https://bugs.exim.org/show_bug.cgi?id=2344
Philip Hazel <ph10@???> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
Status|NEW |RESOLVED
--- Comment #3 from Philip Hazel <ph10@???> ---
(In reply to Carlo Marcelo Arenas Belón from comment #2)
> if the whole objective of it was to avoid the SEGFAULT triggered by
> extending the stack over the soft limit then running instead the following
> in the script might be better:
>
> ulimit -Ss $(ulimit -Hs)
The idea is to check whether pcre2test -S is working; in other words, whether
the code that calls setrlimit() inside pcre2test is working. That code does
check for extending the code over the hard limit (and gives an orderly failure
if that is attempted). So using some other test isn't quite as convincing.
Also I don't know how portable your ulimit command is, nor how portable $(...)
is in various shells.
I'll change that test to use the same value as is used for -bigstack. That
seems logical ... check that -S 64 works before using it for any of the tests.
In fact, -bigstack is only needed (on my box) when compiling with clang and
-sanitize options.
--
You are receiving this mail because:
You are on the CC list for the bug.
