In <[🔎] 20090704120642.GH6047@kunpuu.plessy.org>, Charles Plessy wrote: >we have a patch in the Debian package mummer for which we lost origin and >detailed description. I would like to forward it upstream, but I would > prefer to know what it does before ;) Would somebody have a few minutes > to throw an eye on it? It is in our SVN, and here is a copy: > >http://svn.debian.org/wsvn/debian-med/trunk/packages/mummer/trunk/debian/p >atches/01sm_src_tigr.diff > >+ assert ( D_buf = (int *) calloc ( (M+1)*(N+1), sizeof(int) ) ) ; >+ D = &D_buf ; >+ assert ( Op_buf = (char *) calloc ( (M+1)*(N+1), sizeof(char) ) ) ; >+ Op = &Op_buf ; >+ assert ( Show_A = (char *) calloc ( 2*(M+1) , sizeof(char) ) ) ; >+ assert ( Show_B = (char *) calloc ( 2*(N+1) , sizeof(char) ) ) ; Four out of these six lines are bugs waiting to happen. If NDEBUG is defined, the calls to calloc will be completely elided. Then you'll have a number of uninitialized pointers that are used here: > D [0] [0] = 0; > Op [0] [0] = 'a'; This patch is a good idea -- fixed length buffers are rarely considered user-friendly. However, it is executed poorly; it's not ready for upstream in its current state. -- Boyd Stephen Smith Jr. ,= ,-_-. =. bss@iguanasuicide.net ((_/)o o(\_)) ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-' http://iguanasuicide.net/ \_/
Attachment:
signature.asc
Description: This is a digitally signed message part.