Begasus_32 has quit [Quit: Vision[0.10.3]: Gone to the dogs!]
<Begasus>
not fast enough to capture space bar on boot here ;)
<humdinger>
still doing BIOS/MBR boot here. holding SHIFT is easier
<Begasus>
crossing fingers ...
AlaskanEmily has quit [Remote host closed the connection]
<Begasus>
nope, back on R1B2 but issue remains
<Begasus>
heading out for now, dogschool, cu later (thanks for the hint though humdinger ) :)
<humdinger>
:(
<humdinger>
play with the dogs then... :)
humdinger has quit [Quit: Vision[0.10.3]: Oi, with the poodles already!]
<nephele[x]>
Hmm, I'm on beta3 now, my alt key doesn't work and i cant use my left mouse button until i bring up the team monitor, but this happened before too on the nightly branch (as in right before i updated), so not sure how related that is...
<nephele[x]>
I suppose this is still the same issue with lower version gcc only doing std::filesystem with some magic invocation, i haven't been able to fix that for me, but i am a bit wondering why it seems to work for you though
<PulkoMandy>
nephele[x], you are on a 64bit system, right?
<PulkoMandy>
that could explain differences in the build, I think
<nephele[x]>
Yes, i am
<nephele[x]>
I'm going to try building it with clang 12 now, perhaps that is different, would certainly narrow down if it's indeed the gcc version
<nielx[m]>
is it linking to stdc++?
<nielx[m]>
I don't see it anywhere
<PulkoMandy>
you need to link stdc++fs (a separate library) for this, and yes, it should be doing it already
<nielx[m]>
it's not in the list of objects on the pasted output
<PulkoMandy>
nephele[x], did you do a clean build, or is this after updating an existing build?
<PulkoMandy>
the involved files are cmake/WebKitCompilerFlags.cmake, cmake/OptionsCommon.cmake, and WTF/wtf/CmakeLists.txt, and cmake variables STD_FILESYSTEM_IS_AVAILABLE and HAVE_STD_EXPERIMENTAL_FILESYSTEM
<nephele[x]>
clean build now
<nephele[x]>
with llvm12 i get the same error, it links with gcc too
<PulkoMandy>
stdc++fs should be added to WTF_LIBRARIES if it's needed, not sure what happens of that variable but I expect it should be used when linking WTF into the other libs
<PulkoMandy>
as far as I can see, the cmake scripts are doing things correctly, first trying to compile something without stdc++fs, and if that fails, adding stdc++fs to the linked libs
<nephele[x]>
I added HAVE_STD_EXPERIMENTAL_FILESYSTEM:BOOL=on to the cmake cache, and now i get a different error at least
<nephele[x]>
/boot/system/develop/tools/bin/../lib/gcc/x86_64-unknown-haiku/8.3.0/../../../../x86_64-unknown-haiku/bin/ld: Source/JavaScriptCore/shell/CMakeFiles/jsc.dir/__/jsc.cpp.o: relocation R_X86_64_PC32 against symbol `_ZTVN3JSC25WebAssemblySourceProviderE' can not be used when making a shared object; recompile with -fPIC
<nephele[x]>
/boot/system/develop/tools/bin/../lib/gcc/x86_64-unknown-haiku/8.3.0/../../../../x86_64-unknown-haiku/bin/ld: final link failed: bad value
<nephele[x]>
/boot/system/develop/tools/bin/../lib/gcc/x86_64-unknown-haiku/8.3.0/../../../../x86_64-unknown-haiku/bin/ld: final link failed: bad value
<nephele[x]>
I can add stdc++fs to WTF_LIBRARIES and do a clean build?
<nephele[x]>
also interesting from the initial configure log: -- Performing Test STD_EXPERIMENTAL_FILESYSTEM_IS_AVAILABLE - Failed
<Not-5835>
[haiku/website] pulkomandy pushed 1 commit to master [+0/-0/±1] https://git.io/JcJgy
<Not-5835>
[haiku/website] pulkomandy 1293bde - Attempt to fix markdown a bit on repo change announcement.
<PulkoMandy>
there is a CMakeOutput.log and CMakeErrors.log in WebKiBuild/Release/CMakeFiles
<PulkoMandy>
that could tell more about what failed exactly
<PulkoMandy>
here, this test is passing
<PulkoMandy>
STD_FILESYSTEM_IS_AVAILABLE fails, but STD_EXPERIMENTAL_FILESYSTEM_IS_AVAILABLE passes, and that results in stdc++fs being linked in, as expected
<nephele[x]>
that test fails with the same fpic issue as the build does, maybe that is on clang, i'll do a clean build with gcc again
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
orealis has quit [Ping timeout: 480 seconds]
<nephele[x]>
Forza: in reference to the linked pdf, it sais BeOS was a competitor to os x
<jmairboeck>
Hi, is Ghostscript known to be broken on Haiku? Lilypond nearly works! But I had to make a postscript output from it and convert it to PDF on Linux: http://0x0.st/-pOK.pdf
<Forza>
Apple went to buy next step or beos to be osx. We know they went with nextstep...
ADS_Sr_ has quit [Remote host closed the connection]
ADS_Sr has joined #haiku
Maturion has quit [Ping timeout: 480 seconds]
Maturion has joined #haiku
roiredX has joined #haiku
rodolphoeck has quit [Quit: gone]
<roiredX>
good day
ClaudioM has quit [Quit: leaving]
<PulkoMandy>
nephele[x], both are correct. BeOS started before osX but osX was released in 2001, while BeOS was stil alive. So they did compete at various points (before OSX was even started, and later when it was released)
<PulkoMandy>
jmairboeck, I had ghostscript working well enough for what I needed it for at some point, that was converting some postscript files from Sun to PDF or viewing them, while working on the sparc port
<PulkoMandy>
I did not touch it since and even then I did not test all the features
<jmairboeck>
lilypond produced a weird error, don't know the exact message any more, and ps2pdf manually afterwards also didn't work
<jmairboeck>
I now try building it with gs-api (linking with libgs) instead of invoking it as a command, let's see if this works better
<jmairboeck>
building the documentation also doesn't work, presumably because of the same problem
Maturion has quit [Ping timeout: 480 seconds]
<jmairboeck>
also, I noticed a little packaging weirdness with guile: the guile main package (i.e. guile 2) provides the same binaries as guile1_tools. Shouldn't they be also put into a tools subpackage, and the two tools packages conflict each other? How does the package system handle this when multiple packages contain the same files?
<PulkoMandy>
yes, they should probably be put into separate packages or in the _dev package
<PulkoMandy>
the package system will pick one of the two files if there are conflicts, but I'm not sure it will let you install both packages side by side in this case (because it will see the conflict in the PROVIDES for the two packages)
<jmairboeck>
it did let me install guile without uninstalling guile1_tools first
<jmairboeck>
but I uninstalled that later before trying to use it, so I'm not sure which files were available in between
<Vidrep_64>
What's the best way to test korli's intel_pstates patch? Build the module and put in non-packaged?
<nephele[x]>
PulkoMandy: do you have any other advice to try and figure out if the amd64 build issue for webkit is specific to my machine, or to the architecture?
<nephele[x]>
if the std filesystem and std experimental filesystem checks both fails it still builds but has no way to link it in
Skipp_OSX has joined #haiku
<PulkoMandy>
it seems strange that the check fails, but then everything else is building fine?
<nephele[x]>
Up to the linking of WTF, yes
<nephele[x]>
so the includes are working, the linking just isn't
AlaskanEmily has joined #haiku
jadedctrl has quit [Quit: Vision[znc]: i've been blurred!]
<roiredX>
bye all
roiredX has quit [Quit: Vision[0.10.3]: travelling through KOCMOC!]
jadedctrl has joined #haiku
Begasus has quit [Quit: Ex-Chat]
ADS_Sr has quit [Read error: No route to host]
bbjimmy has joined #haiku
ADS_Sr has joined #haiku
odnes has joined #haiku
<nephele[x]>
Well, defining STD_EXPERIMENTAL_FILESYSTEM_IS_AVAILABLE in cmakecache manually did go by that error, but would fail the linking with the same error the initial test failure failed with
<nephele[x]>
so i assume that is the only problem that needs to be figured out, why the linking fails in the original test
Maturion has joined #haiku
odnes has quit [Read error: No route to host]
ClaudioM has joined #haiku
<PulkoMandy>
normally all code is compiled with -fPIC by default for Haiku
<PulkoMandy>
so I don't understand why you can get this error
<PulkoMandy>
but then, I don't know a lot about x86_64 and I don't really want to learn nmore about it at the moment
vdamewood has quit [Read error: Connection reset by peer]
jmairboeck has quit [Quit: Konversation terminated!]
<Vidrep_64>
PulkoMandy, would our non-working CDDB_lookup be considered a regression, even though it was the CDDB server that changed and not anything in Haiku?
<Vidrep_64>
I'm just getting ready to test out the intel_pstates patch on the test image you linked to. bb shortly with the result...
Vidrep_64 has quit [Quit: Vision[0.10.3]: i've been blurred!]
tuaris has joined #haiku
vdamewood has joined #haiku
<nephele[x]>
Pulkomandy: Okay, i'll try to add fpic to CXXFLAGS and see if that makes a difference
<nephele[x]>
Well, adding that seems to make no difference, hmm.
<nephele[m]>
I'll ask the ml list tommorow
Huggy has joined #haiku
illwieckz has quit [Read error: No route to host]
Vidrep_64 has joined #haiku
illwieckz has joined #haiku
<Vidrep_64>
The Beta 3 test iso with the Intel p_states patch applied now boots to installer/live desktop and installs OK
<Not-5835>
[haikuports/haikuports] diversys pushed 1 commit to master [+1/-1/±0] https://git.io/JcTQq
<Not-5835>
[haikuports/haikuports] enzo1982 d1549bb - boca: bump to 1.0.5 (#6022)
<Not-5835>
[haikuports/haikuports] diversys pushed 1 commit to master [+1/-1/±0] https://git.io/JcTQ3
<Not-5835>
[haikuports/haikuports] enzo1982 3dd3e60 - freac: bump to 1.1.5 (#6023)
<Vidrep_64>
Do we have a preview release of WebPositive available to test, or is it going to be exactly the same as the rebased branch I have been using all along?
Maturion has quit [Ping timeout: 480 seconds]
<nephele[x]>
haikuwebkit on amd64 fails to build vidrep_64
<nephele[x]>
but pretty much all webpositive improvements that were in the nightly are in the beta3 already, and the haikuwebkit stuff is the same as our repo now (when the builder managers to build it anyhow) , though i don't know what you mean by a rebased branch
<nephele[m]>
Compared to the old repo there is at least dark mode support, bigger scrollbars and fixed css rotations (with respect to viewport scrolling)
<nephele[m]>
and other webkit stuff :)
<nephele[m]>
32bit should already be possible to test now, assuming the build there worked (i am assuming it did but havent checked yet :g)
<nephele[m]>
once the haikuwebkit build om the builder is fixed on amd64 webpositive can be tested on amd64
<Vidrep_64>
I'm just cloning the source
<nephele[m]>
i'll go to sleep now, school tommorow, night :)
Vidrep_64 has quit [Quit: Vision[0.10.3]: i've been blurred!]
<Not-5835>
[haikuports/haikuporter] kallisti5 pushed 1 commit to master [+0/-0/±2] https://git.io/Jckem
<Not-5835>
[haikuports/haikuporter] kallisti5 4925b1d - buildmaster/tools: Update for Python3 env
mmu_man_ has joined #haiku
mmu_man has quit [Ping timeout: 480 seconds]
ADS_Sr has quit [Ping timeout: 480 seconds]
mmu_man_ has quit [Ping timeout: 480 seconds]
Huggy has quit [Read error: Connection reset by peer]
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]