[15:25] meld sucks and is slow... kdiff3 doesn't have undo. I guess I'll try p4merge? Are there really no great merge tools, only passable ones? [15:37] If it can't be done using diff and patch it's not worth doing [15:38] also: good morning [15:41] well, that simply isn't true. [15:42] what happens when your base changes and the patch no longer applies? start over? [16:04] sorry, I use meld when I need something more than vimdiff [16:04] so nothing better for you, beyondcompare is supposed to be nice but don't have it [16:06] maybe I'll try vimdiff. ugh. [16:06] I think this is the largest conflicting merge i've ever done. it always sucks. [16:19] oh wow... VS Code detects unmerged changes and offers a beautiful gui. [18:37] Yeah, I use vimdiff as well [18:37] but yeah, things can get messy quickly [18:37] Also I think VSCode is borrowing heavily from Atom iirc [18:39] ugh... some tool I used merged wrong and I'm still recoverying. [18:39] gonna be like 8hrs of merging. [18:44] what the hell happened? [18:45] just two big changes to the same codebase [18:45] are they already committed? Sounds like they're so large a human readable diff is worthless [18:46] its not that big. it will get reviewed in gerrit. [18:47] its almost like every part of the patch had merge conflict [18:48] ahh [18:49] and wow, ya'll chose gerrit? we have a love/hate relationship with it here [18:49] as most people have with most code review systems [18:49] I hate gerrit [18:49] :) [18:49] but it is what we use. [18:49] and its probably better than whatever is in BB [18:49] because atlassian is the literal worst. [18:50] I tried to get us to migrate to Diffrential (Phabricator) but there was enough crabbiness from oldtimers we didn't [18:50] :) [18:50] i honestly didn't think atlassian could be that bad until I started using their products. [18:50] I guess someone has to give IBM/Rational a run for worst. [18:50] Wow, that's a nice bar [19:36] finally done. WHEW [19:36] good times! [19:36] only 3.5hrs, not quite as bad as it felt. [19:37] and since greg-g asked, the diff is +3255, -115, but 1303 of that +3255 is generated code, so really only +2000 loc [19:40] 2000?! still huge [19:40] Differential doesn't show diffs that big because it's outside realistic human review [19:41] really? [19:41] its really not that big. [19:42] I mean... mabye that is huge for python or PHP [19:42] that's pretty big, I think when we did limits 800ish was as big as we'd allow [19:42] but for go, remember 75% of that looks like: if err != nil { return err } [19:42] on 4 lines [19:42] lol [19:43] I forget the line number cut off, but here's their reasoning: https://secure.phabricator.com/book/phabricator/article/differential_large_changes/ [20:21] Yipes, taht's a lot of LoC for diffs [20:24] its really not that big. [20:28] https://lwn.net/Articles/737937/ +3407 -30, for comparison :p [20:29] https://lwn.net/Articles/738170/ +3359 -28 for comparison to add all of SMBD to the kernel. LOL