On Mon, Feb 18, 2008 at 07:01:47AM +0100, Nicolas Évrard wrote:When using liferea, I get sometimes weird crashes that appear to happend randomously.Version: 1.4.11-1Has this started to happen recently, or have you seen it in previous versions too? Is there some way you could try to see if an update in some other package is related to this?
No it has been there for quite some time now, but I finally took the time to use liferea-dbg only today (and so far it had only crash once).
Here's the output from gdb: (gdb) bt #0 0x00002abd3c5d2162 in ?? () from /lib/libc.so.6 #1 0x00002abd3c5d430c in ?? () from /lib/libc.so.6 #2 0x00002abd3c5d5960 in malloc () from /lib/libc.so.6 #3 0x00002abd3bbe32bb in g_malloc () from /usr/lib/libglib-2.0.so.0#4 0x0000000000433f7e in NetIO (host=0x2aaab8047c80 "formats-ouverts.org", url=0x2aaab8047b5a "/rss.php", cur_ptr=0x2aaab8044910, authdata=0x0, proto=NETIO_PROTO_HTTP, suppressoutput=0) at netio.c:1096I've taken a look at that code and there's nothing strange there. Moreover, it's code that has not changed since two years ago. - netio.c:1095-1098 -------------------------------------------------- length = recvbufused - (nextstr-netbuf); body = g_malloc(length+1); memcpy(body, nextstr, length); body[length] = '\0'; ---------------------------------------------------------------------- The only thing there that's not quite kosher is that length is an unsigned int, but g_malloc wants an unsigned long. Even if those are different types on your arch there's no way I see that can be a problem, because as far as I can see from the code, length will always be less than 4096. This is weird.
Indeed ... the randomness is also a hint that something strange is going on there. I would input that on faulty hardware if it was not the only app misbehaving on my laptop.
-- (°> Nicolas Évrard / ) Liège - Belgique ^^
signature.asc
Description: Digital signature