💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › freenode-kisslinux-2020-10-14.txt captured on 2024-05-12 at 16:07:24.
⬅️ Previous capture (2021-12-17)
-=-=-=-=-=-=-
2020-10-14T00:13:05 #kisslinux <himmalerin> To anyone who hasn't updated clang recently and uses wayland/sway don't update 2020-10-14T00:13:13 #kisslinux <himmalerin> Sway won't start if you do 2020-10-14T04:55:02 #kisslinux <dylanaraps> himmalerin: I cannot reproduce. Send logs? 2020-10-14T05:03:03 #kisslinux <muevoid> I saw the logs I haven't had any issues on latest clang either I just switched over to clang only 2020-10-14T05:04:00 #kisslinux <muevoid> unrelated note does gdb segfault for anyone? I thought it was my filesystem at first but this reinstall with a normal filesystem i have the same issue 2020-10-14T05:07:27 #kisslinux <muevoid> o/ 2020-10-14T05:07:50 #kisslinux <dylanaraps> E5ten: https://github.com/dylanaraps/str/runs/1250268571#step:3:103 + https://github.com/dylanaraps/str/blob/master/str.c#L143 2020-10-14T05:08:13 #kisslinux <dylanaraps> Under glibc snprintf returns two different values for these snprintf calls. Under musl, the values are identical 2020-10-14T05:08:38 #kisslinux <dylanaraps> Function is called here: https://github.com/dylanaraps/str/blob/master/test.c#L137-L139 2020-10-14T05:08:42 #kisslinux <dylanaraps> Any idea why this happens? 2020-10-14T05:09:05 #kisslinux <dylanaraps> POSIX manual says: Upon successful completion, the snprintf() function shall return the number of bytes that would be written to s had n been sufficiently large excluding the terminating null byte. 2020-10-14T05:09:11 #kisslinux <dylanaraps> And: If the value of n is zero on a call to snprintf(), nothing shall be written, the number of bytes that would have been written had n been sufficiently large excluding the terminating null shall be returned, and s may be a null pointer. 2020-10-14T05:09:28 #kisslinux <dylanaraps> muevoid: o/ 2020-10-14T05:11:09 #kisslinux <dylanaraps> First call to snprintf (NULL, 0, "%d", ...) returns '4', second call (buf, len, "%d", ...) returns '8' basically. 2020-10-14T05:11:31 #kisslinux <dylanaraps> But I've allocated memory according to the NULL, 0 call's return value. 2020-10-14T05:11:33 #kisslinux <dylanaraps> Make sense? 2020-10-14T05:18:06 #kisslinux <muevoid> thats strange dylanaraps 2020-10-14T05:22:11 #kisslinux <kinkinkijkin> hey konimex, I am using kiss-llvm and cannot build libelf, i get the error "make: make: don't know how to make elf_version.os. Stop" 2020-10-14T05:22:13 #kisslinux <dylanaraps> The only difference I can see is that I do len + 1 in the second call. That makes the first value '5' rather than '4'. musl still returns '4' but maybe glibc is rounding 5 up to 8 for whatever reason? 2020-10-14T05:22:55 #kisslinux <konimex> kinkinkijkin: you're using KISS or wyverkiss? 2020-10-14T05:23:04 #kisslinux <kinkinkijkin> yes 2020-10-14T05:23:23 #kisslinux <dylanaraps> Send the output of 'kiss s libelf' 2020-10-14T05:23:26 #kisslinux <dylanaraps> kinkinkijkin: 2020-10-14T05:23:34 #kisslinux <kinkinkijkin> i have kiss, and wyverkiss is added as a repo with less importance than kiss-llvm 2020-10-14T05:23:53 #kisslinux <konimex> kinkinkijkin: send me the output of kiss l make 2020-10-14T05:24:11 #kisslinux <konimex> wait, scratch that, kiss s make 2020-10-14T05:24:54 #kisslinux <kinkinkijkin> libelf is in kiss-llvm, wyverkiss, and repo/extra 2020-10-14T05:25:03 #kisslinux <konimex> I have a feeling the "make" you used is bmake and not gmake (Wyverkiss' build script has this edited to explicitly required gmake) 2020-10-14T05:25:05 #kisslinux <kinkinkijkin> make is in wyverkiss, repo/core, and installed 2020-10-14T05:25:36 #kisslinux <kinkinkijkin> alright, ill swap for gmake quickly and try again 2020-10-14T05:25:39 #kisslinux <konimex> yep, you used wyverkiss' make instead of KISS, you need to put wyverkiss priority higher than kiss-llvm 2020-10-14T05:25:54 #kisslinux <konimex> so it should be wyverkiss, kiss-llvm, and repo/extra 2020-10-14T05:29:42 #kisslinux <konimex> also, if you used Wyverkiss, it's redundant to include repo/extra, repo/core, etc. since Wyverkiss already include that with submodules 2020-10-14T05:30:36 #kisslinux <dylanaraps> muevoid: musl (in vasprintf) does the same thing but omits the len1 == l2 check. http://git.musl-libc.org/cgit/musl/plain/src/stdio/vasprintf.c 2020-10-14T05:30:42 #kisslinux <dylanaraps> len2* 2020-10-14T05:32:02 #kisslinux <mcf> dylanaraps: i think you need to re-initialize the va_list (or use va_copy) in between the vsnprintf calls. 2020-10-14T05:33:18 #kisslinux <dylanaraps> mcf: The musl link above uses a single va_copy with no va_start. Thanks for the suggestion, will try it. 2020-10-14T05:34:16 #kisslinux <mcf> that's because it takes a va_list as a parameter, whereas yours is a varargs function 2020-10-14T05:34:30 #kisslinux <dylanaraps> I see that, yeah. 2020-10-14T05:37:52 #kisslinux <dylanaraps> Works. Thanks mcf: 2020-10-14T05:39:28 #kisslinux <mcf> va_list is usually implemented internally as something like `typedef __va_list_struct va_list[1]`, so when you do `va_list ap`, you allocate storage, but if you use va_list as a function parameter, it decays into a pointer 2020-10-14T05:40:17 #kisslinux <mcf> the standard says this: "The object ap may be passed as an argument to another function; if that function invokes the va_arg macro with parameter ap, the value of ap in the calling function is indeterminate and shall be passed to the va_end macro prior to any further reference to ap." 2020-10-14T05:42:29 #kisslinux <dylanaraps> Right 2020-10-14T05:54:34 #kisslinux <kinkinkijkin> now i am trying to build ncurses and come across "sh: cd: line 5: can't cd to include: no such file or directory" and i assume this is a minor packaging issue 2020-10-14T05:55:18 #kisslinux <kinkinkijkin> don't have any repos overriding repo/extra for this one 2020-10-14T05:55:48 #kisslinux <dylanaraps> 'kiss s ncurses' shows repo/extra first? 2020-10-14T05:56:00 #kisslinux <kinkinkijkin> shows repo/extra *only* 2020-10-14T05:57:19 #kisslinux <kinkinkijkin> wait a sec, havent kiss u'd in a couple days since llvm took forever to build 2020-10-14T05:57:45 #kisslinux <dylanaraps> That shouldn't make a difference 2020-10-14T05:57:59 #kisslinux <dylanaraps> Does it fail during ./configure or during make? 2020-10-14T05:58:02 #kisslinux <dylanaraps> Send a build log. 2020-10-14T05:58:03 #kisslinux <kinkinkijkin> during make 2020-10-14T05:58:26 #kisslinux <kinkinkijkin> i gotta figure the logistics to get a build log to you 2020-10-14T05:58:37 #kisslinux <kinkinkijkin> this is on a laptop, installing using grml as the liveusb 2020-10-14T05:58:46 #kisslinux <dylanaraps> nc termbin.com:9999 < /path/to/log 2020-10-14T05:59:01 #kisslinux <dylanaraps> ncurses builds fine here. 2020-10-14T06:00:47 #kisslinux <kinkinkijkin> https://termbin.com/obp0g 2020-10-14T06:04:27 #kisslinux <muevoid> ncursea needs gmake 2020-10-14T06:04:32 #kisslinux <muevoid> ncurses* 2020-10-14T06:04:47 #kisslinux <muevoid> wyverkiss uses netbsd-curses by default which doesn't 2020-10-14T06:05:15 #kisslinux <muevoid> so you could edit build and replace make with gmake 2020-10-14T06:05:36 #kisslinux <muevoid> At least im pretty sure thats problem 2020-10-14T06:06:42 #kisslinux <kinkinkijkin> trying that 2020-10-14T06:07:12 #kisslinux <kinkinkijkin> that worked 2020-10-14T06:07:33 #kisslinux <muevoid> kk 2020-10-14T06:42:34 #kisslinux <E5ten> dylanaraps: did your stdarg fixes fix it? 2020-10-14T06:43:03 #kisslinux <E5ten> Because yeah you can't reuse the arguments without va_end and another va_start 2020-10-14T06:43:33 #kisslinux <dylanaraps> Yeah 2020-10-14T06:43:49 #kisslinux <dylanaraps> In the end I swapped to using __VA_ARGS__ + a function-like macro though 2020-10-14T06:44:12 #kisslinux <E5ten> I prefer that too usually 2020-10-14T06:59:54 #kisslinux <mcf> what's the advantage of using a macro? 2020-10-14T07:02:19 #kisslinux <dylanaraps> No need for va_ stuff as it happens at compile time. macro(arg, arg2, ...) func(arg, arg2, arg3, __VA_ARGS_) -> func(arg, arg2, arg3, arg4, arg5) 2020-10-14T07:02:35 #kisslinux <dylanaraps> I botched that example lol 2020-10-14T07:02:36 #kisslinux <mcf> it is probably a bit surprising that str_printf(&s, "%d", f()) calls f twice, though 2020-10-14T07:02:57 #kisslinux <dylanaraps> mcf: Yeah. I noticed this a few days ago. 2020-10-14T07:03:29 #kisslinux <mcf> also, if you have no arguments, you'd get a syntax error 2020-10-14T07:04:58 #kisslinux <E5ten> No arguments as in no format or no arguments in addition to the format? For macros I always have the format as part of VA_ARGS for that reason 2020-10-14T07:05:20 #kisslinux <dylanaraps> This causes compilation error: str_printf(&s2, "hello"); 2020-10-14T07:05:22 #kisslinux <E5ten> dylanaraps: oh I see you didn't do that, imo you should 2020-10-14T07:05:24 #kisslinux <mcf> yeah, you could combine the format with the arguments 2020-10-14T07:05:29 #kisslinux <dylanaraps> Will do 2020-10-14T07:06:02 #kisslinux <mcf> but still, this is just to avoid using va_list? it gets used eventually in snprintf anyway 2020-10-14T07:06:49 #kisslinux <mcf> in fact, i think it would be a good idea to provide str_vprintf for more flexibility (you could call it from other varargs functions), and have str_printf call str_vprintf 2020-10-14T07:06:53 #kisslinux <E5ten> I think the reason is because it adds a fair bit of clutter to the code, not being against its use because of like overhead or something 2020-10-14T07:07:40 #kisslinux <mcf> the macro adds more clutter in my opinion. do-while loop, and escaped newlines 2020-10-14T07:08:05 #kisslinux <mcf> and you have to put the whole implementation in a header 2020-10-14T07:08:11 #kisslinux <dylanaraps> Fixed, macro. 2020-10-14T07:08:24 #kisslinux <mcf> the double evaluation is my main concern though 2020-10-14T07:08:49 #kisslinux <dylanaraps> Yup. Will see how well str_vprintf idea works 2020-10-14T07:11:04 #kisslinux <muevoid> does anyone here use gdb? 2020-10-14T07:11:12 #kisslinux <mcf> another concern with the macro: what if you one of the arguments you passed to str_printf is named l1 or l2? your arguments would get shadowed by the macro's internal variables 2020-10-14T07:11:25 #kisslinux <mcf> yes 2020-10-14T07:11:35 #kisslinux <muevoid> you do mcf? 2020-10-14T07:11:47 #kisslinux <E5ten> mcf: yeah double evaluation is a problem, in most cases this would be for like a log function or something and the va_args would only be used once 2020-10-14T07:11:52 #kisslinux <dylanaraps> mcf: Another good point. 2020-10-14T07:12:07 #kisslinux <mcf> muevoid: yeah, i use it quite a bit 2020-10-14T07:12:21 #kisslinux <muevoid> You aren't on kiss though are you? 2020-10-14T07:12:51 #kisslinux <mcf> no, but if you have questions about using gdb i might be able to help 2020-10-14T07:13:23 #kisslinux <muevoid> Nah I was using it previously for backtraces but now gdb segfaults for me and im not sure why :< 2020-10-14T07:13:44 #kisslinux <muevoid> maybe ill try using strace on it here in a few 2020-10-14T07:14:40 #kisslinux <mcf> oh. can't help you with that. i guess you'd need to gdb gdb :P 2020-10-14T07:14:46 #kisslinux <muevoid> Lmao yeah 2020-10-14T07:15:10 #kisslinux <muevoid> Ill probs contact the maintainer later to see if they have the same issue 2020-10-14T07:16:23 #kisslinux <muevoid> I do have one question for when I get it working is there a way to set breakpoints in vim and check them with gdb? 2020-10-14T07:17:47 #kisslinux <mcf> i'm not sure since i don't use vim, but i'd be surprised if there wasn't some kind of vim plugin that could do that 2020-10-14T07:18:13 #kisslinux <muevoid> will try to find one thanks 2020-10-14T07:18:28 #kisslinux <dan[m]1> but then again vim is just a text editor, how could it translate the position in src to the binsry 2020-10-14T07:18:33 #kisslinux <dan[m]1> * but then again vim is just a text editor, how could it translate the position in src to the binary 2020-10-14T07:19:18 #kisslinux <muevoid> found one 2020-10-14T07:19:20 #kisslinux <muevoid> https://github.com/albfan/vim-breakpts 2020-10-14T07:19:40 #kisslinux <dan[m]1> cool 2020-10-14T07:19:42 #kisslinux <muevoid> actually no 2020-10-14T07:19:43 #kisslinux <muevoid> nvm 2020-10-14T07:19:51 #kisslinux <muevoid> hold on 2020-10-14T07:19:53 #kisslinux <mcf> you can set breakpoints by the source file and line number, so i guess it could just pass that to gdb somehow 2020-10-14T07:20:36 #kisslinux <dan[m]1> idk, possibly 2020-10-14T07:20:58 #kisslinux <muevoid> i found another one called termdebug which just uses gdb but has them show up in vim 2020-10-14T07:21:40 #kisslinux <mcf> muevoid: you can also use `layout src` to get a source code listing in gdb, and set breakpoints that way 2020-10-14T07:21:45 #kisslinux <muevoid> its actually just in vim 2020-10-14T07:21:53 #kisslinux <muevoid> oh cool i didnt know that 2020-10-14T07:22:07 #kisslinux <mcf> i'm not too familiar with the tui features of gdb though 2020-10-14T07:22:18 #kisslinux <muevoid> do you use a frontend? 2020-10-14T07:22:56 #kisslinux <dan[m]1> would you call gdb a tui? isnt it more of a cli 2020-10-14T07:23:38 #kisslinux <mcf> i use it as a cli, but it does have tui features 2020-10-14T07:23:48 #kisslinux <muevoid> Really? 2020-10-14T07:23:57 #kisslinux <muevoid> Ive never dug that much into gdb to be honest 2020-10-14T07:25:08 #kisslinux <muevoid> On another side note sorry for all the questions do you have experience with valgrind? specifically callgrind tool from valgrind 2020-10-14T07:25:58 #kisslinux <mcf> i use valgrind all the time, but i've only used callgrind once or twice 2020-10-14T07:26:05 #kisslinux <muevoid> Gotcha 2020-10-14T07:26:23 #kisslinux <muevoid> Trying to find some way to visualize the logs in terminal other then just reading the file 2020-10-14T07:26:35 #kisslinux <muevoid> Currently I use gprof2dot which creates a pdf file 2020-10-14T07:54:47 #kisslinux <muevoid> Well webkit2gtk for me now every site I use works fully even gitlab 2020-10-14T07:55:07 #kisslinux <dylanaraps> Nice 2020-10-14T08:03:05 #kisslinux <muevoid> Anyone know a good font looking to switch it up with a nerdfont but not sure which one I should go with 2020-10-14T08:28:38 #kisslinux <muevoid> I think I found one I like 2020-10-14T09:04:17 #kisslinux <muevoid> hey dylanaraps I know you use ALE. Is there a way for it to find files in a seperate directory for example i'm editing something in src/<> can I make it so it can detect files in include/<>? 2020-10-14T09:12:31 #kisslinux <dylanaraps> I set 'g:ale_c_cc_options' to whatever CFLAGS I need 2020-10-14T09:13:15 #kisslinux <muevoid> Thanks! 2020-10-14T09:13:23 #kisslinux <dylanaraps> You can do this via autocmds if you don't want to globally set 2020-10-14T09:13:44 #kisslinux <muevoid> I'll just do it globally for now as most of my projects use the same CFLAGS 2020-10-14T09:13:50 #kisslinux <dylanaraps> 1I also set: let g:ale_lint_on_text_changed = 'never' 2020-10-14T09:13:58 #kisslinux <dylanaraps> And let g:ale_fix_on_save = 1 2020-10-14T09:14:18 #kisslinux <dylanaraps> Some of the tools ale calls can slow things down if they run while you're editing text 2020-10-14T09:14:29 #kisslinux <dylanaraps> Depends on which ones are enabled of course 2020-10-14T09:15:00 #kisslinux <muevoid> Yeah 2020-10-14T09:15:17 #kisslinux <muevoid> for the cflags one do I just set it to a string for example like "-Iinclude" 2020-10-14T09:43:31 #kisslinux <muevoid> I decided I would use a complimation database instead with ccls 2020-10-14T10:36:31 #kisslinux <dylanaraps> mcf: E5ten: Is it worth replacing all pointer dereferences ( (*s)->err ) with a temporary variable ( str *s2 = *s ) or will the compiler just optimize this all away? 2020-10-14T10:37:29 #kisslinux <dylanaraps> (and I shouldn't worry about it) 2020-10-14T13:46:11 #kisslinux <kqz> yay, finally got blender packaged and built on (wyver)kiss 2020-10-14T14:18:46 #kisslinux <konimex> how many dependencies to get blender built? 2020-10-14T14:27:13 #kisslinux <kqz> quite a lot :D but actually not as many as I was expecting, a lot of features are optional and auto turned off 2020-10-14T14:27:34 #kisslinux <kqz> https://git.sr.ht/~kqz/my-repo/tree/master/blender/depends very messy depends file and package 2020-10-14T14:32:11 #kisslinux <kqz> on top of those, I also packaged explicitly for blender opencolorio, openimageio, pugixml, freeglut, glew, pybind11, onetbb, and embree, but still investigating what can be culled 2020-10-14T15:27:19 #kisslinux <micr0> um, is extra/firefox buildable? I'm getting some errors that I think are python 3.9 related 2020-10-14T15:35:12 #kisslinux <Fulton> micr0: try firefox-bin 2020-10-14T15:35:35 #kisslinux <micr0> Fulton firefox-bin won't launch in sway for me, even with GDK_BACKEND=wayland 2020-10-14T15:35:52 #kisslinux <micr0> Error: no DISPLAY environment variable specified 2020-10-14T15:36:06 #kisslinux <micr0> and when I do specify a display (:0 or :1) it won't launch because I don't actually have xwayland 2020-10-14T15:36:35 #kisslinux <Fulton> Oh - interesting, can I see a pastebin of the build error? 2020-10-14T15:37:30 #kisslinux <Fulton> Or termbin if you don't have a browser 2020-10-14T15:37:31 #kisslinux <micr0> Fulton http://ix.io/2AHA 2020-10-14T15:38:10 #kisslinux <micr0> Fulton note that I was able to make it a bit further by manually commenting out some lines in python/mozbuild/mozbuild/configure/__init__.py 2020-10-14T15:38:33 #kisslinux <micr0> basically the for loop had set arg to null or undefined or whatever python's equivalent is 2020-10-14T15:38:38 #kisslinux <micr0> instead of skipping the entire loop 2020-10-14T15:38:49 #kisslinux <micr0> but then I get an error that I am missing libpulse 2020-10-14T15:38:56 #kisslinux <micr0> and I didnt wanna go down the pulseaudio rabbit-hole 2020-10-14T15:39:52 #kisslinux <Fulton> It shouldn't need pulse. give me a second to run a couple of test on my machine 2020-10-14T15:40:46 #kisslinux <Fulton> Never mind it needs rust which takes a solid hour and a half to build 2020-10-14T15:41:12 #kisslinux <micr0> Fulton, if you test, please test with --enable-default-toolkit=cairo-gtk3-wayland as well 2020-10-14T15:41:38 #kisslinux <micr0> Just to be sure, I am upgrading llvm from 10->11 and then rebuilding rust and then firefox to see if that fixes it 2020-10-14T15:42:00 #kisslinux <micr0> llvm takes about the same build time as rust for me, which looks like it will be 45 minutes or so 2020-10-14T15:42:38 #kisslinux <Fulton> If that doesn't work post an issue here: https://github.com/kisslinux/repo 2020-10-14T15:48:29 #kisslinux <micr0> Fulton gonna need a browser to do that or maybe github cli 2020-10-14T16:04:17 #kisslinux <micr0> k, packaged gh in github.com/jedahan/kiss-repo so I can create an issue from the commandline 2020-10-14T16:04:55 #kisslinux <Fulton> You may need a browser to login though 2020-10-14T16:06:38 #kisslinux <Fulton> If it doesn't work I can post the issue and link the pastebin 2020-10-14T16:07:58 #kisslinux <Fulton> You could also use the github app for iOS or Android 2020-10-14T16:08:39 #kisslinux <micr0> I was able to post using the cli - https://github.com/kisslinux/repo/issues/231 2020-10-14T16:09:01 #kisslinux <micr0> thanks for offerring assistance though, appreciate it 2020-10-14T16:32:54 #kisslinux <dilyn> I am having a truly bizarre issue with automake 1.16.1 2020-10-14T16:33:14 #kisslinux <micr0> but its a .1 release - it should be flawless! xD 2020-10-14T16:33:25 #kisslinux <dilyn> I'm building it like how we do in firefox's build, and it doesn't error out at any point. indeed, it successfully reports that it's been installed in the correct directory 2020-10-14T16:33:31 #kisslinux <dilyn> but... it's not there. there's nothing there 2020-10-14T16:33:53 #kisslinux <dilyn> if only micr0... xD 2020-10-14T16:33:59 #kisslinux <micr0> dilyn are you testing kiss 6? 2020-10-14T16:34:05 #kisslinux <dilyn> mmhmm 2020-10-14T16:34:19 #kisslinux <dilyn> but this is separate from kiss. it does this when i do it manually 2020-10-14T16:34:25 #kisslinux <micr0> oh 2020-10-14T16:34:34 #kisslinux <dilyn> it's mindboggling 2020-10-14T16:34:50 #kisslinux <dilyn> it prints stuff like " chmod +x '/home/dilyn/root/.cache/kiss/sources/linux-pam/automake-1.16.1/../junk/share/automake-1.16/config.guess'" 2020-10-14T16:34:58 #kisslinux <micr0> should I just try making automake? 2020-10-14T16:34:59 #kisslinux <dilyn> but there's NOTHING there. there isn't even a folder junk 2020-10-14T16:35:20 #kisslinux <dilyn> I think you should hire an exorcist for this ghost in my machine! 2020-10-14T16:35:33 #kisslinux <micr0> also why .1? there is .2 out I think 2020-10-14T16:35:37 #kisslinux <dilyn> there is 2020-10-14T16:36:06 #kisslinux <dilyn> would you believe me if I told you that linux-pam will error out immediately if you have 1.16.2 instead of 1.16.1? and that's why I'm in this surreal hell hole? 2020-10-14T16:36:44 #kisslinux <micr0> i just untarrd' 1.16.2 in a temp dir and ./configure && make worked fine 2020-10-14T16:36:53 #kisslinux <micr0> but didnt realize you are building automake as part of linux-pam 2020-10-14T16:36:57 #kisslinux <micr0> so not sure how that works 2020-10-14T16:37:49 #kisslinux <dilyn> I'm just not sure how alpine gets away with using automake 1.16.2 here, unless nobody's noticed 2020-10-14T17:38:21 #kisslinux <E5ten> dylanaraps: what's the intended purpose of that? 2020-10-14T17:40:55 #kisslinux <E5ten> dylanaraps: oh I see, not sure if that would have any significant effect 2020-10-14T19:46:51 #kisslinux <himmalerin> dylanaraps: Here's the sway crashing log you asked for yesterday: http://0x0.st/iGIx.txt 2020-10-14T19:47:19 #kisslinux <himmalerin> Maybe I need to rebuild something to work with llvm/clang 11? I tried wlroots and sway w/o success 2020-10-14T19:48:30 #kisslinux <kqz> himmalerin: try rebuilding mesa 2020-10-14T19:51:32 #kisslinux <himmalerin> I'll bet thats it, looks like the main repo's mesa got a revbump for llvm/clang 11 2020-10-14T19:51:40 #kisslinux <himmalerin> rebuilding now, thanks for the suggestion 2020-10-14T19:54:32 #kisslinux <himmalerin> That was it, thank you! 2020-10-14T19:54:41 #kisslinux <himmalerin> Time to add a revbump to my wayland repo 2020-10-14T19:57:33 #kisslinux <micr0> himmalerin are you also maintaining the firefox wayland package? would be nice to have that bumped to 81 2020-10-14T19:58:03 #kisslinux <himmalerin> I'll do that too, thanks for the reminder 2020-10-14T19:58:04 #kisslinux <micr0> I am currently just adding --enable-default-toolkit=cairo-gtk3-wayland to configure in the community/extra repo 2020-10-14T19:58:16 #kisslinux <micr0> <3 2020-10-14T19:58:42 #kisslinux <micr0> honestly if the only difference between wayland native and not is that one line, I wonder if we could get it accepted into community as firefox-wayland 2020-10-14T20:03:19 #kisslinux <dilyn> it would have wayland dependencies which at this point bar that 2020-10-14T20:05:24 #kisslinux <micr0> ahh so its two lines: the above change in build, and adding 'wayland' to the depends. When I check for 'wayland' depends, everything else is in core/extra or community 2020-10-14T20:05:50 #kisslinux <micr0> So first wayland would have to make it into community, then we could have firefox-wayland 2020-10-14T20:05:54 #kisslinux <dilyn> mmhm 2020-10-14T20:06:11 #kisslinux <dilyn> which may one day happen when dylan decides that it's the right time, but for now it remains on the exclusion list 2020-10-14T20:08:21 #kisslinux <micr0> firefox build error involving some file checksum mismatch: http://ix.io/2AJ3 2020-10-14T20:09:31 #kisslinux <himmalerin> same error here 2020-10-14T20:11:24 #kisslinux <micr0> im trying without our current patches to see if that fixes things 2020-10-14T20:15:29 #kisslinux <himmalerin> In the meantime I've pushed the mesa update 2020-10-14T20:28:36 #kisslinux <micr0> so removing the two patches seems to have gotten past the checksum mismatch issue 2020-10-14T20:28:57 #kisslinux <micr0> if it builds correctly, im gonna fork and push to my personal repo and you can decide if you wanna use that for wayland repo too 2020-10-14T20:36:38 #kisslinux <micr0> himmalerin you can look at firefox-wayland from github.com/jedahan/kiss-repo to see what i am building now 2020-10-14T20:37:21 #kisslinux <micr0> its core/extra/firefox with --enable-default-toolkit=cairo-gtk3-wayland added, the patches/ dir removed, and gtk+2 and wayland added as depends. 2020-10-14T21:10:18 #kisslinux <micr0> an hour in, build is still going. generally, this is why I avoid distros with patches that aren't pushed upstream 2020-10-14T21:11:06 #kisslinux <micr0> like, regardless of how insane a project's build system or deps are, messing with that has always ended in pain for maintainers and end-users 2020-10-14T21:20:52 #kisslinux <himmalerin> I'll go look at your repo 2020-10-14T21:21:04 #kisslinux <himmalerin> I wander how dylan builds his wayland firefox 2020-10-14T21:23:02 #kisslinux <dilyn> it's in his wayland-experiment repo 2020-10-14T21:23:44 #kisslinux <himmalerin> Yeah i found it, doesn't look like he patches out gtk+2 for wayland 2020-10-14T21:23:54 #kisslinux <himmalerin> So maybe thats where the checksum error comes from? 2020-10-14T21:24:09 #kisslinux <dilyn> I think he tested that patch on wayland and it worked 2020-10-14T21:24:39 #kisslinux <himmalerin> oh nvm I don't know what I was looking at, he does patch it out 2020-10-14T21:24:51 #kisslinux <himmalerin> he doesn't have the rust 1.47 patch 2020-10-14T21:40:35 #kisslinux <himmalerin> :micr0: I'm fine with re-adding the gtk+2 dep to get Firefox building, hopefully it gets removed from upstream soon 2020-10-14T22:09:16 #kisslinux <micr0> k, updated mesa, bumping firefox to 81.0.2 and seeing if that will build with no patches 2020-10-14T22:09:47 #kisslinux <micr0> it takes about an hour to fail, so I will know in an hour if we need the 1.47 patch or not 2020-10-14T22:10:09 #kisslinux <micr0> sucks that we disable ccache for firefox builds 2020-10-14T22:10:40 #kisslinux <micr0> of note: 81.0.1 failed for me with no patches