ Jifty::Util->require or die $UNIVERSAL::require::ERROR;
+ my $root = Jifty::Util->app_root;
+ unshift @INC, "$root/lib" if ($root);
+}
+use Jifty;
+my $cl = Jifty::ClassLoader->new (base => "Pocpoc");
+$cl->require;
+my $j = new Jifty; # sets up database connection and other stuff
+
+use FindBin ();
+use XML::LibXML;
+
+my $logdir = shift || "$FindBin::Bin/../logs";
+opendir my $dh, $logdir or die "cannot opendir '$logdir': $!";
+
+my $p = XML::LibXML->new;
+my $i = 0;
+$|=1;
+SESSION: for my $dirent (sort { $b cmp $a } readdir $dh) {
+ next if $dirent =~ /^\./;
+ my $abs = "$logdir/$dirent";
+ next unless $abs =~ /(?:^|\/)megainstall\.(\d+T\d+)\.d(?:\/|$)/;
+ my $starttime = $1;
+ opendir my $dh2, $abs or die "cannot opendir: '$abs': $!";
+ my($total,$failed);
+ $total = 0;
+ $failed = 0;
+ my @readdir2 = sort grep { /\.xml$/ } readdir $dh2;
+ print "\@";
+ my $s = Pocpoc::Model::Session->new(handle => Jifty->handle);
+ TESTRUN: for my $i (0..$#readdir2) {
+ my $dirent2 = $readdir2[$i];
+ my $xfile = "$abs/$dirent2";
+ my $xml = $p->parse_file($xfile);
+ my($ok,$seq,$perl,$distro,$branch,$patchlevel);
+ $ok = $xml->findvalue("/distro/\@ok");
+ $seq = $xml->findvalue("/distro/\@seq") || 0;
+ $perl = $xml->findvalue("/distro/\@perl");
+ $distro = $xml->findvalue("/distro/\@distro");
+ ($branch,$patchlevel) = $perl =~ m|/installed-perls/(.*?)/p.*?/perl-5.*?@(\d+)|;
+ $total++;
+ $failed++ unless $ok eq "OK";
+ $s->load_or_create(
+ starttime => $starttime,
+ perl => $perl,
+ branch => $branch,
+ patchlevel => $patchlevel,
+ );
+ if ($i == 0) {
+ my $total = $s->total || 0;
+ if ( $total == @readdir2 ) {
+ print "_";
+ next SESSION;
+ } else {
+ print "($dirent)";
+ }
+ }
+ print "+"; # unless $i % 64;
+ my $t = Pocpoc::Model::Testrun->new(handle => Jifty->handle);
+ $t->load_or_create(
+ distro => $distro,
+ testsession => $s,
+ );
+ $t->set_seq($seq);
+ $t->set_testresult($ok);
+ }
+ $s->set_failed($failed);
+ $s->set_total($total);
+ my $vtotal = $s->total;
+ unless ($vtotal == $total) {
+ die "Sanity check broke: total[$total]vtotal[$vtotal]";
+ }
+ my $vfailed = $s->failed;
+ unless ($vfailed == $failed) {
+ die "Sanity check broke: failed[$failed]vfailed[$vfailed]";
+ }
+}
+
+
+=pod
+
+/usr/bin/perl -le 'use DBI; my $db = shift or die;my $dbh = DBI->connect("dbi:SQLite:dbname=$db","","") or die;my $sth = $dbh->prepare("select * from testruns");$sth->execute;while (my @row = $sth->fetchrow){ print "row[@row]";}' poce | head
+row[1 /home/src/perl/cpan-sql-stuff/CPAN-SQLite/. 1 0 OK]
+row[2 /home/src/perl/tk/SVN/. 1 0 OK]
+row[3 ABH/XML-RSS-1.22.tar.gz 1 0 OK]
+row[4 ABIGAIL/Regexp-Common-2.120.tar.gz 1 0 OK]
+row[5 ABW/Class-Singleton-1.03.tar.gz 1 0 OK]
+row[6 ACALPINI/Lingua-Stem-It-0.01.tar.gz 1 0 OK]
+row[7 ADAMK/Algorithm-Dependency-1.102.tar.gz 1 0 OK]
+row[8 ADAMK/AppConfig-1.64.tar.gz 1 0 OK]
+row[9 ADAMK/Archive-Zip-1.18.tar.gz 1 0 OK]
+row[10 ADAMK/CPAN-Inject-0.05.tar.gz 1 0 NOT OK]
+
+==cut
@@ -0,0 +1,6 @@
+---
+comment: "http://rt.cpan.org/Ticket/Display.html?id=26971"
+match:
+ distribution: "^YSAS/SWF-Builder-0.16.tar.gz"
+patches:
+ - "ANDK/patches/SWF-Builder-0.16-ANDK-01.patch.gz"

0 comments on commit c154fd5

Please sign in to comment.