[Buildroot] 2009.02-rc3 released

[Buildroot] 2009.02-rc3 released

[Peter Korsgaard]

Hi,

-rc3 is out:

http://buildroot.uclibc.org/downloads/buildroot-2009.02-rc3.tar.gz

or

http://buildroot.uclibc.org/downloads/buildroot-2009.02-rc3.tar.bz2

Please give it a spin and report back issues to the mailing list or
the bug tracker.

A nice side effect of the recent cleanups / removal of ancient stuff
is that the tarballs are 25% smaller. Unless big issues are found,
expect this to be the last release candidate before the release - So
give it a good test, so we can shake out the remaining issues by
Friday where I hope to release the final 2009.02.

I would also like to remind all devs that we're deeply frozen, and I
DON'T want to see anything else than bugfixes checked into svn.

-- 
Bye, Peter Korsgaard

[Buildroot] 2009.02-rc3 released

[Steve Calfee]

Hi Peter,

Good job. First time in years I could download a buildroot and get it to build.

So now I thought why not try and see if it runs. I built all the defaults, so I ended up with a 686 target. Fine, that is what I have on as my host. 

Googling around I found mention of running the image as a vmware bootable disk. Good I have vmware installed. However now I get stuck.

The scripts directory has a build-ext3-img script with tantalizing hints it might build what I need. It also appears it needs to be run as root? Trying sudo in ubuntu did not work. However, the script has no docs and it is talking in some kind of code :) It talks about the path to the image - what is that the output file path? Then the name of the image again - the output of the script? And then an enigmatic message "echo "Enter the path and filename for the root filesystem"
echo "tarball that you want to install into the image"" I don't see any applicable .tar type file.


First somewhere in the docs there should be a description of the disk layout for buildroot. From my buildroot dir:

binaries - a couple directories in there is an intriguingly named file "rootfs.i686.ext2" Would that be echo "Enter the path and filename for the root filesystem"
echo "tarball that you want to install into the image"
from the build-ext3-img script?


dl - documented       

package - looks like every known package that can be configured and built?

scripts - ok dir name is clear, script usage is not 

toolchain - not clear but probably build tool sources?

build_i686 - Not clear what it holds?

docs - Nice html docs      

project - my project make/config stuff?            

target - unclear, more my project specific build stuff?

toolchain_build_i686 - looks like where the toolchain got built too. Where do the binaries go? /bin only has sed?

project_build_i686 - appears to be where my project is built, but under uclibc there are 3 mysterious directories autotools-stamps  buildroot-config  busybox-1.13.2. But then there is an intriguing directory called root. Is this an NFS mountable point for my target?

Anyway, my confusion getting started cannot be totally unique. I appreciate any help I can get.

Steve


      

[Buildroot] 2009.02-rc3 released

[Steve Calfee]

After successfully building, I tried to run the built image in vmware. I found out that the kernel was not built.

I used "make menuconfig" and just exited and then "make"

What do I need to do to get the kernel built? Is there a generic x86 defconfig? If I go to the kernel (in toolchain_build_i386/linux....) and run its menuconfig, the build breaks because I need to set a path for buildroot?

Thanks, Steve


      

[Buildroot] 2009.02-rc3 released

[Claus Klein]

Hi,

I  tried to build the default busybox on mac-OX, but it failed.

A year ago, I tried to build a buildsystem based on a minmal system  
like DSL using busybox tools.
but it failed too.

I want to build a cross-development tool-chain and the embedded  
destination system at a small virtual vmware image.

So I want to ask if it is possible to use only posix conform options  
when building a system with buildroot?
At leased, the options available at the busybox should be enough, or  
not?

And to start, there is a small patch to fix my first problem I had:

claus-kleins-macbook-pro:buildroot-2009.02-rc3 clausklein$ diff -u  
package/editors/vim/vim.mk package/editors/vim/vim.mk.ck
--- package/editors/vim/vim.mk	2009-01-31 22:30:46.000000000 +0100
+++ package/editors/vim/vim.mk.ck	2009-02-07 22:39:28.000000000 +0100
@@ -9,7 +9,7 @@
  VIM_SOURCE_SITE:=$(VIM_SITE)/unix
  VIM_PATCH_SITE:=$(VIM_SITE)/patches/7.1
  VIM_DIR:=$(BUILD_DIR)/vim71
-VIM_PATCHES:=$(shell cat package/editors/vim/patches | sed -s 's:\(.\+ 
\):$(DL_DIR)/vim/$(VIM_VERSION).\1:')
+VIM_PATCHES:=$(shell cat package/editors/vim/patches | sed -e 's:\(.\+ 
\):$(DL_DIR)/vim/$(VIM_VERSION).\1:')
  VIM_CONFIG_H:=$(VIM_DIR)/src/auto/config.h
  VIM_CONFIG_MK:=$(VIM_DIR)/src/auto/config.mk

claus-kleins-macbook-pro:buildroot-2009.02-rc3 clausklein$


//BR

Claus


Am 01.02.2009 um 07:37 schrieb Peter Korsgaard:

> Hi,
>
> -rc3 is out:
>
> http://buildroot.uclibc.org/downloads/buildroot-2009.02-rc3.tar.gz
>
> or
>
> http://buildroot.uclibc.org/downloads/buildroot-2009.02-rc3.tar.bz2
>
> Please give it a spin and report back issues to the mailing list or
> the bug tracker.

[Buildroot] 2009.02-rc3 released

[Peter Korsgaard]

>>>>> "Claus" == Claus Klein <claus.klein@arcormail.de> writes:

 Claus> Hi,
 Claus> I  tried to build the default busybox on mac-OX, but it failed.

 Claus> A year ago, I tried to build a buildsystem based on a minmal system
 Claus> like DSL using busybox tools.
 Claus> but it failed too.

 Claus> I want to build a cross-development tool-chain and the embedded
 Claus> destination system at a small virtual vmware image.

 Claus> So I want to ask if it is possible to use only posix conform options
 Claus> when building a system with buildroot?
 Claus> At leased, the options available at the busybox should be enough, or
 Claus> not?

Sure, but as none of the regular BR developers build on macosx,
chances are that you'll find broken stuff. Cleanup patches are
ofcourse welcome.

 Claus> And to start, there is a small patch to fix my first problem I had:

I cleaned up the sed invocation instead in r25264, could you give it a
try please?

-- 
Bye, Peter Korsgaard