Re: [Angstrom-devel] (midnight?) issues with angstrom-version: lsb_version does not end up in ${S}
Hello, On Thu, Sep 9, 2010 at 2:21 PM, Steve Sakoman sako...@gmail.com wrote: On Fri, Aug 27, 2010 at 4:57 PM, Leon Woestenberg Hello, angstrom-version fails to fetch and unpack the lsb_version file: [...] At first I thought I had a race condition about the date incrementing, as my build was just crossing the midnight border: Any progress on finding the source of this error? I (and quite a few others) run into this issue regularly. I've looked I am not quite sure if the over midnight build is related or not. To further investigate this: Maybe a few people should start a CLEAN build early in the morning, i.e. a build that completely finishes BEFORE the date changes. Report in this thread please, Thanks, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] (midnight?) issues with angstrom-version: lsb_version does not end up in ${S}
Hello, On Sun, Sep 12, 2010 at 7:40 PM, Leon Woestenberg leon.woestenb...@gmail.com wrote: To further investigate this: Maybe a few people should start a CLEAN build early in the morning, i.e. a build that completely finishes BEFORE the date changes. Make sure you are running a fully parallel build (i.e. do not use the proposed workaround by disabling parallelism). i.e. Use parallel make and parallel bitbake. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] i.mx51
Hello, On Thu, Aug 12, 2010 at 7:10 PM, Yuri Bushmelev jay4m...@gmail.com wrote: 2010/8/12 Maxim Podbereznyy lisar...@gmail.com: Hi folks! I've just got IMX51EVK and want to compile full set of utilities like u-boot etc using OE environment. Does OE support imx51? AFAIK, Leon Woestenberg (likewise) have support of imx51 for OE in his TODO :) Not sure about current progress. I'm waiting for this too with Efika MX. OpenEmbedded developer Khem Raj beat me with providing a kernel for Efika MX. Not sure if this is well tested, as Khem mentioned this kind of work happens in the 25th hour of the day. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] (midnight?) issues with angstrom-version: lsb_version does not end up in ${S}
Hello, angstrom-version fails to fetch and unpack the lsb_version file: when I happened to build angstrom-version (just before midnight), the lsb_version could not be found. NOTE: Running task 2185 of 3066 (ID: 454, /home/leon/sandbox/ion/openembedded/openembedded.git/recipes/angstrom/angstrom-version.bb, do_install) ERROR: function do_install failed ERROR: log data follows (/home/leon/sandbox/ion/openembedded/tmp/work/ion-angstrom-linux/angstrom-version-1_2010.7-test-20100827-r5/temp/log.do_install.3484) | install: cannot stat `/home/leon/sandbox/ion/openembedded/tmp/work/ion-angstrom-linux/angstrom-version-1_2010.7-test-20100827-r5/lsb_release': No such file NOTE: Task failed: /home/leon/sandbox/ion/openembedded/tmp/work/ion-angstrom-linux/angstrom-version-1_2010.7-test-20100827-r5/temp/log.do_install.3484 At first I thought I had a race condition about the date incrementing, as my build was just crossing the midnight border: tmp/work/ion-angstrom-linux: drwxr-xr-x 3 leon leon 4096 2010-08-26 23:59 angstrom-version-1_2010.7-test-20100826-r5 However, cleaning and retrying gave the same results, Will investigate further tomorrow, just wanted this reported. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] lzma compressed initramfs
Hello Andrea, On Sun, Dec 13, 2009 at 11:32 AM, Andrea Adami andrea.ad...@gmail.com wrote: Second (minor) problem is some sed in linux-kexecboot.inc. Opinions from core-devs ? (I'm not allowed to touch classes...:=) Yes you are. You are not allowed to *commit* them without proper review etc. :-) If you find an elegant solution (or OE accepted hack, for that matter) to your problem without introducing regressions elsewhere your solution will probably be brought aboard. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] [RFC] [PATCH do not apply] Use ?= on PREFERRED_VERSION_linux-libc-headers
Hello, would this patch be accepted in angstrom? I am adding a new architecture in OpenEmbedded which has no upstream kernel support and must override this version. This is the only disruptive change for Angstrom, the second and last patch for Angstrom is exemption from GNU hash style. conf/distro/angstrom-2008.1.conf #This is unrelated to the kernel version, but userspace apps (e.g. HAL) require a recent version to build against -PREFERRED_VERSION_linux-libc-headers = 2.6.31 +PREFERRED_VERSION_linux-libc-headers ?= 2.6.31 Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] Minicom (Serial Port Comm) equivalent for Angstrom...
Hello, On Mon, Nov 9, 2009 at 7:24 AM, Pankaj Chawla pankaj...@gmail.com wrote: Hi Is there a serial communication program on the lines of Minicom that works on Angstrom. I need to do some testing on the serial ports connected to Besides minicom, I often use picocom. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] Minicom (Serial Port Comm) equivalent for Angstrom...
Hello, On Sat, Nov 21, 2009 at 4:33 PM, Pankaj Chawla pankaj...@gmail.com wrote: Is minicom available on the device or connected over serial port to your host machine. I need on device minicom equivalent. I think all those programs can be built for, and run on the target device: sandbox/openrd-base/openembedded/openembedded.git/recipes/microcom: microcom microcom_1.02.bb sandbox/openrd-base/openembedded/openembedded.git/recipes/minicom: minicom-2.1 minicom_2.1.bb minicom-2.3 minicom_2.3.bb sandbox/openrd-base/openembedded/openembedded.git/recipes/picocom: picocom picocom_1.3.bb picocom_1.4.bb I suggest you try to build or install one using an ipkg. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] Website outage
Hello Philip, On Wed, Sep 9, 2009 at 1:29 PM, Philip Balisterphi...@balister.org wrote: It looks like the Angstrom website is back. Sorry for the outage, we had a server move with short notice, and I have been on vacation. Let us know if there is anything else not working. Thanks for the work! Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] [PATCH] Support for powerpc-gnuspe for angstrom-eglibc
The calamari machine uses a PowerPC E500 core, which has the (GNU)SPE
ABI for it's signal processing extensions.
This is the default ABI for Linux, so let's enable it.
This should not affect any other machines and/or architectures.
Tested-by: Leon Woestenberg l...@sidebranch.com
Signed-off-by: Leon Woestenberg l...@sidebranch.com
diff --git a/conf/distro/include/angstrom-eglibc.inc
b/conf/distro/include/angstrom-eglibc.inc
index 92e2233..e947833 100644
--- a/conf/distro/include/angstrom-eglibc.inc
+++ b/conf/distro/include/angstrom-eglibc.inc
@@ -3,6 +3,8 @@ require conf/distro/include/eglibc.inc
TARGET_OS = linu...@['','-gnueabi'][bb.data.getVar('TARGET_ARCH',d,1)
in ['arm', 'armeb'] and bb.data.getVar('MACHINE',d) not in
['collie','h3600', 'h3800', 'simpad', 'ht
+TARGET_OS = linu...@['','-gnuspe'][bb.data.getVar('TARGET_ARCH',d,1)
in ['ppc', 'powerpc'] and bb.data.getVar('MACHINE',d) in
['calamari']]}
+
#TARGET_OS = linux-gnuspe
#mess with compiler flags to use -Os instead of -O2
--
Leon
___
Angstrom-distro-devel mailing list
Angstrom-distro-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] Angstrom defaults to gcc-4.3.3 for powerpc - so some early test results
Hello, for some reason my powerpc cross-compiler defaults to gcc-4.3.3, probably after this commit: --- commit 448b9103e494765fe589283d399c90e9b6e9cd84 Author: Koen Kooi k...@openembedded.org Date: Wed May 13 13:36:43 2009 +0200 angstrom 2009.X: toolchain updates: * make gcc 4.3.3 the default compiler (except for ppc and avr32) * make binutils 2.18 the default (except for armv7a and avr32) --- So without my knowing I started testing gcc-4.3.3 for powerpc, which is good. :-) Here are my findings so far, in case I drift off to other duties or holidays: However, gcc-4.3.x is not able to compile kernel 2.6.24 and earlier, at least not for all non-x86 architectures. kernel/built-in.o: In function `getnstimeofday': (.text+0x253cc): undefined reference to `__umoddi3' kernel/built-in.o: In function `getnstimeofday': (.text+0x253ec): undefined reference to `__udivdi3' kernel/built-in.o: In function `do_gettimeofday': (.text+0x254fe): undefined reference to `__udivdi3' kernel/built-in.o: In function `do_gettimeofday': (.text+0x2551c): undefined reference to `__umoddi3' For powerpc, I came across this fix, not sure if it affects only powerpc: Add -fno-tree-scev-cprop to the KBUILD_CFLAGS This fix is tricky, because earlier versions of the kernel do not employ KBUILD_CFLAGS at all in it's makefile, and any attempt to modify other similar build variables with the flags -fno-tree-scev-cprop have failed. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [oe] ANGSTROM_MODE renamed to ANGSTROMLIBC
Hi Koen, On Mon, Dec 8, 2008 at 8:27 PM, Leon Woestenberg [EMAIL PROTECTED] wrote: On Mon, Dec 8, 2008 at 11:39 AM, Koen Kooi [EMAIL PROTECTED] wrote: Due to changes in how bitbake handles environment vars ANGSTROM_MODE has been renamed to ANGSTROMLIBC: What is the causing change in bitbake? Sorry, disregard my question, I was reading my email in reverse time order... This has been discussed on the ml already. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] Treo 600 available for OE
Hello, Some friendly person is submitting a Treo 600 to an OpenEmbedded developer who intends to make run any OpenEmbedded distribution on it. Any takers for the sucker? If noone responds, I will volunteer myself. Others go first, as my spare time is limited. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [oe] [RFC] Changing stable branch commit policy
Hello all, On Sat, Mar 29, 2008 at 11:48 AM, Koen Kooi [EMAIL PROTECTED] wrote: The mailinglist: And viewmtn: The commit policy (currently under RFC at the above mailinstlist): Koen, thanks for all the effort you put into on this. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] Changing stable branch commit policy
Hello all, Do you want to write a short blurb on this we can stick on the oe and angstrom websites? regards, Koen Sure, a first shot at that (and sorry for cross-posting, but this really is OE related :-) --- The OpenEmbedded project has started to maintain a stable branch, next to its development branch, on March 28th 2007. Besides org.openembedded.dev there now is the org.openembedded.stable branch. The plumbing work on the development branch (long-term improvements such as packaged staging, creating stand-alone SDK toolchains and a new package manager) colides with the goal of offering a stable base for OpenEmbedded-derived distributions. Some projects, most notably the Angstrom distribution, already maintained a stable branch for this reason. The effort of maintaining a stable branch is now centralized back into the OpenEmbedded project. The main goal of the stable branch is to provide a branch that progresses in time with a minimum of disruptive changes, especially to the framework, which include classes, configuration namespace, package management, and tooling such as the bitbake tool. Changes to the stable branch are selected and reviewed according to an agreed upon policy before being committed to the stable branch. This will require extra effort from the OpenEmbedded developer community. The stable branch is not intended to keep up with the pace of development on the development branch. Planned is to create new stable branches every year; the next stable branch is scheduled to branch of the development tree during 2008.12. Old stable branches will go in unmaintained mode 3 months after a new stable branch is created; the current stable will retire 2009.2. OpenEmbedded projects and distributions now have the choice of at least a .dev and .stable branch and the OpenEmbedded developers can keep on plumbing. --- Leon. ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] Changing stable branch commit policy
Hello all, On Tue, Mar 25, 2008 at 12:23 AM, Philip Balister [EMAIL PROTECTED] wrote: Koen Kooi wrote: If we drop the bookkeeping part, getting things into the stable tree will be much easier: ... objections, additions? +1 At some point in time, a stable and development branch will diverge. We should focus Angstrom stable on being usable and not tracking .dev. If not done yet, we have to decide to go for one of two approaches. (1) .stable tracks .dev, but lags. Takes lots of efforts. d s | /| |/ | | /| |/ | | /| |/ | (2) .stable diverges from .dev, not keeping up at some point. The fish bone branch-off approach. d s2 | / |/ | s1 | / |/ Yes, so instead of a .stable branch that lags behind .dev, we have a fish bone approach, were we split from .dev for each stable branch. I would prefer (2), and I think Angstrom decided for this as well given the 2007-1 naming?? Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] [RFC] Changing stable branch commit policy
Hello Koen, On Thu, Mar 27, 2008 at 9:23 PM, Koen Kooi [EMAIL PROTECTED] wrote: | If not done yet, we have to decide to go for one of two approaches. | (1) .stable tracks .dev, but lags. Takes lots of efforts. | (2) .stable diverges from .dev, not keeping up at some point. The | fish bone branch-off approach. | I would prefer (2), and I think Angstrom decided for this as well | given the 2007-1 naming?? At the moment we try to do (1), but the long term plan is to do (2). The ~ main issue with (2) is that it requires either picking a alomost random (broken) branchpoint, or enforce a (brief) period of stabilizing .dev before branching of. And some planning ahead is needed: - Announce when we branch off a new stable branch 1 month ahead (roughly). - Announce when a stable branch goes in unmaintained mode 3 months ahead (roughly). - Select a branch interval, proposal is once per year or half year. Say, 2008-7 or 2009.1, but at least afer meta/sdk/opkg/packaged-staging has stabilized enough in .dev? I am very much in favor of keeping the tooling/framework/classes largely feature-frozen within a stable branch, and selectively upgrade packages. Should we create an oe-stable-branch mailinglist were patches get presented, reviewed and signed-off on? As .stable derives from Angstrom, I suggest merging all the effort/people/processes from Angstrom over to OpenEmbedded-wide stable should be the ideal goal. Is that achievable? Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] Unmutable PREFERRED_VERSION_xyz in angstrom-2007.1.conf
Hello Koen, On 10/21/07, Koen Kooi [EMAIL PROTECTED] wrote: Is it possible to change this upstream, or is there a reason to make some package versions fully unmutable through local.conf? Some of the '=' ones were meant to make angstrom 'unbreakable', but several developer insist on giving enough rope to hang themselves, so we can change them to '?=' and start preparing responses like If you put nonsense in your local.conf you void your warranty or Why are you surprised it breaks with that local.conf? Maybe it's better to keep as-is, for the reason you mentioned. I was just wondering if it was intentional, or not. Developers can cope. Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
[Angstrom-devel] Unmutable PREFERRED_VERSION_xyz in angstrom-2007.1.conf
Hello all, while testing newer package I noticed I could override some, and some not: conf/distro/angstrom-2007.1.conf uses one of PREFERRED_VERSION ?= PREFERRED_VERSION = for package versions. I would like to be able to override these easily from my local.conf, and so ?= is preferred for me. Is it possible to change this upstream, or is there a reason to make some package versions fully unmutable through local.conf? Regards, -- Leon ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
