pugs: aborting smoke tests
Moritz Lenz
for a few weeks now I haven't been able to complete a single "make
smoke" in the pugs repository.
The last test producing an output is:
t/builtins/strings/uc.t
# Actual: ß
at that point there are no more messages for ~ 10 minutes, and there are
instances of pugs running, one after the next with different PIDs. An
"strace -p `pidof pugs`" results in the following output:
write(6, "a", 1) = 1
...
write(6, "a", 1) = 1
write(6, "a", 1) = 1
futex(0x8999710, FUTEX_WAKE, 1) = 1
futex(0x899970c, FUTEX_WAKE, 1) = 1
futex(0x899973c, FUTEX_WAKE, 1) = 1
futex(0x8990f44, FUTEX_WAKE, 1) = 1
write(6, "a", 1) = 1
...
write(6, "a", 1) = 1
repeated many times. After a while no instance of pugs is running
anymore, and after another minute the "make smoke" exits silently.
I don't think that t/builtins/strings/uc.t itself is the problem,
because it runs in a few seconds if executed manually.
Can anybody reproduce the problem or suggest anything that I could do
about it?
Cheers,
Moritz
--
Moritz Lenz
http://moritz.faui2k3.org/ - http://perl-6.de/
Larry Wall
the embedded/tunneled Perl 5, but it's hard to guess without more
information like OS, locale, etc. most of which would be supplied by:
pugs -V
perl -V
env
If you provide that sort of info, we can compare with our own setups
and think about the differences.
Larry
Moritz Lenz
you're right, I forgot my good manners while writing "bug reports" ;-)
I'm using Debian 4.0 "etch", with the "offical" perl debian packages,
perl -v says:
This is perl, v5.8.8 built for i486-linux-gnu-thread-multi
env is:
TERM=xterm
SHELL=/bin/bash
HOSTDISPLAY=rena:0.0
WINDOWID=52428815
LC_ALL=C
XTERM_SHELL=/bin/bash
USER=moritz
PATH=/bin:/sbin:/usr/bin:/usr/sbin:/home/moritz/bin:/usr/games:/usr/local/Eiffel54/studio/spec/linux-glibc2.1/bin:/usr/bin/X11:/usr/local/bin:/usr/local/Wolfram/Mathematica/5.0/Executables/:/mnt/ex/moritz/matlab/bin
PWD=/home/moritz/pugs
LANG=EN_US.UTF-8
PS1=\u@\[\033\133\061\155\033\133\063\061\155\]\h\[\033\133\060\155\033\133\060\155\]:\w>
HISTCONTROL=ignoredups
XTERM_VERSION=XTerm(222)
FIGNORE=~:.class:.o
SHLVL=2
(a few irrelavant things stripped, like SSH stuff)
output of pugs -V:
This is Perl6 User's Golfing System, version 6.2.13, October 17, 2006
(r15872) built for i486-linux-gnu-thread-multi
Summary of pugs configuration:
archlib: /usr/lib/perl/5.8
archname: i486-linux-gnu-thread-multi
bin: /usr/bin
cc: gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)
embedded: /usr/bin/perl noparrot nohaskell
exe_ext:
file_sep: /
ghc: /usr/bin/ghc 6.6 readline nohsplugins threads
installarchlib: /usr/lib/perl/5.8
installbin: /usr/bin
installman1dir: /usr/share/man/man1
installman3dir: /usr/share/man/man3
installprivlib: /usr/share/perl/5.8
installscript: /usr/bin
installsitearch: /usr/local/lib/perl/5.8.8
installsitebin: /usr/local/bin
installsitelib: /usr/local/share/perl/5.8.8
installsiteman1dir: /usr/local/man/man1
installsiteman3dir: /usr/local/man/man3
osname: linux
pager: /usr/bin/sensible-pager
path_sep: :
perl5_path: /usr/bin/perl
perl_compiler: pugs
perl_revision: 6
perl_subversion: 0
perl_version: 0
prefix: /usr
privlib: /usr/share/perl/5.8
pugs_revision: 15872
pugs_version: Perl6 User's Golfing System, version 6.2.13, October 17,
2006 (r15872)
pugs_versnum: 6.2.13
pugspath: /usr/bin/pugs
regex_engine: default
scriptdir: /usr/bin
sitearch: /usr/local/lib/perl/5.8.8
sitebin: /usr/local/bin
sitelib: /usr/local/share/perl/5.8.8
siteprefix: /usr/local
sitescript: /usr/local/bin
sourcedir: /mnt/ex/moritz/pugs
uname: Linux 2.6.20.6 i686
@*INC:
/usr/lib/perl/5.8
/usr/share/perl/5.8
/usr/local/lib/perl/5.8.8
/usr/local/share/perl/5.8.8
/usr/share/perl/5.8/auto/pugs/perl6/lib
/usr/local/share/perl/5.8.8/auto/pugs/perl6/lib
.
in case it matters, the "phiscal" directory lies at /mnt/ex/moritz/pugs,
usually I work with a symlink in my home dir (that is /home/moritz/pugs
-> /mnt/ex/moritz/pugs)
Cheers,
Moritz
--
Moritz Lenz
http://moritz.faui2k3.org/ - http://sudokugarden.de/ - http://perl-6.de/
Moritz Lenz
Moritz Lenz wrote:
> for a few weeks now I haven't been able to complete a single "make
> smoke" in the pugs repository.
>
> The last test producing an output is:
> t/builtins/strings/uc.t
> # Actual: ß
>
> at that point there are no more messages for ~ 10 minutes, and there are
> instances of pugs running, one after the next with different PIDs.
I just found out that after the smoke test a new smoke.yml is in the
repository's root, but no smoke.html.
The last few lines of smoke.yml are:
-
actual_ok: 0
diagnostics: ' Failed (TODO bug) test
(t/xx-uncategorized/value_ali
as_readonly.t line 47, column 5-89)\n'
name: aliases returned by $pair.values should be rw (2)
<pos:t/xx-un
categorized/value_alias_readonly.t line 47, column 5-89>
ok: 1
reason: bug
type: todo
exit: 0
max: 8
ok: 8
passing: 1
seen: 8
todo: 8
wait: 0
revision: 15894
smoker: ~
timing:
duration: 1434
end: 1176407683
start: 1176406249
I guess that t/xx-uncategorized/value_alias_readonly.t is indeed the
last test, so the problem seems to be the generation of the HTML report.
Can I somehow convert smoke.yml to smoke.html without doing the full smoke?
Cheers,
Moritz
--
Moritz Lenz
http://moritz.faui2k3.org/ - http://sudokugarden.de/ - http://perl-6.de/
Agent Zhang
> I just found out that after the smoke test a new smoke.yml is in the
> repository's root, but no smoke.html.
[...]
> I guess that t/xx-uncategorized/value_alias_readonly.t is indeed the
> last test, so the problem seems to be the generation of the HTML report.
> Can I somehow convert smoke.yml to smoke.html without doing the full smoke?
>
Not sure whether this journal is related or not:
http://use.perl.org/~chromatic/journal/31865
Well, I have no clue if chromatic's patch has been applied on the Pugs side.
Hope this helps.
Cheers, agentz
Agent Zhang
> On 4/13/07, Moritz Lenz <mor...@casella.verplant.org> wrote:
> > I just found out that after the smoke test a new smoke.yml is in the
> > repository's root, but no smoke.html.
>
> [...]
>
> > I guess that t/xx-uncategorized/value_alias_readonly.t is indeed the
> > last test, so the problem seems to be the generation of the HTML report.
> > Can I somehow convert smoke.yml to smoke.html without doing the full smoke?
> >
>
>
Okay, I think I've fixed the bug by defaulting to YAML.pm instead of
YAML::Syck. As confirmed by Ingy++, The current version of YAML::Syck
on CPAN is having problem with single-quotes in its emitter:
http://colabti.de/irclogger/irclogger_log/perl6?date=2007-04-17,Tue&sel=65#l108
I've noticed that this issue didn't affect Larry's recent smokes.
Maybe he was using an older YAML::Syck? or he simply didn't install
YAML::Syck at all?
Cheers, agentz
P.S. Moritz has confirmed on #perl6 that his smoke works again. Yay!
Audrey Tang
在 Apr 20, 2007 7:45 PM 時，Agent Zhang 寫到：
Oops, it was a regression introduced in YAML::Syck 0.84. I've fixed
the issue and released 0.85,
which should again serialize strings containing single-quotes correctly.
Thanks!
Audrey