BBC: Blead Breaks IO::All #18019

cjg-cguevara · 2020-08-02T22:30:33Z

This is a bug report for perl from "Carlos Guevara" [email protected],
generated with the help of perlbug 1.42 running under perl 5.33.1.

[Please describe your issue here]

BBC: Blead Breaks IO::All

Please see http://fast-matrix.cpantesters.org/?dist=IO-All

[Please do not change anything below this line]

Flags:
category=core
severity=low

Site configuration information for perl 5.33.1:

Configured by cpan at Fri Jul 31 09:52:26 EDT 2020.

Summary of my perl5 (revision 5 version 33 subversion 1) configuration:
Commit id: b22e6ad
Platform:
osname=openbsd
osvers=6.7
archname=OpenBSD.amd64-openbsd-thread-multi
uname='openbsd cjg-openbsd6 6.7 generic#5 amd64 '
config_args='-des -Dprefix=~~/bin/perl-blead -Dscriptdir=~~/bin/perl-blead/bin -Dusedevel -Duse64bitall -Duseithreads'
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='cc'
ccflags ='-pthread -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include -D_FORTIFY_SOURCE=2'
optimize='-O2'
cppflags='-pthread -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include'
ccversion=''
gccversion='4.2.1 Compatible OpenBSD Clang 8.0.1 (tags/RELEASE_801/final)'
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='cc'
ldflags ='-pthread -Wl,-E -fstack-protector-strong -L/usr/local/lib'
libpth=/usr/lib /usr/local/lib
libs=-lpthread -lm -lutil -lc
perllibs=-lpthread -lm -lutil -lc
libc=/usr/lib/libc.so.96.0
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 -fPIC -L/usr/local/lib -fstack-protector-strong'

@inc for perl 5.33.1:
/home/cpan/bin/perl-blead/lib/site_perl/5.33.1/OpenBSD.amd64-openbsd-thread-multi
/home/cpan/bin/perl-blead/lib/site_perl/5.33.1
/home/cpan/bin/perl-blead/lib/5.33.1/OpenBSD.amd64-openbsd-thread-multi
/home/cpan/bin/perl-blead/lib/5.33.1

Environment for perl 5.33.1:
HOME=/home/cpan
LANG (unset)
LANGUAGE (unset)
LC_ALL=C
LD_LIBRARY_PATH (unset)
LOGDIR (unset)
PATH=/home/cpan/bin/perl-blead/bin:/home/cpan/bin:/home/cpan/bin:/bin:/sbin:/usr/bin:/usr/sbin:/usr/X11R6/bin:/usr/local/bin:/usr/local/sbin:/usr/games
PERL_BADLANG (unset)
SHELL=/usr/local/bin/bash

The text was updated successfully, but these errors were encountered:

karenetheridge · 2020-08-02T22:45:29Z

I see both passing and failing reports from your systems under 5.33.1. What was the difference between those perl builds? For example, knowing what commits you built each of those from would be helpful as a starting point for bisection.

xenu · 2020-08-02T23:04:17Z

perl -V output in the cpantesters reports contains the commit id. It seems that both failing linux reports are using 4a1bbd3 while the latest successful report was using f4b02d7.

jkeenan · 2020-08-02T23:58:34Z

This is a bug report for perl from "Carlos Guevara" [email protected],
generated with the help of perlbug 1.42 running under perl 5.33.1.

[Please describe your issue here]

BBC: Blead Breaks IO::All

Bisection points to this commit:

$ gitshowf 894d8b10212a906402f4db9f9aac9efe9fa084fd
commit 894d8b10212a906402f4db9f9aac9efe9fa084fd (HEAD, refs/bisect/bad)
Author:     Ryan Voots <[email protected]>
AuthorDate: Tue Nov 12 14:57:57 2019 -0500
Commit:     Karl Williamson <[email protected]>
CommitDate: Thu Jul 30 15:25:17 2020 -0600

    Small changes to test on bugzilla and improve compatibility

diff --git a/dist/Safe/Safe.pm b/dist/Safe/Safe.pm
index 1ed1ee9b19..44b0786c5c 100644
--- a/dist/Safe/Safe.pm
+++ b/dist/Safe/Safe.pm
@@ -67,7 +67,7 @@ require utf8;
 # particular code points don't cause it to load.
 # (Swashes are cached internally by perl in PL_utf8_* variables
 # independent of being inside/outside of Safe. So once loaded they can be)
-do { my $a = pack('U',0x100); $a =~ m/\x{1234}/; $a =~ tr/\x{1234}//; };
+do { my $a = pack('U',0x100); utf8->import; require 'utf8_heavy.pl' };
 # now we can safely include utf8::SWASHNEW in $default_share defined below.
 
 my $default_root  = 0;
diff --git a/dist/Safe/t/safeutf8.t b/dist/Safe/t/safeutf8.t
index c58f7e5ad1..cda98cba2c 100644

Here is the tail of the bisection output:

...
 /tmp/tHQHrLMkwM/lib/perl5/5.33.1/pod/perltooc.pod
  /tmp/tHQHrLMkwM/lib/perl5/5.33.1/pod/perltoot.pod
Can't locate utf8_heavy.pl in @INC (@INC contains: /home/jkeenan/.cpan /tmp/tHQHrLMkwM/lib/perl5/site_perl/5.33.1/x86_64-linux /tmp/tHQHrLMkwM/lib/perl5/site_perl/5.33.1 /tmp/tHQHrLMkwM/lib/perl5/5.33.1/x86_64-linux /tmp/tHQHrLMkwM/lib/perl5/5.33.1) at /tmp/tHQHrLMkwM/lib/perl5/5.33.1/Safe.pm line 70.
Compilation failed in require at /tmp/tHQHrLMkwM/lib/perl5/5.33.1/CPAN.pm line 57.
BEGIN failed--compilation aborted at /tmp/tHQHrLMkwM/lib/perl5/5.33.1/CPAN.pm line 57.
Compilation failed in require.
BEGIN failed--compilation aborted.
HEAD is now at 894d8b1021 Small changes to test on bugzilla and improve compatibility
bad - non-zero exit from /tmp/tHQHrLMkwM/bin/perl -I /home/jkeenan/.cpan -MCPAN::MyConfig -MCPAN -e $CPAN::Config->{build_dir}=q{/tmp/HPetnZKo4u}; -e install('IO::All'); die unless CPAN::Shell->expand(Module => 'IO::All')->uptodate;
894d8b10212a906402f4db9f9aac9efe9fa084fd is the first bad commit
commit 894d8b10212a906402f4db9f9aac9efe9fa084fd
Author: Ryan Voots <[email protected]>
Date:   Tue Nov 12 14:57:57 2019 -0500

    Small changes to test on bugzilla and improve compatibility

:040000 040000 bbc79ad1911370eccd62c230cafcd4d183ea8ab2 c4f6ec8cb549d7adf06840d5d7ec641cb5cdeadc M	dist
bisect run success
That took 1519 seconds.

@khwilliamson @simcop2387 Can you take a look?

Thank you very much.
Jim Keenan

Leont · 2020-08-04T17:01:10Z

Bisection points to this commit: 894d8b1

That doesn't make sense, IO::All isn't using Safe.

atoomic · 2020-08-04T17:02:23Z

all the most this change was reverted

karenetheridge · 2020-08-04T18:11:50Z

It looks like the build failed at that commit, so the bisection script decided that this was the significant commit. Instead, that commit should be skipped during the bisection so it can resume and continue testing against IO::All.

ppisar · 2020-08-06T06:17:48Z

This is triggered by:

commit 89341f87f9fc65c4d7133e497bb04586e86b8052
Author: Tony Cook <[email protected]>
Date:   Tue May 12 10:29:17 2020 +1000

    make $fh->error report errors from both input and output
    
    For character devices and sockets perl uses separate PerlIO objects
    for input and output so they can be buffered separately.
    
    The IO::Handle::error() method only checked the input stream, so
    if a write error occurs error() would still returned false.
    
    Change this so both the input and output streams are checked.
    
    fixes #6799

ppisar · 2020-08-06T07:30:01Z

A reproducer:

$ perl -e 'open my $f, q{<}, q{/etc/hosts} or die; print qq{error\n} if $f->error'
error

It seems the patch breaks IO::Handle error() method for regular files.

89341f8 fix for GH Perl#6799 introduced a regression when calling error() on an IO::Handle object that was opened for reading a regular file: $ perl -e 'open my $f, q{<}, q{/etc/hosts} or die; print qq{error\n} if $f->error' error In case of a regular file opened for reading, IoOFP() returns NULL and PerlIO_error(NULL) reports -1. Compare to the case of a file opened for writing when both IoIFP() and IoOFP() return non-NULL, equaled pointer. This patch fixes handling the case of the NULL output stream. GH Perl#18019

89341f8 fix for GH #6799 introduced a regression when calling error() on an IO::Handle object that was opened for reading a regular file: $ perl -e 'open my $f, q{<}, q{/etc/hosts} or die; print qq{error\n} if $f->error' error In case of a regular file opened for reading, IoOFP() returns NULL and PerlIO_error(NULL) reports -1. Compare to the case of a file opened for writing when both IoIFP() and IoOFP() return non-NULL, equaled pointer. This patch fixes handling the case of the NULL output stream. GH #18019

jkeenan · 2020-09-01T00:48:35Z

http://fast-matrix.cpantesters.org/?dist=IO-All;perl=5.33.2;reports=1 is showing all GREEN for IO::All. Closing this ticket.

Thank you very much.
Jim Keenan

cjg-cguevara added the Needs Triage label Aug 2, 2020

jkeenan added BBC Blead Breaks CPAN - changes in blead broke a cpan module(s) and removed Needs Triage labels Aug 2, 2020

ppisar mentioned this issue Aug 6, 2020

IO::Handle: Fix a spurious error reported for regular file handles #18035

Merged

khwilliamson added a commit that referenced this issue Aug 19, 2020

perldelta: Note GH #18019 is fixed

ca316df

jkeenan closed this as completed Sep 1, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

BBC: Blead Breaks IO::All #18019

BBC: Blead Breaks IO::All #18019

cjg-cguevara commented Aug 2, 2020

karenetheridge commented Aug 2, 2020

Uh oh!

xenu commented Aug 2, 2020 •

edited

Loading

Uh oh!

jkeenan commented Aug 2, 2020

Uh oh!

Leont commented Aug 4, 2020

Uh oh!

atoomic commented Aug 4, 2020

Uh oh!

karenetheridge commented Aug 4, 2020

Uh oh!

ppisar commented Aug 6, 2020

Uh oh!

ppisar commented Aug 6, 2020

Uh oh!

jkeenan commented Sep 1, 2020

Uh oh!

BBC: Blead Breaks IO::All #18019

BBC: Blead Breaks IO::All #18019

Comments

cjg-cguevara commented Aug 2, 2020

[Please do not change anything below this line]

Flags: category=core severity=low

karenetheridge commented Aug 2, 2020

Uh oh!

xenu commented Aug 2, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jkeenan commented Aug 2, 2020

Uh oh!

Leont commented Aug 4, 2020

Uh oh!

atoomic commented Aug 4, 2020

Uh oh!

karenetheridge commented Aug 4, 2020

Uh oh!

ppisar commented Aug 6, 2020

Uh oh!

ppisar commented Aug 6, 2020

Uh oh!

jkeenan commented Sep 1, 2020

Uh oh!

Flags:
category=core
severity=low

xenu commented Aug 2, 2020 •

edited

Loading