Update of patch #6092 (project pspp):
Status: Ready For Test/Review => Works For Me
_______________________________________________________
Follow-up Comment #1:
It should be safe enough I think. The two error cases are where the file has
more cases than it claims, in which case the "extra" cases will never be read,
which seems OK; and where the file has fewer cases than it claims, in which
case we'll prematurely return end-of-file. In the latter case we should
issue an error message and mark the reader as tainted, I think, but that's
the only addition I'd make to your patch.
Oh, except for one nit: I'd like case_cnt to be a signed type, probably
int32_t, because you compare it against -1. Due to C integer promotion rules
this comparison actually works, I think, but I don't like it.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/patch/?6092>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
_______________________________________________
pspp-dev mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/pspp-dev
