Re: set_field_buffer on 5.3

[Thomas Dickey]

On Wed, 1 Feb 2006, Mike Aubury wrote:

> I'm getting an error with valgrind in my app (which happens to be seg
> faulting ) :
>
>
> ==3977== Invalid write of size 1
> ==3977==    at 0x4041B714: memccpy (in /lib/i686/libc-2.3.2.so)
> ==3977==    by 0x41730832: set_field_buffer (in /usr/lib/libform.so.5.3)
> ..
> ==3977==    Address 0x413E57A5 is 0 bytes after a block of size 41 alloc'd
> ==3977==    at 0x4015F3AF: malloc (in /usr/lib/valgrind/valgrind.so)
> ==3977==    by 0x4172C602: (within /usr/lib/libform.so.5.3)
> ==3977==    by 0x4173091B: set_field_buffer (in /usr/lib/libform.so.5.3)
>
>
> (The field its displaying to is a dynamic field if that matters)
>
> Is there/were there any bugs in there that might be causing this - or am I
> breaking it earlier with some memory errors ?

not that I recall - you might be exercising it differently from the
way I've been doing, and found a bug.  Is that really ncurses 5.3?

All I recall doing in that area is adding some logic for the 
wide-character flavor - no real bug-fixes other than getting 
wide-characters to work reliably.

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net