severity 375559 wishlist
thanks
Hi,
On 26 Jun 2006, Rob Sims stated:
> If in a clean git sourced tree,
> make-kpkg --append-to-version -test1 kernel_image
> is run, interrupted, and then re-run, subsequent builds fail:
>
> ...
> ====== making stamp-configure because of ======
> The changelog says we are creating 2.6.17-test1-gfcc18e83
> However, I thought the version is 2.6.17-test1-gfcc18e83-dirty
> exit 4
> make: *** [sanity_check] Error 4
Currently there is no support in kernel-package for anything
that changes the version number mid-build.
> In addition, make-kpkg clean does not "undirty" the kernel. The
Hmm. I guess I can look into that.
> "-dirty" in the version string may cause kernel developers to ignore
> bug reports due to using apparently non-pristine sources.
> -dirty is appended by scripts/setlocalversion, which flags the
> following:
>
> git diff-files :100644 100644
> 7c434e037e7f658bdc088cbc760d4eaba36f785d
> 0000000000000000000000000000000000000000 M scripts/package/Makefile
> :100644 000000 6edb29f2b4a6c8aa056a748168a70583897a9d3c
> 0000000000000000000000000000000000000000 D scripts/package/builddeb
>
> Are the changes to these files necessary, or just insurance that the
> built-in package generator is not used?
They are necessary.
manoj
--
Anthony's Law of Force: Don't force it; get a larger hammer.
Manoj Srivastava <[EMAIL PROTECTED]> <http://www.datasync.com/%7Esrivasta/>
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
