< gmaxwell>
sipa: POWER9 has a random instruction... it's called "darn" ... Deliver A Random Number.
< sipa>
haha
< achow101>
fucking git deleted all my worktrees..
< sipa>
achow101: define deleted?
< achow101>
pruned all of the worktrees. the actual directories were untouched. it was mid rebase though so I thought a lost a bunch of commits
< midnightmagic>
Just for future reference: since I was told that the vmbuilder thingy is broken on modern systems I attempted to manually upgrade the gitian base image for bionic -- from the old trusty that I was using, via lxc-execute in the base image. however, the grab-packages script attempts to force a re-download of all installed packages but will fail in an upgrade scenario due to the "deinstall
< midnightmagic>
ok" possible state for some of the
< midnightmagic>
upgraded packages (cgmanager in particular.) Additionally, the query issued by grab-packages will return installed packages which are leftovers from the upgrade -- libapt-inst1.5 for example. These have to be manually apt-get remove'd. gsign on the newer platform now (sigh) gets to deal with a pinentry timeout, but otherwise I think this is about it for doing an upgrade rather than a new
< midnightmagic>
install
< provoostenator>
meshcollider: you're right, it's a seperate field now, but sill in the comment section, so we could move that to the left.
< meshcollider>
I thought thats what Luke's PR did
< provoostenator>
#11803 does the formatting work too, we only to move it to the left
< warren>
midnightmagic: vmbuilder has been broken for me for years now, but fortunately dongcarl is close to making a much better deterministic environment to replace gitian.
< wumpus>
eh re gmaxwell | sipa: POWER9 has a random instruction... it's called "darn" ... Deliver A Random Number.
< wumpus>
achow101: ok
< harding>
achow101: my git worktrees for the bitcoin repository also broke inexplicably two days ago in between my uses of them. I thought maybe something on my system broke them (I use Qubes 4 + Debian 9), e.g. some symlink was lost, but now I'm wondering if it was something else.
< achow101>
harding: i'm also using qubes 4 with debian 9. Maybe a qubes problem..
< harding>
achow101: do you use worktrees with any other projects? I do, and I just checked a few and didn't find any of them broken (even ones on the same qube as my bitcoin repository). I don't know what would distinguish bitcoin from other repositories I use; the only thing I can think of is that it's the largest (size and object count) that I use with worktrees.
< achow101>
harding: i do. I think it's because bitcoin has a lot of objects which the garbage collector cleans up every so often.
< harding>
achow101: interesting. I have another repository on a separate qube that has lots of objects and gets routinely compacted. I don't have a worktree of it, but I'll create one now and check back in on it in a week or so.
< wumpus>
no problems with my worktrees here (ubuntu 18.04)
< wumpus>
(on that system I have a seperate worktree for every branch)
< wumpus>
I think that concludes the feature-merging for 0.18
< wumpus>
the rest in the milestone I'd consider fixes
< sipa>
\o/
< bitcoin-git>
[bitcoin] harding opened pull request #15425: [Doc] add missing newline to listunspent help for witnessScript (master...2019-02-help-listunspent-missing-newline) https://github.com/bitcoin/bitcoin/pull/15425
< meshcollider>
\o/
< bitcoin-git>
[bitcoin] harding opened pull request #15426: [Doc] importmulti: add missing description of keypool option (master...2019-02-importmulti-keypool) https://github.com/bitcoin/bitcoin/pull/15426