[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: assert() crasher in circuitlist.c:414



> Thanks for sending the assert line. Below is a patch
so you can change
> your code and let me know if that fixes it. (Or if
you want us to build
> a new exe with the patch in it, let me know and we will.)

Only win32-compiler I\'ve around is gcc on cygwin. Maybe
it would not be that wise to use it cause it differes
to much with what you used to compile it (could leed to
more/less/other bugs). So, a new exe or some hint about
the used compiler would be nice :)
 
> One question for you, though -- were the [warn] lines
above printed with
> about the same timestamp, or were there many minutes
between them?

[err] and the last [warn] had exactly the same
timestamp. The [warn]\'s before don\'t.
 
> +      log_fn(LOG_WARN,\"Unexpected state %d\",cp->state);
>        tor_assert(0);

There exists a few cases where only a tor_assert(0) got
called. Maybe it would be nice to have there some more
output as well for possible future probs?!
And while we are on it, is there a reason to have the
ugly lines

int err;
FILE *f;
f = fopen(\"d:\\\\foo.txt\", \"w\");
fprintf(f, \"POINT 1\\n\");
fclose(f);

in the function nt_service_body(int argc, char **argv)
at file src/or/main.c? Some kind of checking if d: is
writeable? and if not just crash (no f==NULL check)?

your magicjoe