Catch emit_library_call being called with a NULL parameter
Nick Clifton
nickc@cygnus.com
Wed Mar 8 18:24:00 GMT 2000
Hi Richard,
: I consider them equally nice behaviour. Especially in this
: case where the segv comes so close to where the abort would.
:
: What do you see as better from an abort? Would you prefer
: that the C front end inherited the C++ front end's trapping
: of segv, transmuting it to a printed ICE message?
No not really.
It is not that important, but it is just that to me, an abort
indicates that code that generated the trap is actually correct, and
that the bug must lie elsewhere, whereas a seg fault tends to imply
that there is something wrong with the code itself.
I am a big believer in paranoid programming and I like code that
checks all of its assumptions and all of its inputs. That is while I
like aborts and asserts.
Cheers
Nick
More information about the Gcc-patches
mailing list