Oswald Buddenhagen
Oswald Buddenhagen
it might be bit "too smart" (too subtle, too hard to discover), though. it would be workable if all patterns had an escape prefix to have them matched literally. (and...
this seems like a somewhat radical behavior change which is likely to cause some consternation. please elaborate why you're changing it, and why you've chosen this particular behavior. anyway, i...
i haven't looked at the code too closely, but if b->filename is coming from the message's content-disposition and isn't already sanitized, then you're introducing a shell injection vulnerability.
if the executable doesn't have any side effects (that directly affect systemd, anyway), then invoking it isn't qualitatively different than testing for a file's existence - it's just another syscall,...
i see what you mean. would "pulling up" the logic actually add complexity (or even introduce semantic anomalies), or would it merely move the existing complexity? oh, i clearly missed...