Patchwork [meta-oe] bianutils-2.20.1: fix checksums

login
register
mail settings
Submitter Eric BENARD
Date Sept. 6, 2011, 8:32 p.m.
Message ID <1315341136-25548-1-git-send-email-eric@eukrea.com>
Download mbox | patch
Permalink /patch/11119/
State Superseded
Headers show

Comments

Eric BENARD - Sept. 6, 2011, 8:32 p.m.
the source archive seems to have changed on the 26 aug 2011, thus
the checksums have to be updated.

Signed-off-by: Eric Bénard <eric@eukrea.com>
---
 .../recipes-devtools/binutils/binutils_2.20.1.bb   |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
Martin Jansa - Sept. 7, 2011, 7:58 a.m.
This is probably caused by replacing 2.20.1 tarballs with links to 2.20.1a.

IMHO would be better to change recipes to say 2.20.1a and force
everybody to redownload it with new name (otherwise breaks checksums
for everybody with old 2.20.1 in downloads dir)

Also please provide at least "diff -rq" when sending fix checksums
patches, otherwise checksums are quite useless.

The same applies to "fix checksums for binutils-2.18"..

Regards,

2011/9/6 Eric Bénard <eric@eukrea.com>:
> the source archive seems to have changed on the 26 aug 2011, thus
> the checksums have to be updated.
>
> Signed-off-by: Eric Bénard <eric@eukrea.com>
> ---
>  .../recipes-devtools/binutils/binutils_2.20.1.bb   |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb b/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
> index 2558bda..cfceab3 100644
> --- a/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
> +++ b/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
> @@ -36,7 +36,7 @@ SRC_URI = "\
>      file://202_elflink_noaddneeded_vs_weak.patch \
>      "
>
> -SRC_URI[md5sum] = "9cdfb9d6ec0578c166d3beae5e15c4e5"
> -SRC_URI[sha256sum] = "228b84722d87e88e7fdd36869e590e649ab523a0800a7d53df906498afe6f6f8"
> +SRC_URI[md5sum] = "2b9dc8f2b7dbd5ec5992c6e29de0b764"
> +SRC_URI[sha256sum] = "71d37c96451333c5c0b84b170169fdcb138bbb27397dc06281905d9717c8ed64"
>
>  BBCLASSEXTEND = "native"
> --
> 1.7.6
>
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>
Paul Menzel - Sept. 7, 2011, 8:06 a.m.
Dear Martin, dear Eric,


Eric, there is a typo in the commit summary: An »a« slipped in binutils.


Am Mittwoch, den 07.09.2011, 09:58 +0200 schrieb Martin Jansa:

> This is probably caused by replacing 2.20.1 tarballs with links to 2.20.1a.

the archives were replaced because of the “GPL violation” I think.

http://nickclifton.livejournal.com/9067.html

> IMHO would be better to change recipes to say 2.20.1a and force
> everybody to redownload it with new name (otherwise breaks checksums
> for everybody with old 2.20.1 in downloads dir)

Good point.

> Also please provide at least "diff -rq" when sending fix checksums
> patches, otherwise checksums are quite useless.

Do you mean of the archives? To see the differences in the source?

This is the first time I heard of that, but it is a reasonable
suggestion.

> The same applies to "fix checksums for binutils-2.18"..


Thanks,

Paul
Martin Jansa - Sept. 7, 2011, 8:20 a.m.
On Wed, Sep 07, 2011 at 10:06:11AM +0200, Paul Menzel wrote:
> Dear Martin, dear Eric,
> 
> 
> Eric, there is a typo in the commit summary: An »a« slipped in binutils.
> 
> 
> Am Mittwoch, den 07.09.2011, 09:58 +0200 schrieb Martin Jansa:
> 
> > This is probably caused by replacing 2.20.1 tarballs with links to 2.20.1a.
> 
> the archives were replaced because of the “GPL violation” I think.
> 
> http://nickclifton.livejournal.com/9067.html
> 
> > IMHO would be better to change recipes to say 2.20.1a and force
> > everybody to redownload it with new name (otherwise breaks checksums
> > for everybody with old 2.20.1 in downloads dir)
> 
> Good point.
> 
> > Also please provide at least "diff -rq" when sending fix checksums
> > patches, otherwise checksums are quite useless.
> 
> Do you mean of the archives? To see the differences in the source?
> 
> This is the first time I heard of that, but it is a reasonable
> suggestion.

Yes, if you're able to find "old" tarball then "diff -rq" on unpacked
new and old tarball gives good idea what was changed, if it was just
some formality or something fishy..

You can find it at least in my "fix checksum" commits, ie:
http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=79d295c31e64e972608d5dde5d57e6250d7ca77c
http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=e341fb2ca67ea7670ca15379050992e81efa5697
http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=31fa1970ca305ff72afd3cf193b85550c6f3aa88

Regards,

Patch

diff --git a/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb b/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
index 2558bda..cfceab3 100644
--- a/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
+++ b/meta-oe/recipes-devtools/binutils/binutils_2.20.1.bb
@@ -36,7 +36,7 @@  SRC_URI = "\
      file://202_elflink_noaddneeded_vs_weak.patch \
      "
 
-SRC_URI[md5sum] = "9cdfb9d6ec0578c166d3beae5e15c4e5"
-SRC_URI[sha256sum] = "228b84722d87e88e7fdd36869e590e649ab523a0800a7d53df906498afe6f6f8"
+SRC_URI[md5sum] = "2b9dc8f2b7dbd5ec5992c6e29de0b764"
+SRC_URI[sha256sum] = "71d37c96451333c5c0b84b170169fdcb138bbb27397dc06281905d9717c8ed64"
 
 BBCLASSEXTEND = "native"