<wpwrak>
now we need also need a process that can move a little from A to B. but that "a little" has slightly different characteristics than the things in "everything"
atk has joined #neo900
<Joerg-Neo900>
o.O
<wpwrak>
so we have to see what the impact of these differences is. maybe it's negligible. maybe we hit unexpected issues, like today. maybe it turns out to be very hard, harder than what we solved today.
<Joerg-Neo900>
now that made it absolutely clear ;-)
<wpwrak>
in the latter case, a fallback solution could be to just re-generate periodically or such. that would create certain issues, too, though. so that needs to be evaluated.
<Joerg-Neo900>
please don't overengineer!
<wpwrak>
in any case, atk seems to be optimistic that it's now smooth sailing. i never worked with those hooks while he has.
<Joerg-Neo900>
we don't need hooks to populate the empty ee repo with what's already there in ee-full
<Joerg-Neo900>
we only need to run that job, no?
<Joerg-Neo900>
hooks will be needed when somebody does next push
<wpwrak>
yes
<Joerg-Neo900>
so, why don't we already populate the repo with what we got so far?
<wpwrak>
becasue there are more details missing. e.g., the definition of the committer. the committers influences the hash. if the committer changes, and the repo needs regenerated, all hashes change too. this will cause git to freak out for everyone who has cloned the repo.
<wpwrak>
so as i said, we can do that, but it's not very civilized
<Joerg-Neo900>
ok, so the big job is not ready for primetime. Simple answer
louis__ has joined #neo900
<wpwrak>
precisely
louisdk has joined #neo900
<Joerg-Neo900>
it seems my brain isn't fit anymore for elaborate long answers, neither for "simple" ones with little things and big things
<wpwrak>
a nap may help :) or, in my case, dinner ... and then a nap (plumbers announced their visit for tomorrow. of course, that probably means that i'll wait in vain, once again)
<Joerg-Neo900>
I'm pretty sure it's not a nap that's needed here
<wpwrak>
seep helps to detox (chemically) the brain. that's scientifically proven :)
<Joerg-Neo900>
it's probably I never will adopt that terminology and view to the world that comes with excessive use of git
<wpwrak>
sLeep even
<wpwrak>
yes, it was all easier with steam-driven punchcard readers :)
<Joerg-Neo900>
or simply put: I git this wrong: [2016-12-19 Mon 00:21:09] <wpwrak> yeah ! so we now recreate a perfect copy of the "censored" repo
<Joerg-Neo900>
got even
<wpwrak>
nice freudian ;-))
<Joerg-Neo900>
no, mechanical failure
<Joerg-Neo900>
I got motoric issues
<wpwrak>
yup, we have the copy. now we just have to make sure it stays current. plus a script or two to wrap it all up.
<Joerg-Neo900>
you just said we don'
<Joerg-Neo900>
t
<Joerg-Neo900>
sorry, this is too bizarre for me now
<wpwrak>
anyway, that's for tomorrow
<Joerg-Neo900>
either we got a perfect copy then it's ok to have it available via webgit for public, and it will not even be a drama when there might be a push of new stuff in two days since we could re-initialize and resync the whole thing in no time. Or the copy is not yet perfect then I got a two or three statements wrong
<Joerg-Neo900>
anyway I'd like to have *some* content on webgit rather sooner than later
<Joerg-Neo900>
it's pretty embarrassing to have an empty repo that we made such fuzz about all the time
silviof has quit [Ping timeout: 240 seconds]
cc___ has quit [Ping timeout: 246 seconds]
mzki has quit [Ping timeout: 265 seconds]
xes has quit [Remote host closed the connection]
louisdk has quit [Ping timeout: 256 seconds]
louis__ has quit [Ping timeout: 264 seconds]
knttl has quit [Ping timeout: 264 seconds]
knttl has joined #neo900
xes has joined #neo900
goiken_ has quit [Ping timeout: 250 seconds]
goiken_ has joined #neo900
xmn has quit [Quit: Leaving.]
goiken_ has quit [Ping timeout: 260 seconds]
goiken_ has joined #neo900
xmn has joined #neo900
goiken_ has quit [Ping timeout: 250 seconds]
goiken_ has joined #neo900
xmn has quit [Read error: Connection reset by peer]
xmn has joined #neo900
goiken_ has quit [Ping timeout: 265 seconds]
goiken_ has joined #neo900
Kabouik has quit [Remote host closed the connection]
paulk-collins has quit [Ping timeout: 250 seconds]
xmn has quit [Quit: Leaving.]
neo900 has joined #neo900
Joerg-Neo900 is now known as Guest59525
neo900 is now known as Joerg-Neo900
Guest59525 has quit [Killed (verne.freenode.net (Nickname regained by services))]
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined #neo900
goiken_ has quit [Ping timeout: 250 seconds]
goiken_ has joined #neo900
goiken_ has quit [Ping timeout: 246 seconds]
goiken_ has joined #neo900
tsuggs has quit [Ping timeout: 245 seconds]
goiken_ has quit [Ping timeout: 258 seconds]
tsuggs has joined #neo900
goiken_ has joined #neo900
qwazix has quit [Ping timeout: 268 seconds]
qwazix has joined #neo900
silviof has joined #neo900
goiken_ has quit [Ping timeout: 264 seconds]
goiken_ has joined #neo900
goiken_ has quit [Ping timeout: 248 seconds]
goiken_ has joined #neo900
qwazix has quit [Ping timeout: 268 seconds]
qwazix has joined #neo900
paulk-collins has joined #neo900
goiken_ has quit [Ping timeout: 265 seconds]
goiken_ has joined #neo900
treeman has quit [Ping timeout: 250 seconds]
treeman has joined #neo900
jonsger has joined #neo900
cc___ has joined #neo900
goiken_ has quit [Ping timeout: 252 seconds]
goiken_ has joined #neo900
galiven__ has joined #neo900
galiven_ has quit [Ping timeout: 268 seconds]
Pali_ has joined #neo900
<wpwrak>
Joerg-Neo900: agreed. i was quite surprised to see you guys "pull the plug".
<wpwrak>
how900, atk: now, about setting the committer name. do you know if there is any sort of convention the kind of automated committer we'd have in this case ?
goiken_ has quit [Ping timeout: 268 seconds]
goiken_ has joined #neo900
<wpwrak>
from what i gather, having an e-mail address (at least syntactically) in there, is required. (e.g., the behaviour of "git commit --author=AUTHOR" strongly suggests this)
mzki has joined #neo900
ecloud is now known as ecloud_wfh
jefrite has joined #neo900
SylvieLorxu has joined #neo900
<how900>
wpwrak: it's set with git config user.name foo ; git config user.email foo@bar.com
<how900>
wpwrak: so the git config of the user running the script, or for each commit...
jonsger1 has joined #neo900
jonsger has quit [Ping timeout: 246 seconds]
jonsger has joined #neo900
jonsger1 has quit [Ping timeout: 256 seconds]
Pali_ is now known as Pali
jonsger1 has joined #neo900
jonsger has quit [Ping timeout: 240 seconds]
jonsger1 is now known as jonsger
<wpwrak>
how900: yes, in our case, there isn't really a proper "user" who is the committer (well, if someone runs things manually, then there technically is, but that's spurious). so it would be nice to use something a) less misleading, and b) properly reproducible
<Wizzup>
can I ask approx how many orders does neo900 have atm?
<wpwrak>
how900: could be "whatever <nobody@neo900.org>", or such. so i wonder if there's some precedent for such names we could follow, or if we get to make up our own
<wpwrak>
how900: another option could be to use the author for the committer. but it seems that this would require more elaborate tweaking.
<wpwrak>
maybe something like "Generated from internal repo <nobody@neo900.org>" ? (to indicate 1) that it's not a true source, 2) that it's no a proper subset either, 3) roughly where it comes from. i.e., it's basically what you'd expect of a regular repo, but you'll find quirks if you look really closely)
<wpwrak>
Wizzup: should be around 350, i think
<Wizzup>
cool, thanks
<Wizzup>
so for the final prod you are looking for 450 more
<how900>
Wizzup: yup
<how900>
anyone going to 33c3? We should organize a meeting...
<Wizzup>
I am not - already had other (hard to change/family) plans, and no ticket
<atk>
wpwrak: 350 now? that sounds like it has gone up from 250 since a good while ago.
<atk>
wpwrak: is the target still 500 or has that gone up?
<atk>
wpwrak: I could have a look at what committer info linux has for pre-bitkeeper times
<wpwrak>
atk: it was some ~320 a year ago. so that's +10%, with only collateral PR (i.e., hellekin's newsletters sometimes motivate lurkers. but we haven't really tried to reach out to new people in 2016)
<atk>
ah
<atk>
Linux-2.6.12-rc2 is the earliest commit in mainline
<atk>
They didn't bother importing history before that.
<bencoh>
iirc there'is a history tree
<wpwrak>
grr. so how will i build my 0.12 kernel now ?
<atk>
Just make the committer "Auto Commit <nobody@neo900.org>"
<wpwrak>
bencoh: thanks ;-)
<atk>
I'm impressed that the history stretches that far back.
<atk>
and if you host this on cgit, explain in the "about" page that this tree is a copy tracking the internal tree and why some parts are being temporarily censored.
<atk>
Anyway, I need to go, see you this afternoon.
<wpwrak>
Joerg-Neo900: does "Generated from internal repo <nobody@neo900.org>" for the committer sound okay ? once we've picked a name, i'll process "ee-full" with it and push it into "ee" on neo900.org/git
* ceene
wonders why are you doing such strange things
<wpwrak>
ceene: the filtering ? because joerg doesn't want the kicad_pcb file(s) to appear in the public ee repo. so we need to filter this from ee-full.
SylvieLorxu has joined #neo900
<ceene>
i thought the way to do that was gonna be gitolite partial copy
<Joerg-Neo900>
wpwrak: yes
<wpwrak>
ceene: naw, gitolite is the wrong layer. there, you can hide things from a web browser, but not from git clone
<wpwrak>
ceene: i.e., you can hide it from a lawyer, but not from an engineer :)
<ceene>
partial copy wouldn't copy just master branch, leaving aside all other commits not related to this branch?
<Joerg-Neo900>
I think it's partial-copy is poorly documented, fragile, and depends massively on discipline op committers
<Joerg-Neo900>
of
<Joerg-Neo900>
and you have no control over stuff a posteriori
<ceene>
the kind of loops wpwrak seems to be doing doesn't look any less fragile, though
<ceene>
discipline committer is a real issue, though
<ceene>
i've committed too many things i didn't want to commit
<ceene>
:/
<Joerg-Neo900>
ceene: we simply "copy" the complete repo, and we can apply filters during that
<Joerg-Neo900>
no loops
<ceene>
but i wonder if git format-patch | git am the best way to do that...
<Joerg-Neo900>
well, it seems really simple, standard, and it's already tested and seems to do what we want
<ceene>
don't know, i think i would have used filter-branch
<Joerg-Neo900>
the guys considered all that
<Joerg-Neo900>
don't worry, you won't notice either
<ceene>
in any case, "if it is working don't touch it", as the engineers' maxim says
<wpwrak>
ceene: there are no loops now :) and it's git log, for the precious -m option. that's what makes it all work :)
<Joerg-Neo900>
and it works with scripts, which I like ;-)