Notorious b.s.d. posted:my posting persona is eternal
|
|
# ? Nov 24, 2019 03:27 |
|
|
# ? May 26, 2024 11:48 |
|
mystes posted:Maybe they insist on running build servers on os x vm's rather than having a secret linux port of their toolchain but they still are willing to use some standard hypervisor? that would be a fascinating outcome, which is why i asked
|
# ? Nov 24, 2019 04:08 |
|
Notorious b.s.d. posted:the license allows virtualization the macOS Hypervisor.framework is just the bits you need to implement virtualization efficiently in user space, without a kernel extension, it doesn’t contain implementations of any virtual devices xhyve is a fork of bhyve atop Hypervisor.framework so it does have devices, I think QEMU and Parallels also support it, same with Docker
|
# ? Nov 24, 2019 04:25 |
|
Tankakern posted:and, for the first time i've found a subject in which the arch wiki is outdated - current state of optimus on linux. fedora's got you covered: https://rpmfusion.org/Howto/Optimus tested it now! i had to replace my xorg.conf (wouldn't be the authentic linux experience otherwise) but now i can set and forget some env variables in my steam.desktop and game away to my heart's content cool
|
# ? Nov 25, 2019 14:58 |
|
|
# ? Dec 2, 2019 16:55 |
|
idk I kind of like that raspberries pi have protected memory and network support
|
# ? Dec 2, 2019 18:33 |
|
im curious if anyone has ever managed to use TempleOS and HolyC professionally e: the lack of networking support would be a problem I guess maybe could be done by running it in a VM and writing an interface layer to the host OS? Private Speech fucked around with this message at 19:26 on Dec 2, 2019 |
# ? Dec 2, 2019 19:21 |
|
Private Speech posted:im curious if anyone has ever managed to use TempleOS and HolyC professionally Well, good news! https://github.com/minexew/Shrine klafbang fucked around with this message at 20:00 on Dec 2, 2019 |
# ? Dec 2, 2019 19:58 |
|
HolyC is an excellent pun at least
|
# ? Dec 2, 2019 20:29 |
|
i see michael keeps us posted over what esr is up to The GCC Git Conversion Heats Up With Hopes Of Converting Over The Holidays
|
# ? Dec 8, 2019 14:11 |
|
an almost arbitrarily bad import in 2017 (when the effort started), as long as it got the tips of branches correct, would by now be way better than this mess.
|
# ? Dec 8, 2019 15:35 |
|
is linux on the desktop yet?
|
# ? Dec 8, 2019 15:38 |
|
pram posted:is linux on the desktop yet?
|
# ? Dec 8, 2019 15:56 |
|
pram posted:is linux on the desktop yet? yes. we did it, we can stop now
|
# ? Dec 8, 2019 16:25 |
|
Tankakern posted:i see michael keeps us posted over what esr is up to it must not be that important because otherwise someone competent would come along and do it
|
# ? Dec 8, 2019 16:50 |
|
akadajet posted:it must not be that important because otherwise someone competent would come along and do it
|
# ? Dec 8, 2019 17:01 |
|
Isn't clang still on Subversion? Maybe they could ask esr to lend a hand.
|
# ? Dec 8, 2019 19:35 |
|
no, they switched to git
|
# ? Dec 8, 2019 19:43 |
|
akadajet posted:it must not be that important because otherwise someone competent would come along and do it someone earlier this year did take over doing it, and ESR is very loudly whining about how they don't trust it
|
# ? Dec 8, 2019 19:54 |
|
Beamed posted:someone earlier this year did take over doing it, and ESR is very loudly whining about how they don't trust it lol seems the new guy quickly whipped up some scripts and is closer to achieving a usable conversion than esr reminder that esr has farted around for two years and begged up enough donations to build a really big computer (lots of cores and ram) which he claimed was required to make progress on the conversion
|
# ? Dec 9, 2019 22:21 |
|
Didn't the other dude literally just use the standard svn to git import scripts that come with git, and this yielded something that was basically usable? Apparently that is not galaxy brained enough for ESR, who managed to find the one real world situation where some dumb algorithms and data structures interview question optimization wank needed to be done (if you ignore the inconvenient fact that git-svn already exists) and then proceeded to poo poo his pants in front of everyone Like what was the big complexity surrounding this import job? Anything in particular about GCC's development processes that require a nonstandard solution?
|
# ? Dec 9, 2019 22:32 |
|
Sapozhnik posted:Didn't the other dude literally just use the standard svn to git import scripts that come with git, and this yielded something that was basically usable? nah apparently he only uses it for a minor part of the conversion and the rest is his own scripts this is what esr is hand-wringing about btw, in his mind the fact that the pre existing migration tool falls over on a repository like gcc means it can’t be trusted at all and even minor use of it taints the conversion. instead you should trust esr’s dilatory wanking around in python that he later translated to go (for performance (lol it’s still slow because esr is bad at this)) quote:Like what was the big complexity surrounding this import job? Anything in particular about GCC's development processes that require a nonstandard solution? a lot of the issues have to do with subversion being poop from a butt. the prebuilt conversion tool assumes popular conventions about how svn projects are structured and branched, and is doomed to fail on anything as old and big as the gcc repo since its guaranteed that with that much history, many people did many nonstandard things reminder that translating svn’s notion of branches to git is nontrivial because svn doesn’t really have branches, just ~special~ subtree copies. if you’ve never used it, branching is literally a svn command to recursive copy a subtree you want to branch, while invisibly marking it such that it can later merge to/from the original subtree. since svn just has all branches live at the equivalent of git HEAD, and each “branch” covers only a subset of the files in the repo, it’s easy to see why anything which breaks conventions for branch structure could be difficult to migrate to git branches in a sensible way
|
# ? Dec 9, 2019 22:56 |
|
i've tried to use git-svn on several occasions and every time it fell over, due to issues described already or because svn decided that certain revs were forbidden to speak of. which hosed up the import ofc. happily it was never my responsibility to fix anything so that's all i know about that
|
# ? Dec 9, 2019 23:03 |
|
quote:if you’ve never used it look man i was young and I needed the money
|
# ? Dec 9, 2019 23:21 |
|
is esrs plan to literally load every single commit into memory before converting?
|
# ? Dec 10, 2019 05:24 |
|
I must be missing something because the implication of that article is he was hamstrung for several months due to the lack of ram for his personal computer, which may be one of the dumbest things I've ever heard
|
# ? Dec 10, 2019 07:07 |
|
abigserve posted:I must be missing something because the implication of that article is he was hamstrung for several months due to the lack of ram for his personal computer, which may be one of the dumbest things I've ever heard does it make more sense if you read it as a justification given by esr for his lack of progress
|
# ? Dec 10, 2019 07:18 |
|
no, because esr is barely even the one writing reposurgeon. he's the least active team member by far lol https://gitlab.com/esr/reposurgeon/commits/master
|
# ? Dec 10, 2019 07:25 |
|
Suspicious Dish posted:no, because esr is barely even the one writing reposurgeon. he's the least active team member by far lol 5,935 commits at some point you gotta step back and think hey maybe this is the wrong way to go about an svn to git migration
|
# ? Dec 10, 2019 08:14 |
|
reposurgeon seems an initially simple project having gone way way out of scope, ton of complexity parsing raw svn data with various heuristics and fixups not necessarily to fit git, but to fit details of the internal model of reposurgeon. seems like if one were to go through all this effort step 1 should probably have been to fork svn which has the by-definition-correct code for all that already.
|
# ? Dec 10, 2019 10:29 |
|
How is esr seen by the people who actually work with him? I don't mean the weird cult-ish and bizarrely numerous commenters on his blog, but people like the GCC and Emacs developers.
|
# ? Dec 10, 2019 10:32 |
|
read up on esr’s involvement with ncurses, fetchmail, and ntpd here’s a taste quote:Eric's response was to make an (unspecified) threat. eschaton fucked around with this message at 10:57 on Dec 10, 2019 |
# ? Dec 10, 2019 10:48 |
|
someone refused to merge a ncurse patch from esr bc it was garbage:quote:Eric replied, saying
|
# ? Dec 10, 2019 11:22 |
|
Zlodo posted:someone refused to merge a ncurse patch from esr bc it was garbage: lömãrf
|
# ? Dec 10, 2019 14:38 |
|
echaton: seems like the graphics guy, vetter, shares your concerns about linux abis, at least for the graphics subsystem https://blog.ffwll.ch/2019/12/upstream-too-little-too-late.html
|
# ? Dec 10, 2019 15:25 |
|
eschaton posted:read up on esr’s involvement with ncurses, fetchmail, and ntpd Pretty funny that RMS of all people had to step in and be the adult, in particular giving social advice on how to cater to ESRs ego. Actually, that last bit surprises me a little. ESR has unwarranted ideas about his own ability, but I never felt he was defensive about them. He has very little insecurity, and is instead just deluded.
|
# ? Dec 10, 2019 18:26 |
|
eschaton posted:read up on esr’s involvement with ncurses, fetchmail, and ntpd lol wow esr is even worse than I thought. he’s the IRL version of that comic where the guy says “I made this”
|
# ? Dec 10, 2019 21:10 |
|
Tankakern posted:echaton: seems like the graphics guy, vetter, shares your concerns about linux abis, at least for the graphics subsystem The graphics supernova will not last forever and is already coming to an end. Each vendor's 2019 model looks a lot like their 2015 model, rtx gimmickry notwithstanding.
|
# ? Dec 10, 2019 21:27 |
|
after like 4 years using arch linux i've come to the conclusion that it's a piece of poo poo and pacman has got to be the shittest package manager of em all
|
# ? Dec 10, 2019 22:00 |
|
|
# ? May 26, 2024 11:48 |
|
i don't know exactly when NVIDIA is going to write off rtx as a failure, like they did 3D displays and VR and all those other things. it should be happening soon though.
|
# ? Dec 10, 2019 22:01 |