Re: [bug-gawk] Segfault on invalid program

[Steve Kemp]

> Thanks for the bug report.  Someone reported this recently to the
> bug list; my response is the same; gawk doesn't exit after diagnosing
> the first syntax error in the hope that it can diagnose more than one.

  I can appreciate that approach, but I still feel like a segfault
 is something to be avoided if possible.

> This can lead to data structures becoming corrupted, and thus the
> segfault.

  That does make sense.  Generally I'd regard a segfault as a bug,
 but if I hit similar issues in the future I will avoid reporting
 them to cut down on the noise.

  (Sorry for duplicating this issue, it was only after I sent the
 mail I realized that it would go to a list, and then I saw the
 previous report.)

> I have thought of a way to avoid the segfault and will work on it.

  There was a patch submitted in the original Debian bug report,
 perhaps it is not ideal, as it was written with little knowledge
 of the GAWK internals, but it was sufficient for that specific
 problem, and seemed to cause no averse affects.

  Regardless it is nice to see the new release!

Steve
--