Smoking 5.12.0-RC0 against 5.10.1 -*- mode:org; coding:utf-8; -*-
https://rt.cpan.org/Ticket/Display.html?id=33929
https://rt.cpan.org/Ticket/Display.html?id=33844
https://rt.cpan.org/Ticket/Display.html?id=56060
engine.c:15:2: error: #error Unsupported PERL_API_VERSION engine.c: In function ‘get_regclass_desc’: engine.c:182: error: ‘regexp’ has no member named ‘data’ engine.c: In function ‘find_internal’: engine.c:428: error: ‘regexp’ has no member named ‘program’ engine.c: In function ‘rc_regcomp’: engine.c:599: warning: passing argument 1 of ‘Perl_pregcomp’ from incompatible pointer type engine.c:599: warning: passing argument 2 of ‘Perl_pregcomp’ makes integer from pointer without a cast engine.c:599: error: too many arguments to function ‘Perl_pregcomp’ engine.c:599: warning: assignment from incompatible pointer type
-> side-effect of changing regexp interna
http://rt.cpan.org/Public/Bug/Display.html?id=33531
https://rt.cpan.org/Ticket/Display.html?id=56129 check not implemented for reftype==REGEXP
http://rt.cpan.org/Public/Bug/Display.html?id=43115
https://rt.cpan.org/Ticket/Display.html?id=56061 does not compile anymore, changed regexp structure?
https://rt.cpan.org/Ticket/Display.html?id=56074 does not compile anymore, changed regexp structure?
-> reported by mail to p5p -> was deprecated in 5.10.0
http://rt.cpan.org/Public/Bug/Display.html?id=43686
http://rt.cpan.org/Public/Bug/Display.html?id=43248
Unsupported private C API functions are now declared “static” to prevent leakage to Perl’s public API.
-> Undefined symbol “Perl_fold_constants” -> made static in b7783a124ffeaab87679eba041dd9997f4d5372a
Author: Nicholas Clark <nick@ccl4.org> Date: Wed Nov 26 19:36:06 2008 +0000
force_list(), fold_constants() and gen_constant_list() can be static.
-> https://rt.cpan.org/Ticket/Display.html?id=56031
See also: Message-ID: <20090118154803.GY2919@plum.flirble.org>
Ah, a rt ticket for perl exists already: http://rt.perl.org/rt3/Public/Bug/Display.html?id=56536
-> Undefined symbol “Perl_fold_constants” -> https://rt.cpan.org/Ticket/Display.html?id=56032
https://rt.cpan.org/Ticket/Display.html?id=56052 -> fold_constants problem, see elsewhere
https://rt.cpan.org/Ticket/Display.html?id=55384 (linklist was made static)
https://rt.cpan.org/Ticket/Display.html?id=36165
https://rt.cpan.org/Ticket/Display.html?id=56123 Invalid version format (version required) at /usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm line 467.
Invalid version format (dotted-decimal versions require at least three parts) at /usr/perl-5.12.0-RC0/lib/5.12.0/version.pm line 150. -> https://rt.cpan.org/Ticket/Display.html?id=56033
(without RT ticket, because it’s not really the author’s fault!)
The fatal error was: Invalid version format (non-numeric data) at /usr/perl-5.12.0-RC0/lib/5.12.0/Module/Build/ModuleInfo.pm line 348, <GEN0> line 6. (again HTML::Table 2.08a)
https://rt.cpan.org/Ticket/Display.html?id=56221 Prototype after ‘@’ for File::Attributes::_foreach_plugin : @& at /var/tmp/CPAN-build/File-Attributes-0.04-WJHVKr/blib/lib/File/Attributes.pm line 27. Invalid version format (non-numeric data) at t/override-notapplicable.t line 19.
t/override-notapplicable.t ..
In 5.10.1 the version error/warning looked different:
Version string ‘foo-o-matic’ contains invalid data; ignoring: ‘foo-o-matic’ at t/override-notapplicable.t line 19.
https://rt.cpan.org/Ticket/Display.html?id=56073 XML::Tidy has a non-strict and non-lax version number -> maybe version.pm should be more lax in lax mode?
At least the poor error message of Module::Install needs to be fixed:
*** Installing dependencies… Invalid version format (non-numeric data) at /usr/perl-5.12.0-RC0/lib/5.12.0/Module/Load/Conditional.pm line 301. Compilation failed in require at inc/Module/Install/AutoInstall.pm line 37.
https://rt.cpan.org/Ticket/Display.html?id=56081 -> ??? Invalid version format (non-numeric data)
http://rt.cpan.org/Public/Bug/Display.html?id=54132 but there’s already a (different) patch in the perl core version of XSLoader
http://rt.cpan.org/Public/Bug/Display.html?id=55723 test was actually checking that dates beyond 2038 did not work on 32 bit systems
-> reported by mail to p5p -> “Deprecations warn by default”
https://rt.cpan.org/Ticket/Display.html?id=56267 defined %… warning is fatal in test
https://rt.cpan.org/Ticket/Display.html?id=56268 defined %… warning is fatal in test
http://rt.cpan.org/Public/Bug/Display.html?id=54062 really fails because Pod::POM::Node is using defined(%hash) syntax
https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6392 defined(%hash) is deprecated at ../blib/lib/Mail/SpamAssassin/Dns.pm line 757. (Maybe you should just omit the defined()?) Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409. Use of “goto” to jump into a construct is deprecated at ../blib/lib/Mail/SpamAssassin/Plugin/Check.pm line 409.
t/recreate.t …………………. Failed 1/9 subtests
line 86 is: ok($warning == 0);
https://rt.cpan.org/Ticket/Display.html?id=56122
https://rt.cpan.org/Ticket/Display.html?id=56126 fatal error, because author specified “use warnings FATAL”
https://rt.cpan.org/Ticket/Display.html?id=56079 -> does not compile anymore, ‘OP_SETSTATE’ undeclared
https://rt.cpan.org/Ticket/Display.html?id=56078 -> does not compile anymore, ‘OP_SETSTATE’ undeclared
https://rt.cpan.org/Ticket/Display.html?id=56080 stolen_chunk_of_op.c: In function ‘namespace_alias_peep’: stolen_chunk_of_op.c:69: error: ‘OP_SETSTATE’ undeclared (first use in this function) stolen_chunk_of_op.c:69: error: (Each undeclared identifier is reported only once stolen_chunk_of_op.c:69: error: for each function it appears in.)
http://rt.cpan.org/Public/Bug/Display.html?id=34716
-> probably same issue like the similar module
http://rt.cpan.org/Public/Bug/Display.html?id=52643 possibly really a bugfix?
https://rt.cpan.org/Ticket/Display.html?id=56051
Undefined subroutine &MIME::Base64::encode_base64 called at t/03_autodynaload_hook.t line 36.
t/03_autodynaload_hook.t …..
(but why???)
- truss output with 5.10.1:
stat(“/usr/perl-5.10.1-test/lib/5.10.1/i386-freebsd/auto/MIME/Base64/Base64.bs”,{ mode=-r–r–r– ,inode=16322911,size=0,blksize=4096 }) = 0 (0x0) … open(“/usr/perl-5.10.1-test/lib/5.10.1/i386-freebsd/auto/MIME/Base64/Base64.so”,O_RDONLY,027757754150) = 4 (0x4)
- truss output with 5.12.0:
stat(“/usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/i386-freebsd/auto/MIME/Base64/Base64.bs”,0x8158a60) ERR#2 ‘No such file or directory’ … open(“/usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/i386-freebsd/auto/MIME/Base64/Base64.so”,O_RDONLY,027757754270) ERR#2 ‘No such file or directory’
In fact, the .so is located in /usr/perl-5.12.0-RC0/lib/5.12.0/i386-freebsd/auto/MIME/Base64/Base64.so
-> https://rt.cpan.org/Ticket/Display.html?id=56034 (see also Config-Model-0.644)
https://rt.cpan.org/Ticket/Display.html?id=56219
Wide character in print at t/numbers.t line 143.
etc. fails between 5.11.2 and 5.11.4 -> binsect is running -> need to restart bisect
http://rt.cpan.org/Public/Bug/Display.html?id=35246 (probably just a bugfix in perl)
http://rt.cpan.org/Public/Bug/Display.html?id=55902 some Devel::Declare magic not working anymore?
https://rt.cpan.org/Ticket/Display.html?id=56218 -> binsect says: 09330df80caf214f375fcf0c04857347e3b17c69
Fix [perl #66970] Incorrect coderef in MODIFY_CODE_ATTRIBUTES
https://rt.cpan.org/Ticket/Display.html?id=56263 -> ??? hangs, not a single test for newer perls available
https://rt.cpan.org/Ticket/Display.html?id=56259 -> pod-coverage test fails only with 5.12.0
t/02pod-coverage.t ..
https://rt.cpan.org/Ticket/Display.html?id=56053
t/01-ref.t ….
https://rt.cpan.org/Ticket/Display.html?id=56071 probably related to Module::Build and/or version.pm
https://rt.cpan.org/Ticket/Display.html?id=56068 same problem like Params-Profile-0.10
t/List-Filter-Storage.t ………….
https://rt.cpan.org/Ticket/Display.html?id=56042
https://rt.cpan.org/Ticket/Display.html?id=56270
t/pod-coverage.t …
https://rt.cpan.org/Ticket/Display.html?id=56271 Can’t locate object method “bar” via package “Foo” at t/01-basic.t line 12. t/01-basic.t …………..
https://rt.cpan.org/Ticket/Display.html?id=56222 Previous.xs: In function ‘XS_Object__Previous_previous_object_xs’: Previous.xs:61: error: ‘const struct block_sub’ has no member named ‘hasargs’
https://rt.cpan.org/Ticket/Display.html?id=43284
t/02_Params-Profile-Functions.t ..
-> no, it looks like an Test::More::is_deeply problem, not handling equivalent regexps anymore correctly -> actually, Test::More never catched regexp diffs with is_deeply. See https://rt.cpan.org/Ticket/Display.html?id=53469 But now regexps are different, so the bug probably does not appear anymore. And in fact, with Test-Simple-0.95_01 I can reproduce the same problem also with perl 5.10.1
https://rt.cpan.org/Ticket/Display.html?id=56128 probably some internal change: “cx_type is 40 not CXt_SUB”
The path /usr/perl-5.12.0-RC0/bin/perl is not Perl at t/05_perl_discovery.t line 49
t/05_perl_discovery.t ..
http://cpansearch.perl.org/src/ADAMK/PITA-Scheme-0.42/lib/PITA/Scheme/Perl/Discovery.pm
unless ( $output[0] =~ /^This is perl,/ ) {
extra newline? -> https://rt.cpan.org/Ticket/Display.html?id=56007 nope, “This is perl,” -> “This is perl 5,” probably does not need to appear in perldelta
https://rt.cpan.org/Ticket/Display.html?id=56127 maybe some tiearray optimization?
https://rt.cpan.org/Ticket/Display.html?id=56262 Invalid version format (non-numeric data) at /usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm line 467.
-> but the error message is very bad and it is quite fishy that it’s happening at that line which is reading
467: if (!$installed_file) {
and really contains a pathname only.
-> The complete debugging session:
DB<1> s ExtUtils::MakeMaker::WriteMakefile(/usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm:50): 50: Carp::croak “WriteMakefile: Need even number of args” if @_ % 2;
DB<1> b 467
DB<2> c ExtUtils::MakeMaker::new(/usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm:467): 467: if (!$installed_file) {
DB<2> x $installed_file 0 ‘/usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/Class/Std/Utils.pm’
DB<3> s Invalid version format (non-numeric data) at /usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm line 467. at /usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm line 467 ExtUtils::MakeMaker::new(‘MM’, ‘MM=HASH(0x286703b0)’) called at /usr/perl-5.12.0-RC0/lib/5.12.0/ExtUtils/MakeMaker.pm line 57 ExtUtils::MakeMaker::WriteMakefile(‘NAME’, ‘RDF::Converter::CSV’, ‘PREREQ_PM’, ‘HASH(0x2834d110)’) called at Makefile.PL line 2 Debugged program terminated. Use q to quit or R to restart, use o inhibit_exit to avoid stopping after program termination, h q, h R or h o to get additional info.
https://rt.cpan.org/Ticket/Display.html?id=56261 http://www.cpantesters.org/cpan/report/2400368 t/06perlobject….undefined method ‘defined?’ for nil:NilClass (NoMethodError) at t/06perlobject.t line 57, <GEN0> line 1. breakage happened between 5.10.0 and patch 34383
https://rt.cpan.org/Ticket/Display.html?id=56266
https://rt.cpan.org/Ticket/Display.html?id=56269 the test t/02.mark.t ……….. ok takes forever under 5.12.0
http://rt.cpan.org/Public/Bug/Display.html?id=55894
https://rt.cpan.org/Ticket/Display.html?id=56265
not ok 430 # (chr(0x00ad) ~ /[[:cntrl:]]/) = 1
not ok 902 # (chr(0x0085) ~ /[[:print:]]/) = 1
not ok 1459 # (chr(0x00b2) ~ /[[:word:]]/) = 1
not ok 1460 # (chr(0x00b3) ~ /[[:word:]]/) = 1
not ok 1466 # (chr(0x00b9) ~ /[[:word:]]/) = 1
not ok 1469 # (chr(0x00bc) ~ /[[:word:]]/) = 1
not ok 1470 # (chr(0x00bd) ~ /[[:word:]]/) = 1
not ok 1471 # (chr(0x00be) ~ /[[:word:]]/) = 1
not ok 1971 # (chr(0x00b2) ~ /\w/) = 1
not ok 1972 # (chr(0x00b3) ~ /\w/) = 1
not ok 1978 # (chr(0x00b9) ~ /\w/) = 1
not ok 1981 # (chr(0x00bc) ~ /\w/) = 1
not ok 1982 # (chr(0x00bd) ~ /\w/) = 1
not ok 1983 # (chr(0x00be) ~ /\w/) = 1
https://rt.cpan.org/Ticket/Display.html?id=53909 https://rt.cpan.org/Ticket/Display.html?id=54739
Too much skepticism (L3J:3<5:9 L3L:3<5:9). Compilation failed in require at t/descartes.t line 41.
t/descartes.t ..
see also PerlMol below
-> does not compile anymore, ‘DEBUG_S_FLAG’ undeclared
t/stack.t …….. Dubious, test returned 1 (wstat 256, 0x100) Failed 1/29 subtests
-> ??? maybe new UCD? (maybe, UCD 5.2 was part of 5.11.3, for which a pass exists, hmmm)
Dubious, test returned 1 (wstat 256, 0x100)
-> bisect is running … failing :-(
defined %… warning is fatal in test
failure looks similar to Chemistry-File-SMILES-0.46
-> ???
Also:
The ‘getopts.pl’ legacy library is deprecated and will be removed in the next major release of perl. Please use the Getopt::Long or Getopt::Std module instead. at /usr/perl-5.12.0-RC0/lib/5.12.0/getopts.pl line 11.
Number found where operator expected at t/10-base.t line 78, near “onlycalledonce 1” (Do you need to predeclare onlycalledonce?) Number found where operator expected at t/10-base.t line 107, near “foo 1” (Do you need to predeclare foo?) Number found where operator expected at t/10-base.t line 125, near “wut 13” (Do you need to predeclare wut?) syntax error at t/10-base.t line 78, near “onlycalledonce 1” syntax error at t/10-base.t line 107, near “foo 1” syntax error at t/10-base.t line 125, near “wut 13” BEGIN not safe after errors–compilation aborted at t/10-base.t line 158.
t/10-base.t ……………
syntax error at t/array-object-is-enumerable.t line 13, near “Array[”
etc.
t/t1.t …..
already resolved in 0.04_01 (compile problem: Closure.xs:585: error: duplicate case value)
-> fixed in Config-Model-1.001
http://rt.perl.org/rt3/Ticket/Display.html?id=73926 segfaults: @bla = reverse(@bla) -> resolved for RC2
-> issue (5.10.0!) seems to be resolved in 0.20_01
http://rt.cpan.org/Public/Bug/Display.html?id=53574 -> resolved in ParseLex-2.19
https://rt.cpan.org/Ticket/Display.html?id=56059
-> fixed in 0.16
-> fixed in 0.03
-> fixed in 0.1005 (or was it just a port/tcp/whatever issue?)
-> depends on Attribute-Util
-> very probably depends on forks
-> depends on Method-Signatures
-> depends on Data-Dump-Streamer
-> depends on Data-Dump-Streamer
-> depends on Attribute-Util
-> very probably depends on Devel-Cover
-> depends on Encode-JP-Mobile
-> depends on Regexp-Copy
-> most probably depends also on Encode-JP stuff
-> depends on HTML-Template-Compiled
-> depends on FreezeThaw
-> very probably depends on HTTP-Proxy
-> depends on HTML-Template-Compiled
-> depends on Data-Alias
-> depends on Test-YAML-Meta
-> probably same as …-Analyse
-> very probably depends on WWW-CPAN
-> very probably depends on WWW-CPAN
probably because of Pod-PseudoPod
-> most probably depends also on Encode-JP stuff
-> depends on YAML-LibYAML (YAML::XS)
-> depends on rpm-build-perl
-> depends on Module-CPANTS-Analyse
-> very probably depends on Perl-MinimumVersion
-> depends on Devel-REPL
-> depends on Devel-Cover
-> depends on YAML-LibYAML
-> depends on Data-Dump-Streamer
very probably because of forks
-> depends on FreezeThaw
-> very probably depends on iCal-Parser
-> very probably depends on iCal-Parser
-> depends on Pugs-Compiler-Rule
fails occasionally, maybe under high load
probably a race condition, difficult to reproduce
bind: Address already in use at /var/tmp/CPAN-build/AnyEvent-MPRPC-0.06-j2I6Uh/blib/lib/AnyEvent/MPRPC/Server.pm line 97 EV: error in callback (ignoring): Client got error: Connection reset by peer at /usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/AnyEvent/Handle.pm line 570 EV: error in callback (ignoring): Client got error: Unexpected end-of-file at /usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/AnyEvent/Handle.pm line 570 t/02_connect.t ..
see HTTP-Server-Simple-Recorder-0.03 http://www.cpantesters.org/cpan/report/5530275
does not work well if test suites are concurrent, see also see http://www.cpantesters.org/cpan/report/4419029
random results, probably tight timeout?
sent a kill because the test suite took too long … almost two hours! https://rt.cpan.org/Ticket/Display.html?id=56064
fails under 5.10.x, but not 5.12.x
should not pack .git files into repo (and fails tests also)
needs (but does not declare!) FreezeThaw
random results, also with older perls
Undefined subroutine &Test::Dependencies::get_modules_used_in_file called at /usr/perl-5.12.0-RC0/lib/site_perl/5.12.0/Test/Dependencies.pm line 155.
t/00-dependencies.t ..
random results (under load?)
fails occasionally, maybe under high load
random results, it seems
https://rt.cpan.org/Ticket/Display.html?id=56077 Undefined subroutine &main::plan called at t/kwalitee.t line 8. t/kwalitee.t …… Dubious, test returned 2 (wstat 512, 0x200) No subtests run
(fails because Test-Kwalitee does not build on 5.12.0)
https://rt.cpan.org/Ticket/Display.html?id=56260 fails since commit 17cc9359ea8ee1b546aa067b91362160e3c1e1ee
not ok 5 - Line numbers incremented correctly
(But actually I think the module’s philosophy — incrementing the line number counter by the included content — is actually wrong)
http://www.cpantesters.org/cpan/report/7041981
https://rt.cpan.org/Ticket/Display.html?id=56120
-> Storable+Regexp issue
Can’t store REGEXP items at ../../lib/Storable.pm (autosplit into ../../lib/auto/Storable/_store.al) line 264, at blib/lib/HTML/Template/Compiled.pm line 572
t/19_query.t ………………. Dubious, test returned 255 (wstat 65280, 0xff00) Failed 4/4 subtests
https://rt.cpan.org/Ticket/Display.html?id=33869
Storable was never able to serialize REGEXP items, but now it loudly complains.
https://rt.cpan.org/Ticket/Display.html?id=56125 https://rt.cpan.org/Ticket/Display.html?id=56124
https://rt.cpan.org/Ticket/Display.html?id=56069 probably unrelated to 5.12.x, other perls also show failures
https://rt.cpan.org/Ticket/Display.html?id=55724 -> probably fixed in perl’s core, just needs a CPAN release
t/deprecated.t ….
-> fixed in perl’s core, just needs a CPAN release