Remove legacy guard algorithm code
Once we decide to merge the new guard code, there's no real point in keeping the old code around, since it's pretty thoroughly deprecated.
- Show closed items
Activity
-
Newest first Oldest first
-
Show all activity Show comments only Show history only
- Nick Mathewson changed milestone to %Tor: 0.3.0.x-final
changed milestone to %Tor: 0.3.0.x-final
- Author
As we do this, we should grep the code for every call to an instance of a function declared in entrynodes.h or bridges.h, and see if any of those functions can become static.
Trac:
Keywords: N/A deleted, tor-guard added- Author
On my prop271_030_v1 branch, in commit 472a621ccc8a90bf90d8394210519974ad235848, I've wrapped all the legacy code in #ifdefs.
If we plan to remove the legacy code in two alpha releases of 0.3.0 anyway, I would not be opposed to removing it sooner (pre-merge), since I doubt someone is gonna rely on the legacy algorithm for those two alpha releases.
- Author
If it's okay with you, I'd like to do it at-merge, in order to make the merge simpler.
- Author
c52c47ae6f0da5 (merged to master) turns off compilation of this code. There's more to go, though: we'll have to actually remove it.
- Author
Trac:
Status: new to assigned
Owner: N/A to nickm - Author
Planning to do this once 0.3.0.1-alpha has failed to crumble in people's hands for a couple of days.
Trac:
Points: N/A to .1 - Author
Trac:
Points: .1 to .5 - Author
Trac:
Type: defect to task - Author
My branch
remove_legacy_guards
does this.Trac:
Status: assigned to needs_review LGTM!
Please check my branch
remove_legacy_guards
for some more code deletions. I deleted three pieces of code which did:SMARTLIST_FOREACH_BEGIN(guard_contexts, guard_selection_t *, gs) { - if (!strcmp(gs->name, "legacy")) - continue;
Reading the code it should not be possible to instantiate a 'legacy' guard_selection_t anymore, so there is no point in checking for it, right?
Now the only part that mentions 'legacy' in entrynodes.c is:
if (!strcmp(guard->selection_name, "legacy")) { ++n_errors; entry_guard_free(guard); continue; }
I think we might need to keep this there in case someone still has 'legacy' guards in their state file.
Check my branch and feel free to move this to
merge_ready
.- Author
Merged your branch; thanks!
Trac:
Status: needs_review to closed
Resolution: N/A to implemented - Trac closed
closed
- Trac changed time estimate to 4h
changed time estimate to 4h
- Trac moved to tpo/core/tor#20830 (closed)
moved to tpo/core/tor#20830 (closed)