[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #18332 [Tor]: Relays should store HS descriptor without the complicated "am I the right one" logic
#18332: Relays should store HS descriptor without the complicated "am I the right
one" logic
--------------------------------------------+------------------------------
Reporter: dgoulet | Owner:
Type: enhancement | Status: needs_review
Priority: Medium | Milestone: Tor:
Component: Tor | 0.2.8.x-final
Severity: Normal | Version:
Keywords: tor-hs, must-fix-before-028-rc | Resolution:
Parent ID: | Actual Points:
Reviewer: | Points: small
| Sponsor: SponsorR-
| must
--------------------------------------------+------------------------------
Changes (by arma):
* status: needs_revision => needs_review
Comment:
Ok, I just uploaded a ticket18332-try3 branch on my side.
It simply removes the call to {{{public_server_mode}}} from inside
{{{rend_cache_store_v2_desc_as_dir}}}, since as you say that call draws in
all sorts of other stuff that the unit tests aren't prepared for.
There's a tension here between simplicity of code, and the belt-and-
suspenders approach we might take to protect ourselves from future
mistakes. I think David gives a good example of how we'd need to hack the
unit tests to survive the belt-and-suspenders approach here. But even in
his hack, I worry that he has one place where he maybe should have set
{{{options_mutable->ORPort}}} back to 0 but didn't, and also I wonder if
there are new things that we malloc but don't free at the end of the unit
tests. So in the argument for simplicity, I think not making more of a
mess of the unit tests is a good balance.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/18332#comment:15>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs