<Begasus>
at least that's the message in solvespace
HaikuUser has joined #haiku
<x512[m]>
Wayland EGL support is needed for GTK OpenGL.
<Begasus>
k, so we're not there yet :)
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
<Begasus>
PulkoMandy, thanks on the update on webkit, much better now (allthough websocket takes quite some resources from CPU's)
<PulkoMandy>
Begasus: please make a bugreport for that, if you can identify websites that trigger it, it would be very helpful
<PulkoMandy>
It happened once during testing but I couldn't reproduce it
<Begasus>
k, will try to gather some information, mainly got the forum/github/google search, open
<Begasus>
my guess it's with the forum, as it refreshes in the background?
mmu_man has joined #haiku
<Begasus>
or too many tabs open? :)
lkw has quit [Ping timeout: 480 seconds]
lkw has joined #haiku
<nielx[m]>
Begasus: thanks for the test, will update the PR
<Begasus>
np nielx[m], if you got something to test I can launch a build here
<nielx[m]>
PR updated
<nielx[m]>
Appreciated if you could give it another go
<Begasus>
k, build started now, thanks!
<Begasus>
(on that laptop (if it keeps it up) it probably will be done somewhere this evening or tomorrow) :)
<nielx[m]>
I will keep that in mind :-)
<nielx[m]>
I am still wondering whether it is right to build everything twice for jit
<nielx[m]>
But I haven't the time to do any actual testing on whether GCC really is slower
<nielx[m]>
well time maybe, energy no
<Begasus>
if build time isn't an issue, it could be narrowed down to 1 build for all, if performance is an issue it's better to keep it seperated I guess?
<Begasus>
build time/performance ...
<nielx[m]>
Well the argument in the packaging instructions is that building the compiler with --enable-host-shared will enable position independent code, which slows down the executable by 5%
<nielx[m]>
It is one of those cases where a 5% slowdown is not much in most cases, but in aggregate it cost a lot.
<Begasus>
the executable as in "gcc"?
mmu_man has quit [Ping timeout: 480 seconds]
<Begasus>
if things are fine, it only needs to be build once on buildmaster, so that shouldn't be a blocker imho
<PulkoMandy>
nielx[m]: executables on haiku are always position independent, we don't support anything else
<PulkoMandy>
So rebuilding gcc multiple times won't change that, I guess?
<nielx[m]>
In that case, we can simplify the build a lot
<nekobot>
[haiku/haiku] 26bda37ca592 - Update translations from Pootle
<Begasus>
is there a difference in Framebuffer driver vs VESA driver?
<Begasus>
ScreenPrefs shows framebuffer (seen vesa there in the past too)
<Begasus>
KERN: app_server: Finding best mode for * (2 resolutions) and then KERN: app_server: Use 2576x1118 (8) instead in syslog
<PulkoMandy>
Framebuffer is for uefi, vesa is for bios machines
<PulkoMandy>
In vesa you can chage resolutions without rebooting
<Begasus>
ah, thanks
<Begasus>
resolution is native so that's fine
<nielx[m]>
Going to update the gcc recipe Begasus. I will retest/rebuild on x86_64, can you restart your x86 build?
<Begasus>
sure nielx[m], only am running it for about 40minutes :P
<Begasus>
give a yell when you're done, need to pull the changes then
mmu_man has joined #haiku
Diver has joined #haiku
<nielx[m]>
done
<Begasus>
build restarted too :)
<Begasus>
build on 64bit here starting too
<nielx[m]>
ah in that case I will defer to your test
<Begasus>
first run enabling jit was fine, so should be ok too (as I also just did the build once)
<Begasus>
Web+ still OK for now PulkoMandy, scrolling feels smoother also :)
AlaskanEmily has quit [Remote host closed the connection]
<Begasus>
nielx[m], saw it checks for iconv, any reason for not including it in the BUILD_REQUIRES?
mmu_man has quit [Ping timeout: 480 seconds]
<nielx[m]>
I don't know, but my guess is that historically it worked so it was never considered. I do not know if iconv gives any advantage
<Begasus>
don't know too, just saw it passing by ...
<ynoga99[m]>
x512: question out of the blue, do you happen to remember having an issue generating the libsolv_*source* package when bootstraping the riscv64 build ?
<ynoga99[m]>
(I've managed to work around it, so it's more out of curiosity)
mmu_man has joined #haiku
tqh has joined #haiku
mmu_man has quit [Ping timeout: 480 seconds]
lkw has quit [Ping timeout: 480 seconds]
<Begasus>
to support the -finput-charset option (for iconv) (but that's from a thread for gcc4.* ;) )
<Begasus>
build finished succesfully on 64bit nielx[m] :)
mmu_man has joined #haiku
<Begasus>
no mention for iconv when running ./configure --help
<nielx[m]>
great news!
<Begasus>
will post an update when 32bit is finished (one way or the other) :)
<Begasus>
syslibs(devel) and fortran don't depend on the base gcc package?
<Begasus>
ah, no seperate package for jit also I see
<nielx[m]>
I think that that's correct
<nielx[m]>
for syslibs, the dependency is the other way
<nielx[m]>
I do not know enough about fortran to make any judgments, but since it gets bundled with its own binary (gfortran), I am assuming that this can work independently from the base gcc package (which contains the C, C++ and Objective C compiler)
<Begasus>
yeah, should be ok, installed now (backed up previous version) to do some checks
<nielx[m]>
Though it also might be the case that there is an implicit dependency, but since gcc is so ubiquitous nobody notices this
<nielx[m]>
good to check, on my machine I rebuilt the GCC 13 recipe with previous iterations of the GCC 13 package, so I am fairly confident
<Begasus>
yeah, previous check worked out before too, but didn't had it installed then
<Begasus>
hmm ... no mention for jit when running configure on emacs
<Begasus>
on the left the newly build package, on the right one I had done before
<nielx[m]>
is there a gcc_jit package though?
<nielx[m]>
the new recipe should create gcc_jit and gcc_jit_devel packages
<Begasus>
no
<Begasus>
grabbing gcc-13.1.0_2023_06_20-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/gcc-13.1.0_2023_06_20-1-x86_64.hpkg
<Begasus>
grabbing gcc_debuginfo-13.1.0_2023_06_20-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/gcc_debuginfo-13.1.0_2023_06_20-1-x86_64.hpkg
<Begasus>
grabbing gcc_fortran-13.1.0_2023_06_20-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/gcc_fortran-13.1.0_2023_06_20-1-x86_64.hpkg
<Begasus>
grabbing gcc_syslibs-13.1.0_2023_06_20-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/gcc_syslibs-13.1.0_2023_06_20-1-x86_64.hpkg
<Begasus>
grabbing gcc_syslibs_devel-13.1.0_2023_06_20-1-x86_64.hpkg and moving it to /Opslag/haikuports/packages/gcc_syslibs_devel-13.1.0_2023_06_20-1-x86_64.hpkg
<nielx[m]>
That's very odd, are you sure you got all the commits from my PR?
<Begasus>
doesn't seem like I did :)
<Begasus>
arghhh ....
<Begasus>
sorry for the noice ;)
<nielx[m]>
no worries :-)
<Begasus>
lol
<Begasus>
back to building ... :)
<Begasus>
too bad it needs to rebuild the lot
<Begasus>
nielx[m], sure both --enable-shared and --enable-host-shared should be used? ;)
mmu_man has quit [Ping timeout: 480 seconds]
tqh has quit [Remote host closed the connection]
<nielx[m]>
Yes, --enable-shared is just the flag that the platform supports shared libraries. The --enable-host-shared is used to tell the build system to use it for GCC itself, which as PulkoMandy said previously, is moot on Haiku because all our executables are built under this presumption
<Begasus>
oki +1 :)
lkw has joined #haiku
mmu_man has joined #haiku
<Begasus>
k, round 2 :)
<Begasus>
configure: error: The installed libgccjit failed to compile and run a test program using
<nielx[m]>
ok, that's something, anything in the config.log?
<Begasus>
ld: cannot find crtbeginS.o: No such file or director
<Begasus>
there is a test exaple linked in config.log mentioning that file prior
<Begasus>
move to gcclibDir looks ok, the links are broken
vdamewood has joined #haiku
mmu_man has joined #haiku
<Begasus>
although the others all provide link to gcclibDir and have the libraries in libDir
<nielx[m]>
I think what's happening in this case is that libgccjit is figuring out where it is in the file system, and then it adds that path to the library path, which then allows it to find crtbeginS.o
<HaikuUser>
how can i get my gpu working? i have a rx570
<augiedoggie>
either disable the driver for it and use the built-in framebuffer, or try some highly experimental stuff that you have to build yourself
<HaikuUser>
i dont even know is it working or not rn
<augiedoggie>
if you can boot to a working desktop then that's about as good as it gets
tqh has quit [Quit: Leaving]
<HaikuUser>
i was thinking about installing it to my old(about 8 years old) laptop but somewhat it doesnt give screen :/
<augiedoggie>
you probably have to enter the bootloader and select the "fail-safe graphics" option
<HaikuUser>
no i meant
<HaikuUser>
laptop doesnt work
<HaikuUser>
no post
<HaikuUser>
fan works for a few seconds and just stops
<|cos|>
HaikuUser: tried leaving it charging for 24h or so?
<HaikuUser>
yes
<HaikuUser>
battery is full
<HaikuUser>
i also tried powering up without the battery but result is same
<HaikuUser>
laptop is the "acer aspire es1-331
<|cos|>
that model seems to have pre-efi bios (i always blame efi if getting the chance), but you're not able to reach setup by pressing F2?
<HaikuUser>
mo
<HaikuUser>
no*
<HaikuUser>
it doesnt give screen(panel is fine)
<HaikuUser>
i tried bios resetting
<|cos|>
hope you'll manage getting it working, somehow! i'm off to bed, late as it is already...
<HaikuUser>
what is your timezone?
<|cos|>
CET
<|cos|>
or rather CEST, this time of yea
HaikuUser has quit [Ping timeout: 480 seconds]
HaikuUser has joined #haiku
HaikuUser has quit []
mmu_man has quit [Ping timeout: 480 seconds]
FreeFull has quit []
Halian has quit [Remote host closed the connection]
mmu_man has joined #haiku
Halian has joined #haiku
<Halian>
waddlesplash: progress! with hrev57156, I get one screenful of on-screen debug info before Haiku freezes; I'm updating the ticket accordingly. I'll get the stuff for serial debugging when my next disability check rolls in.
<waddlesplash>
sounds ... random, I don't think anyone changed the bootloader
<Halian>
\(o.O)/
<waddlesplash>
make sure you disable "on screen paging"
<Halian>
Oops
<waddlesplash>
see if it continues further
<Halian>
Will do
Halian has quit [Remote host closed the connection]
Halian has joined #haiku
<Halian>
waddlesplash: so if I disable on-screen debug paging, I get no output whatsoever before it freezes o_o (updated ticket accordingly)
<waddlesplash>
now that's interesting...
<Halian>
^
Diver has quit [Quit: Leaving.]
Diver has joined #haiku
Diver has quit []
jess has quit []
smalltalkman has joined #haiku
<Halian>
waddlesplash: do you have any ideas that don't require a serial cable, by chance?
<Halian>
(I should also notice that if I disable APIC, SMP, etc., I get no debug output, either.)