Re: Request Review (testing/helper)
Christian Grothoff <[email protected]> Fri, 27 Sep 2024 14:44:04 +0200
| Newsgroups | gmane.network.gnunet.devel |
|---|---|
| Message-ID | <[email protected]> |
On 9/27/24 14:21, [email protected] wrote: > + // FIXME is this behavior correct/intended? > + // the helper is shut down when there's an exception handler, > + // otherwise it's restarted? > h->exp_cb (h->cb_cls); Yep, that's clearly the idea here, also done everywhere else. The idea is: usually the helper is auto-restarted, but if an application wants to handle errors manually, well, it registers a callback and then does it manually in there. -Christian On 9/27/24 14:34, ch3 wrote: > Hello, > > could someone please have a look at my last commit? > https://git.gnunet.org/gnunet.git/commit/?id=0a0f332664271ea4d53a0276638a407e92b01483 > > Besides general sanity checks, I specifically am curious about: > > 1. Along the lines of the FIXME in helper.c: Is it intended behavior > that a helper gets shut down if it has an exception handler, but if > it doesn't it gets restarted? > > 2. Testing keeps a list of helpers in ns->helpers[]. When one of the > helpers' exception callbacks is called - would it be desirable to > take it out of the array, or is it fine to just mark it as finished? > > Thanks in advance, > ch3 >