we are supposed to receive and return UTF-8 everywhere in the dock. it's valid to do while(*++s) since a nul byte is not permitted in any UTF-8 character (except the NULL one). I also think it's a memory corruption around this function in one of the strings, but I can't find where exactly :-/ 2010/4/23 Tofe > It may be caused by a bad support of some unicode string. Looking at the > error makes me think there's a sort of buffer overflow there: > > applet-amazon.c:_url_encode:220 > > When I look at this location, I see this: > > }while(*++s); // avance d'un cran dans la chaine. Si on est pas à la > fin, on continue... > s = str; > /// la ligne suivane peut planter ... > cd_debug ("allocation of %d bytes...\n", lenght + 1); > t = g_new (gchar, lenght + 1); // Allocation à la bonne taille > > For example, is "while(*++s)" valid for a unicode string? > > Tofe > > On Fri, Apr 23, 2010 at 3:22 PM, Matthieu Baerts