< NicolasDorier>
wumpus: your advise on solving my problem by running configure in another folder worked
< NicolasDorier>
MarcoFalke: I think I noticed why my python tests were working before but not anymore. It seems that the cache built by rpc-tests.py is created in BUILDDIR/qa/cache
< NicolasDorier>
is it ok to move them in /tmp/ ?
< sipa>
NicolasDorier: then we'd lose them across restarts
< sipa>
NicolasDorier: though i think wumpus was talking about regenerating them for every run anyway
< NicolasDorier>
oh this would be perfect
< wumpus>
sipa: well my proposal was to regenerate based on matching bitcoind executable, but deterministically regenerating on every run sgtm too
< wumpus>
I think the only advantage of the persistent cache, as currently implemented, is that it can save some time when manually executing a certain test multiple times and you know for sure that nothing relevant to the cache changed. That would be lost when comparing bitcoind hashes just as well.
< wumpus>
apart from that an abstraction like: the test runner could create the cache before it starts running tests, then remove it afterward - would make most sense as les error-prone too
< bitcoin-git>
[bitcoin] jonasschnelli closed pull request #8888: CConnman: Remove some global usage in the gui (master...no-global-connman) https://github.com/bitcoin/bitcoin/pull/8888
< instagibbs>
NicolasDorier, I nuke my cache anytime I get a failed test
< instagibbs>
Sorry, should have suggested this first :)
< jnewbery>
NicolasDorier: I've had a corrupted qa cache before. Took me ages to figure out why all my tests were failing! I'll happily review any PR to make the cache a bit smarter.
< NicolasDorier>
jnewbery: actually I would just like to put it into /tmp. Not only for this problem but for the fact I can't run tests without some workaround anymore since I dev on a shared folder in docker (yeah I know I am the only one having this issue :p)
< bitcoin-git>
bitcoin/master 568c05a Jonas Schnelli: Allow to opt-into RBF when creating a transaction
< bitcoin-git>
bitcoin/master 838a58e Jonas Schnelli: [Qt] Add simple optin-RBF checkbox and confirmation info
< bitcoin-git>
bitcoin/master c4e4792 Russell Yanofsky: [Qt] Change RBF checkbox to reflect -walletrbf setting...
< bitcoin-git>
[bitcoin] jonasschnelli closed pull request #9592: [Qt] Add checkbox in the GUI to opt-in to RBF when creating a transaction (master...pr/grbf) https://github.com/bitcoin/bitcoin/pull/9592
< bitcoin-git>
bitcoin/master 7759aa2 Russell Yanofsky: Save watch only key timestamps when reimporting keys...
< bitcoin-git>
bitcoin/master 3d857f3 Wladimir J. van der Laan: Merge #9818: Save watch only key timestamps when reimporting keys...
< bitcoin-git>
[bitcoin] laanwj closed pull request #9818: Save watch only key timestamps when reimporting keys (master...pr/multirewatch) https://github.com/bitcoin/bitcoin/pull/9818
< BlueMatt>
maybe we should require people who upgrade to read the release notes before we run......
< luke-jr>
BlueMatt: what if they upgrade from 0.10 to 0.15 in one leap?
< luke-jr>
display all of it?
< BlueMatt>
then it'll take you a while to read through it all
< BlueMatt>
(it wasnt a serious suggestion, just noting that everyone has issues with OOM in 0.14 but its in the release notes)
< luke-jr>
tbh, I dislike bundling an ever-increasing archive of release notes with the source code
< BlueMatt>
luke-jr: i wasnt serious, i was just joking....
< luke-jr>
k
< bitcoin-git>
[bitcoin] jnewbery opened pull request #10024: [trivial] Use log.info() instead of print() in importmulti.py (master...uselogginginimportmulti) https://github.com/bitcoin/bitcoin/pull/10024