Re: [Ltib] PKG_KERNEL_LEAVSRC issues...

[seh]

Hi Peter,

I'm going to need to re-read this email as at first read I'm not exactly
sure what you are running into.  If possible can you reduce it to 2
sentences of; "I expected this... but observe this...."

Just to shed some light on the design, normally .config will override
defconfig, however if you touch defconfig ltib will notice this and use
the updated defconfig.  The purpose of this is so that if you have an
autobuilder and check-in an update to defconfig, the autobuilder will use
the updated defconfig.

Regarding "LEAVESRC" the idea is that this only needs to be transient as
once a package is left unpacked, it will never be removed and so on
subsequence builds the same sources remain there and just get re-built.

Regards, Stuart


> Stuart,
>
> I've pulled LTIB from CVS on 20101216, and am using it for the basis of
> OMAP work, and I'm runing into an issue.
>
> I'm running into trouble due to the code in LTIB that clears out
> PKG_*_LEAVESRC after a build completes.
>
> The issue arises when I wanted to rebuild (from close to scratch) by
> removing everything in rpm/BUILD, rpm/RPMS/arm, and rootfs, then execute
> "./ltib" to rebuild the world .  Unfortunately
> config/platform/<platform>/.config takes precedence over the defconfig
> and requires "./ltib -b --preconfig
> config/platform/<platform>/defconfig" to make things work correctly
> (this is due to the cod ein clear_transient_config clearing
> PKG_KERNEL_LEAVESRC which gets stored in
> config/platform/<platform>/.config and used by "./ltib" ib it exists).
>
> I believe that PKG_KERNEL_LEAVESRC is overloaded as its used not only in
> a one-shot case, but also to force the requirement of kernel source for
> packages that need the kernel source.
>
> I had solved this previously by creating a new PKG_<package>_NEEDSRC
> variable is used in the ltib configuration select mechanisms(e.g):
>
> config PKG_HELLOWORLD_MOD
>      depends CAP_HAS_MMU
>      select PKG_KERNEL_NEEDSRC
>      bool "hello world module example"
>      help
>          simple hello world kernel modules example
>
> and then used either PKG_<package>_LEAVSRC or PKG_<package>_NEEDSRC to
> force ltib to skip the %clean section:
>
>          if($leavesrc || $pcf->{$key . '_LEAVESRC'} || $pcf->{$key .
> '_NEEDSRC'}) {
>              $rpmclean = '';
>              $ENV{$key . '_LEAVESRC'} = 'y';
>          } else {
>              $rpmclean = '--clean --rmsource ';
>          }
>
> and left the code in clear_transient_configs alone along with the use of
> PKG_KERNEL_LEAVESRC in config/platform/<platform>/main.lkc.  I also
> added the PKG_KERNEL_NEEDSRC rebuild key into $config_deps where LEVESRC
> was used - is there a way to have LTIB automatically detect a change in
> PKG_<package>_LEAVESRC/NEEDSRC and automatically force <package> to be
> rebuilt?  I ask partly since in the same project I had to go through
> hoops to have openssl/wpa_supplicant source remain to build the wireless
> driver for the wl127x wireless chip (for its encryption support).
>
> Either my understanding of LTIB is off(no surprise there), and/or
> PKG_<package>_LEAVSRC has subtle side effects.  This drove me nuts until
> I "fixed" it.  Ican create a patch to the current LTIB CVS and send it
> along...
>
> --
> Peter Barada
> address@hidden
>
>
> _______________________________________________
> LTIB home page: http://ltib.org
>
> Ltib mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/ltib
>