| |
| |
| |
|
Page: 1 2 3 4 5 6 7
Comments:
<0> Genesis: OK, I'm rebuilding now with the --disable-accessibility option removed from mozconfig
<1> mikearthur: toshiba gfx chip iirc
<2> hrw|work: cheers
<3> mattcamp: http://rafb.net/paste/results/POaC2b58.html
<4> What is the distro file for?
<5> hrw|work: i have the feeling i fundementally don't understand the process of building a toolchain, but now it's trying to build glibc-intermediate_2.4 and the configure complains about the compiler not having threads support
<1> giel: so do I sometimes
<5> :)
<6> giel: building for non arm?
<5> XorA: no, arm
<5> i have seen this stuff before and i thought i figured it out, but i forgot how :(
<6> giel: eneabling nptl?
<5> XorA: i don't know, should i? isn't it on by default? what is it?
<7> giel: you want to build a glibc 2.4, NTPL, EABI toolchain?
<5> zecke: no, i want to build a non-eabi toolchain, for the rest i don't really care
<5> zecke: because in my understanding i need a newer kernel for eabi
<5> ah nptl is hte thread stuff
<5> i need threads
<7> giel: you want to build glibc2.4 which only has NTPL support
<7> giel: in contrast to < 2.4 which allowed NTPL and LinuxThreads
<8> upgrading 2.6.10 -> 2.6.16/17 should not be implossiblem..
<7> giel: for glibc2.4 your cross initial compiler needs thread support enabled but more I don't know
<5> suihkulokki: yes it is not impossible, but too much effort for now... heaps of out-of-tree drivers
<7> anyway need to finish stuff for IBC
<5> zecke: i do not necessarily want 2.4, but the 2.3.5+cvs and 2.3.3 ones errored out with something about an ***embly error
<7> giel: then take a look at openzaurus-unstable.conf
<7> giel: and set a preferred version of binutils
<5> will do, thanks for the pointer
<7> giel: binutils 2.17 does not work with old glibc :}
<5> is the preferred version of binutils also in openzaurus-unstable?
<5> zecke: ah! thanks!
<7> giel: yes, it is in openzaurus-unstable, there are two lines you can copy into you local.conf
<5> zecke: thanks a lot! will try!
<2> hrw|work: what would the parent of the tosa touchscreen device be?
<5> zecke: btw, we should meet up at the IBC exhibition!
<5> zecke: or aren't you coming?
<7> giel: I won't be there, just need to pack stuff, update phones :(
<7> *back office*
<9> mikearthur: It doesn't have to have a parent
<5> zecke: good luck
<0> koen: What's at the URL that you sent me a few minute ago:
<2> RP: ok, but should I make one anyway?
<1> mikearthur: I'm not familiar with kernel code
<2> hrw|work: no worries
<2> hrw|work: you just seem to know whats going on :D
<0> koen: Oh, I see, it's an strace dump.
<1> koen: index.php | 44 ++++++++++++--------------------------------
<0> koen: Were you running firefox or regxpcom?
<4> in a kernel bb file, how can I run some machine specific stuff to ***emble the kernel into an inage suitbale for falshing?
<3> mattcamp: firefox-bin
<1> Crofton: do_deploy() { if MACHINE == 'my machine' then do magic;fi}?
<7> giel: have you seen http://planet.linuxtogo.org/ ?
<4> ok
<1> Crofton: look at packages/linux/linux-openzaurus.inc
<7> Crofton: but your override should work :}
<7> unless we call do_deploy without doing update_data
<3> mattcamp: regxpcom breaks at the same point
<0> koen: Are you sure the problem is with a dlopen call? Based on the strace output (in particular, the fact that it crashed right after loading a JS file), I suspect a problem with xptcall. xptcall has segfaulted on ARM before.
<3> mattcamp: I'm not sure anymore
<3> mattcamp: the dl_open stuff was probably a remnant of having debug files
<0> I will try to debug it on my device.
<3> which are gone now
<1> koen: pushing new stuff for you
<3> hrw|work: yay!
<0> koen: Incidentally, I've worked with Mozilla on XScale-based hardware before, in a project I did about two years ago.
<10> hrw org.oe.dev * rb7f351a... / (1 contrib/feed-browser/index.php): feed-browser: no need to keep two copies of package listing generation code
<10> hrw org.oe.dev * rbefec029... / (1 contrib/feed-browser/index.php): feed-browser: split maintainers to separate lines (works when ', ' is used as separator
<0> koen: At that time, I had to use gcc 2.95.x because xptcall crashed when I used gcc 3.x.
<3> we had minimo running a while ago
<0> koen: But OE has a patch for that; maybe that's broken in FF 1.5
<3> and FF 1.0
<6> firefox 1.0.7 series run fne
<1> firefox 1.0/1.5 are obsolete
<5> zecke: no, will look at the page
<5> zecke: it seems this is working now, it's still compiling, thanks a lot!
<5> zecke: i would never figured out by myself i needed an older binutils
<5> zecke: ah your blog! will read
<5> dvbsnoop sounds good!
<7> ;)
<7> in contrast to dvbnet it is not oopsing your machine
<5> hehe
<5> zecke: i've got a custom tuntap like thingy
<5> zecke: a kernel module that makes an ethernet device and a device node
<5> zecke: write to the device node, it ends up in the ethernet device
<5> zecke: that's what i use instead of the dvbnet mess
<7> giel: I need to learn the 'dvr' API and finally I'm able to get the full transport stream on my Apple notebook and x86 desktop
<7> I wonder if implementing the MPE FEC is any fun
<5> zecke: yeah, i'm supported by philips semiconductors for that stuff, has advantages and disadvantages, at least i have some hardware and some more-or-less working stuff, but their code is quite horrible
<5> zecke: they really are linux noobs
<0> Has anyone considered alternatives to hard-coding --disable-accessibility in Firefox's mozconfig file? Some users need this feature.
<1> http://openzaurus.org/wordpress/2006/08/24/feed-browser/
<11> hrw|work: cool
<1> cu
<3> mattcamp: someone was probably a bit overzealous with space-saving
<3> hrw|gone: http://tango.freedesktop.org/Tango_Icon_Gallery has some nice icons for feedbrowser
<0> koen: Agreed. I've removed that option from my copy.
<0> koen: If size is really a problem, I think moving components/libaccessibility.so to a separate package would be an acceptable compromise.
<3> yeah
<3> hrw|gone: http://ewi546.ewi.utwente.nl/tmp/fb/?action=details&pnm=a52dec (note the icons)
<11> hey koen, this you gave shows a52dec 0.7.4-r0 twice (once for OZ 3.5.4 and once for OZ 3.5.4.1). isn't that a bit confusing? looks good though
<3> pH5: it is
<3> pH5: it should include the feed name
<3> (or we shouldn't include multiple releases)
<11> oh well, s/this you/this example you/ :)
<11> koen: I'd prefer option (b), if the two versions are identical (and they should be)
<3> pH5: is the dbus-glib cset in already?
<0> koen: regxpcom doesn't segfault for me.
<11> koen: only the part that makes old dbus packages provide dbus-glib
<0> koen: What hardware and what toolchain are you using?
<3> mattcamp: glibc 2.4/gcc4.1.1/EABI pxa270
<11> koen: I'll check in the change for some gpe packages
<0> koen: Ah, OK. We're using glibc 2.3.5+cvs20050627 and gcc 3.4.4.
<12> Hello
<0> koen: Does Firefox 1.0.7 work with the same toolchain?
<11> hi sirfred
<3> mattcamp: haven't really tried
<3> mattcamp: building that one now
<13> so once i've got an updated oe database from mtn, how do i merge it with the existing tree?
<0> koen: I think it would be worthwhile to try. I suspect that the xptcall implementation for ARM is broken when compiled with gcc 4.x, as it was for 3.x about two and a half years ago.
<13> if i try to checkout, it errors stating the folder already exists
<0> koen: Anyway, Firefox 1.5 now runs with my toolchain. (Granted, I can't see it because my device has no screen and my text-to-speech code for Firefox isn't working yet, but that's another story.)
<3> :)
<14> I want to see how to add usb support for the PXA270, I need USB m*** storage and USB serial converter. Is there a good example I can go and look at in one of the existing kernels in OE?
<11> Gerrath: the modules exist. if the pxa270 is the host, you need uhci-hcd, usb-storage and usb-serial, if it's on the other end you need pxa27x_udc, g_file_storage and g_serial
<14> pH5, I'm ***uming I can just compile them all in so there is no need to load/unload modules?
<11> Gerrath: not for the gadget drivers, only one of g_file_storage and g_serial can be loaded at a time.
<10> pH5 org.oe.dev * rf115a3c... / (5 files in 2 dirs): gpe-autostarter: depend on dbus-glib
<10> pH5 org.oe.dev * rc3e7c314... / (5 files in 2 dirs): gpe-beam: depend on dbus-glib
<14> pH5 ok, but if I'm just doing a generic serial and a generic m*** storage then compiling it in should work, correct?
<10> pH5 org.oe.dev * r1935a... / (4 files in 2 dirs): gpe-calendar: depend on dbus-glib
<10> pH5 org.oe.dev * rbb532edd... / (1 packages/gpe-contacts/gpe-contacts_0.45.bb): gpe-contacts-0.45: depend on dbus-glib
<10> pH5 org.oe.dev * r7f913... / (4 files in 2 dirs): gpe-filemanager: depend on dbus-glib
<11> Gerrath: yes. I compiled in pxa27x_udc + g_ether here and it worked just fine.
<10> florian org.oe.dev * rc3e8231... / (1 packages/libhal-nm packages/libhal-nm/libhal-nm_0.0.2.bb): libhal-nm: add 0.0.2 by Milan Plzik.
<10> florian org.oe.dev * ra83944... / (1 packages/hicolor-icon-theme/hicolor-icon-theme_0.9.bb): hicolor-icon-theme: add 0.9 (by Milan Plzik)
<10> florian org.oe.dev * rb5047198... / (5 files in 4 dirs): libnl: add 1.0-pre6 based on an earlier version by Milan Plzik
<10> florian org.oe.dev * rc62f25e... / (8 files in 4 dirs): dhcdbd: Add 1.14 by Milan Plzik
<14> pH5, good, eventially I want to do it modular but it (ipk) wants to install the kernel as well since the kernel is not accessable via the rootfs and I don't want to deal with that issue yet :-)
<3> Gerrath: empty the kernel package
<11> Gerrath: for ipaq-pxa270 it just creates an empty k.. as koen said :)
<14> koen, great, man that is simple. can't believe I didn't think of that one :-/
<14> pH5, I see the option for the OHCI option in the kernel config but not the UHCI, do you know what it is labled as in the kernel config?
<11> Gerrath: my error, ohci is correct for the pxa270 usb host.
<14> pH5, great, thanks.
<13> note to self: before updating oe.dev tree... back-up your local.conf file
<13> grr
<15> bigd0g: er? your local.conf shouldn't be in the tree....
<13> it was in the conf folder. that's the only way i could get bitbake to run it. i'm sure there's another way?
<11> bigd0g: it looks for conf/local.conf in every folder of your BBPATH environment variable
<13> pH5: ohhh
<3> heh
Return to
#oe
or
Go to some related
logs:
linux emacs ddd error csu/crti
#perl
gentoo sha256 sum
#fedora
Invalid command 'php_value', pastebin
failed to start the x server debian
#centos
#web
framebuffer 2048x1536 grub
googleearth Xlib: extension XFree86-DRI missing on display
|
|