New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Setting the SIGFPE handler to SIG_IGN is problematic, especially when embedding perl #12349
Comments
From andres@anarazel.deCreated by andres@anarazel.deThis is a bug report for perl from andres@anarazel.de, ----------------------------------------------------------------- In PERL_SYS_INIT3 the SIGFPE handler unconditionally gets changed to This generally doesn't seem to be a good idea and actually hits undefined In fact linux ignores setting SIGFPE to SIG_IGN and resets it to SIG_DFL I/We hit this case in postgres' plperl module which embeds perl. Postgres uses Should you have postgres with plperl lying arround you can reproduce the issue andres=# SELECT (-(2::numeric^31))::int/-1; In a situation where perl gets embedded it doesn't seem like a good idea to Do you have any suggestions how to handle this? As far as I can see that behaviour is pretty much the same across all supported Perl Info
|
From @jkeenanOn Fri Aug 24 14:37:39 2012, andres@anarazel.de wrote:
Can you identify the place in the Perl 5 source code where this happens? Thank you very much. |
The RT System itself - Status changed from 'new' to 'open' |
From andres@anarazel.deFor reference, this is the thread discussing the issue on the pgsql-hackers |
From andres@anarazel.deOn Saturday, August 25, 2012 04:35:37 AM James E Keenan via RT wrote:
The callpath (including macros) is something like in v5.17.3-85-g3f6a9d3: PERL_SYS_INIT3 (perl.h) I tried to send that before, but it didn't get through, possibly because I Greetings, Andres |
From @LeontOn Fri, Aug 24, 2012 at 11:37 PM, andres@anarazel.de
I agree this is problematic. I've been meaning to do the git
I'd just set it back to SIG_DFL right after PERL_SYS_INIT
Yeah, it's been in since 2001, and on some OSes longer. Leon |
From andres@anarazel.deLeon Timmermans via RT <perlbug-followup@perl.org> schrieb:
Please excuse the brevity and formatting - I am writing this on my mobile phone. |
From @LeontOn Sat, Aug 25, 2012 at 12:18 PM, anarazel@anarazel.de
Longjumping out of a signal handler is undefined and dangerous (see Leon |
From andres@anarazel.deOn Saturday, August 25, 2012 12:37:20 PM Leon Timmermans via RT wrote:
Thanks, Andres |
From @jhi
No recollection, sorry. I'd look around in the p5p archives to see |
From @LeontOn Sat, Aug 25, 2012 at 11:31 AM, Leon Timmermans <fawaka@gmail.com> wrote:
It seems to me like if we want to do anything (not sure we should), we Leon |
From andres@anarazel.deOn Saturday, August 25, 2012 09:44:38 PM Leon Timmermans via RT wrote:
Andres |
Migrated from rt.perl.org#114574 (status was 'open')
Searchable as RT114574$
The text was updated successfully, but these errors were encountered: