T3s|4
dr|z3d: just restarted into the 'latest' from the console - and ended up with 71dfc155 (Build date: 2023-01-31 03:41:23 UTC) - 2.1.0-5+
dr|z3d
oops
dr|z3d
you're right T3s|4, let me rectify that. I must have uploaded the wrong update from the wrong dir.
T3s|4
np
dr|z3d
offending update.zip nuked so it won't happen again.
dr|z3d
that probably explains why you weren't dancing around in the channel proclaiming -60+ :)
dr|z3d
uploaded.
T3s|4
lols, as was running 60+ earlier, on this same rig :)
T3s|4_
dr|z3d: thanks, the 16:37 update provides your latest iteration of 60+
zzz
been saying this for a while, but this time I mean it... we're getting really close to stomping out the memory leaks
zzz
hopefully within a few days
dr|z3d
great, zzz, noticed the bug report.
zzz
I wish, it's scattered across several
zzz
but only at that point will I get to the real meat of the review, which is figuring out if all this subdb stuff does what it claims to
dr|z3d
nothing short of herculean, what you've been doing.
zzz
thanks, it's been an adventure
zzz
also making some progress on working around the 3 second lag problem on persistence
dr|z3d
excellent news. what's the general issue, have you identified that yet?
zzz
I2PSocket.close() is blocking and requires a round-trip, so don't wait for that
dr|z3d
ah, ok, well good news that you're making progress there.
dr|z3d
T3s|4: new update coming soon, fat fingered the current one.
zzz
I'm a little wrapped around the axle because I have a mix of bug fixes and early phase 2 changes all hopelessly tangled up
dr|z3d
don't know what to suggest other than "commit early, commit often" :)
zzz
never been my style but for long, low-priority projects, I definitely need to do something different
zzz
the problem is I like to always be testing the 50 things I'm working on at once. If they're in 50 different branches I can't do that
not_bob
Another site added to the blacklist!
dr|z3d
well, dev branch, commit to that, and then merge the individual commits when you're happy with them to your main branch.
dr|z3d
merge or cherry-pick, whatever suits.
dr|z3d
or if you prefer multiple branches for each bit of code, then a test branch that you can pull all the various commits on different branches into.
dr|z3d
happy days, not_bob!
not_bob
Also, greetings to all!
dr|z3d
ps. we don't care, just keep squashing them :)
not_bob
;)
zzz
yeah it would be some kindof merge/test branch and yank from all the others every time you build
eyedeekay
Granted it's more like 5-10 mods at a time, but what I do is checkout master, then checkout a new branch i2p.i2p.version-testnumber, then git pull idk branchname --no-rebase every branch I want included
dr|z3d
If I'm juggling various branches, I'll do a git lg (or git log) so I have the details in console, and then just grab the commits I need.
zzz
I'm getting a little better at branches since eyedeekay wants me to do MRs
zzz
but I've never tried to merge or cherrypick
dr|z3d
if it's single commits, git cherry-pick -x <commit> is trivial.
dr|z3d
sorry, git cherry-pick -X ..
dr|z3d
-X makes sure the commit you're merging from is referenced in the commit notes.
dr|z3d
eyedeekay: zzz's got some memory leak code in the wings just waiting for you to green light :)
zzz
but my #1 requirement is to test all WIP all the time; I routinely let stuff soak for weeks or months or even a year before checking it in
zzz
dr|z3d, he and I are on the same page, no need to inject, and I'm not waiting for a green light, please don't muddy the waters :)
dr|z3d
ok, good.
eyedeekay
I'm more waiting to green light it myself today, trying to avoid too many midnight checkins :)
zzz
right
zzz
even 5 seconds of self-reflection of 'what was going on when I did that' even for a minor mistake can help you improve your first-checkin-perfection rate :)
zzz
speaking of self-reflection, it's my turn
zzz
somebody tell me wtf I was doing on May 30, 2015
zzz
because I just found a bad one
dr|z3d
commit code crime?
dr|z3d
*committing
zzz
yeah
zzz
our gunzipper wasn't checking the checksum
zzz
so I just fixed that locally
zzz
and now the checksum is failing in the unit test (((((((((
zzz
so guilty of code crime and test crime
dr|z3d
shocking!
zzz
this is back when I was doing mtn test branches
zzz
if I had to guess, I probably put it aside, then came back later and mistakenly thought I was done
dr|z3d
seeing a huge number of these on one router, maybe normal, maybe not? Warning! Sending direct search message in a client netDb context!
zzz
dunno, ask eyedeekay or enter a ticket
zzz
you were right about the last one
dr|z3d
not seen that one much, if at all, before today. only on one neglected router.