various fixes #490
various fixes #490
Conversation
jbenet
commented
Jan 4, 2015
jbenet
commented
- lots of logging
- prefix logger
- a few bugfixes
- disable utp
|
rebased on master |
jbenet
force-pushed
the
more-dht-fixes
branch
2 times, most recently
from
ce59ff7
to
733e281
Compare
|
All tests (including dockertest!) mostly* pass. Fixed #488. I'm going to merge this into master as it's way better than what's there now. Note: this does get rid of utp (see #58 (comment)) Still occasionally get these two failures (merging this PR, but we should fix them elsewhere):
i haven't looked into this. i recently changed the
dht is not returning ErrNotFound properly, but instead getting stuck until context (which is pretty long now) expires. the expected behavior is correct. this needs to be fixed. @whyrusleeping could you take a look at these? |
|
Actually, before i merge, looks like jenkins is almost back to passing: https://build.protocol-dev.com/job/all-the-tests/373/ Remaining issue: https://build.protocol-dev.com/job/sharnessss/385/console -- @briantigerchow thoughts? |
|
rebased on top of master again, to get changes in #492 and maybe diagnose sharness problem. |
this commit adds a logger with prefixes
we shouldn't use an arbitrary timeout here. since Get doesnt take in a context yet, we give a large upper bound. think of an http request. we want it to go on as long as the client requests it.
utp is BROKEN!! it causes tests to fail.
@jbenet @whyrusleeping This bug (missing return) could tie up the client worker and cause operations to come to a halt.
Without `-f`, `make clean` fails on machines that don't have the dir. cc @jbenet
this fixes a failing ipns test which didnt have a "working" routing system
TestGetFailures may just be operating very slowly, instead of completely failing. Right now it gets caught on travis often. not sure if its actually wrong.
mocknet indeterminism screwed this test up. that's twice it's bitten us. let's not let it do it a third time. cc @briantigerchow omg.
This is causing test failures because tests don't usually have "/-/-" format. we can decide whether or not to allow keys without validators, but for now removing. cc @whyrusleeping
We use make test as the measure of correctness. This laxity has let bugs creep into several systems. This commit changes our target to always run expensive tests, unless one specifically runs `make test_short` (we would do well to remove most if not all timing-- that's usually what makes tests take a long time.)
Bitswap doesn't usually care about dialing. the underlying network adapter can make sure of that.
