| *** phlixi is now known as Guest1205 | 04:17 | |
| *** phlixi_ is now known as phlixi | 04:17 | |
| *** Mister_Magister_ is now known as Mister_Magister | 07:29 | |
| mal | rubdos[m]: hi, some build issues with your llvm PR, clang spec needs some fixing | 19:17 |
|---|---|---|
| rubdos[m] | mal thanks for the ping | 19:48 |
| rubdos[m] | My threadripper is out-of-order for building, so I'm building on my laptop... i5-5200U. Takes some hours. | 19:48 |
| rubdos[m] | I'm going to try and build LLVM 16 now, I've pushed the two packaging lines. | 19:48 |
| rubdos[m] | That is, without any testing. If you'd rather have me test that too, that's totally fine. I'll be happy to spin up a Rust 1.52 build against llvm-15 just for the fun of it tomorrow. | 19:49 |
| mal | rubdos[m]: thanks, I trigger new build of devel using that, btw mesa needs updating even for llvm-15, I have a branch for that already so you don't need to care about that | 19:50 |
| rubdos[m] | thanks, good to know | 19:52 |
| rubdos[m] | feel free to ping me here or on Github for anything that comes to mind | 19:52 |
| rubdos[m] | I was about to check flypig's Gecko log to check what Rust he might need | 19:52 |
| mal | ok, we'll see how the build goes with the latest version in PR, will take couple of hours probably to build everything | 19:52 |
| mal | I'll check the result tomorrow, initial check later today to check clang build | 19:53 |
| rubdos[m] | thanks! I've started llvm16 locally now, see how far I get. | 19:53 |
| rubdos[m] | I was already surprised that LLVM15 just built like that, of course I forgot about clang | 19:54 |
| rubdos[m] | who uses C anyway ;-) | 19:54 |
| rubdos[m] | Okay so flypig does ESR 91, let's see. | 19:54 |
| rubdos[m] | ESR 91 is officially built with Rust 1.53 and has MSRV 1.51. Anything over FF 94 will require a Rust push. | 19:55 |
| rubdos[m] | So sounds like I will submit Rust 1.54.0 and 1.74. There's no way I'm going to get that last one to behave before EOW, but we'll see. | 19:56 |
| rubdos[m] | My target is to hit ~1.63 asap, and see where I get from there :-) | 19:57 |
| mal | rubdos[m]: just a tiny nitpick about the recent changes, clang-pseudo is not sorted alphabetically but seems some other sorting issues there already so not sure if important | 20:14 |
| rubdos[m] | yeh, I didn't really see a sorting order anymore... So I threw it in somewhere. Want me to resort to re-sorting? | 20:15 |
| rubdos[m] | Will happily do so | 20:15 |
| mal | maybe sort that part most of those are correctly sorted but the stuff around those offload files are wrong | 20:16 |
| rubdos[m] | force-pushed an amend | 20:18 |
| mal | you can squash all changes into one commit, maybe add some mention of reordering to commit message extra lines | 20:19 |
| rubdos[m] | mal: do I push a commit to sort them? It's quite a tiny change. Can also do it in a separate PR if that's cleaner to your eye. | 20:19 |
| rubdos[m] | Just a single commit with a mention about the sorting, then? | 20:20 |
| mal | yes | 20:21 |
| rubdos[m] | ack | 20:21 |
| mal | also format the commit message to [llvm] Update to version 15.0.7 | 20:21 |
| mal | which is the style used in that repo | 20:22 |
| mal | ah, actually [llvm] Update llvm and clang to 15.0.7 | 20:22 |
| mal | like the previous update commit | 20:22 |
| rubdos[m] | done | 20:23 |
| mal | thanks | 20:23 |
| rubdos[m] | Thank you for the quick help | 20:24 |
| rubdos[m] | I've also amended my WIP LLVM-16 according to the above. | 20:26 |
| rubdos[m] | I'm spewing bs. I shouldn't be porting 1.54, since mainline is already 1.61 :D | 20:39 |
| mal | rubdos[m]: heh, so 1.63 should not be difficult if you really need newer than 1.61 | 21:27 |
| mal | rubdos[m]: I had some branch for 1.64, can't remember what issue it had | 21:28 |
| mal | also for 1.66 it seems | 21:28 |
| mal | probably need to test those again at some point | 21:28 |
| rubdos[m] | I recall you writing something on the tracker, so I'll read that again. Thanks :-) | 21:39 |
| rubdos[m] | 1.63 should be doable and should fix our short-term issues. I'd like to jump slightly past it though too | 21:39 |
| mal | as long as gecko still builds | 21:40 |
| rubdos[m] | Apparently, Firefox builds with stable Rust nowadays, so it shouldn't matter | 21:45 |
| rubdos[m] | but that's indeed something to test | 21:45 |
| mal | rubdos[m]: looks like rust needs updating for new llvm | 22:03 |
| mal | some changes in 1.62 but also up to 1.65 seems to be some changes, not sure which is the minimum required one | 22:05 |
| mal | maybe I need to rebase my test branches to see which one might work | 22:06 |
| rubdos[m] | mal: do you mean that the current 1.61 would require changes, or that 1.62+ needs changes? | 22:07 |
| mal | I mean at least some of the errors I see in rust build log were fixed in 1.62, not sure if all were fixed in that version | 22:07 |
| mal | in 1.65 changelog I see mention of "Update to LLVM 15" | 22:11 |
| rubdos[m] | Aha, so I'm already too far in the future for 1.61... | 22:11 |
| rubdos[m] | either way, then we already have the patches ready. | 22:11 |
| rubdos[m] | That was kinda my goal, having a progression of LLVMs, clangs and Rusts ready for when Jolla is ready to merge them :) | 22:12 |
| rubdos[m] | (and then I can document them too) | 22:12 |
| rubdos[m] | LLVM 15 is the minimum required for current Rust master, that's why I started there. | 22:13 |
| rubdos[m] | I'll take some sleep now, I think. My laptop will build LLVM 16 over night :'-) | 22:14 |
| rubdos[m] | Thanks for the assistance, information, and builds and tests mal! Much appreciated. | 22:14 |
| rubdos[m] | Good night! | 22:14 |
| mal | I rebased on 1.64 branch, let's see if it builds | 22:17 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!