CVS dpkg md5sum fix

Wichert Akkerman

unread,

Dec 31, 2001, 9:14:08 AM12/31/01

to

Previously YAEGASHI Takeshi wrote:
> Here is another fix. Some dpkg scripts fail because newer md5sum adds
> trailing "-" to the output.

Isn't that the same fix that was commited to CVS HEAD on June 20?
The ChangeLog seems to indicate that.

Wichert.

--
_________________________________________________________________
/wic...@wiggy.net This space intentionally left occupied \
| wic...@deephackmode.org http://www.liacs.nl/~wichert/ |
| 1024D/2FA3BC2D 576E 100B 518D 2F16 36B0 2805 3CB8 9250 2FA3 BC2D |

--
To UNSUBSCRIBE, email to debian-dp...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org

YAEGASHI Takeshi

unread,

Dec 31, 2001, 9:11:39 AM12/31/01

to

Hi,

Here is another fix. Some dpkg scripts fail because newer md5sum adds
trailing "-" to the output.

Index: scripts/dpkg-genchanges.pl
===================================================================
RCS file: /cvs/dpkg/dpkg/scripts/dpkg-genchanges.pl,v
retrieving revision 1.14
diff -u -r1.14 dpkg-genchanges.pl
--- scripts/dpkg-genchanges.pl 2001/06/20 16:30:15 1.14
+++ scripts/dpkg-genchanges.pl 2001/12/31 14:02:13
@@ -321,9 +321,10 @@
(@s=stat(STDIN)) || &syserr("cannot fstat upload file $uf");
$size= $s[7]; $size || &warn("upload file $uf is empty");
$md5sum=`md5sum`; $? && subprocerr("md5sum upload file $uf");
- $md5sum =~ m/^([0-9a-f]{32})\s*$/i ||
+ chomp($md5sum);
+ $md5sum =~ s/ *-$//; # Remove trailing spaces and -, to work with GNU md5sum
+ $md5sum =~ m/^[0-9a-f]{32}$/ ||
&failure("md5sum upload file $uf gave strange output \`$md5sum'");
- $md5sum= $1;
defined($md5sum{$f}) && $md5sum{$f} ne $md5sum &&
&error("md5sum of source file $uf ($md5sum) is different from md5sum in $dsc".
" ($md5sum{$f})");
Index: scripts/dpkg-scanpackages.pl
===================================================================
RCS file: /cvs/dpkg/dpkg/scripts/dpkg-scanpackages.pl,v
retrieving revision 1.10
diff -u -r1.10 dpkg-scanpackages.pl
--- scripts/dpkg-scanpackages.pl 2001/06/20 16:30:15 1.10
+++ scripts/dpkg-scanpackages.pl 2001/12/31 14:02:14
@@ -102,6 +102,7 @@

open(C,"md5sum <$fn |") || die "$fn $!";
chop($_=<C>); close(C); $? and die "\`md5sum < $fn' exited with $?\n";
+ s/ *-$//; # Remove trailing spaces and -, to work with GNU md5sum
/^[0-9a-f]{32}$/ or die "Strange text from \`md5sum < $fn': \`$_'\n";
$tv{'MD5sum'}= $_;

Index: scripts/dpkg-scansources.pl
===================================================================
RCS file: /cvs/dpkg/dpkg/scripts/dpkg-scansources.pl,v
retrieving revision 1.5
diff -u -r1.5 dpkg-scansources.pl
--- scripts/dpkg-scansources.pl 2000/11/24 16:01:12 1.5
+++ scripts/dpkg-scansources.pl 2001/12/31 14:02:14
@@ -285,6 +285,7 @@
exec 'md5sum' or xdie "can't exec md5sum:";
}
chomp($md5 = join '', <MD5>);
+ $md5 =~ s/ *-$//; # Remove trailing spaces and -, to work with GNU md5sum
unless (close MD5) {
xwarn_noerror close_msg 'md5sum';
return;
Index: scripts/dpkg-source.pl
===================================================================
RCS file: /cvs/dpkg/dpkg/scripts/dpkg-source.pl,v
retrieving revision 1.26
diff -u -r1.26 dpkg-source.pl
--- scripts/dpkg-source.pl 2001/07/17 15:03:55 1.26
+++ scripts/dpkg-source.pl 2001/12/31 14:02:15
@@ -1071,7 +1071,9 @@
$size= (stat _)[7];
my $md5sum= `md5sum <$filename`;
$? && &subprocerr("md5sum $filename");
- $md5sum =~ s/^([0-9a-f]{32})\n$/$1/ || &failure("md5sum gave bogus output \`$_'");
+ chomp($md5sum);
+ $md5sum =~ s/ *-$//; # Remove trailing spaces and -, to work with GNU md5sum
+ $md5sum =~ m/^[0-9a-f]{32}$/ || &failure("md5sum gave bogus output \`$_'");
$f{'Files'}.= "\n $md5sum $size $filename";
}

--
YAEGASHI Takeshi <t...@keshi.org> <tak...@yaegashi.jp> <yaeg...@dodes.org>

Fumitoshi UKAI

unread,

Dec 31, 2001, 9:28:19 AM12/31/01

to

At Mon, 31 Dec 2001 15:13:41 +0100,

Wichert Akkerman wrote:
>
> Previously YAEGASHI Takeshi wrote:
> > Here is another fix. Some dpkg scripts fail because newer md5sum adds
> > trailing "-" to the output.
>
> Isn't that the same fix that was commited to CVS HEAD on June 20?
> The ChangeLog seems to indicate that.

I don't think so.

| 1.12 (doogie 28-Nov-01): printf("%s %c-\n", digest, bin_mode ? '*' : ' ');

| revision 1.12
| date: 2001/11/28 23:30:31; author: doogie; state: Exp; lines: +1 -1
| When parsing stdin, md5sum now displays '-' as the filename, to match
| textutils md5sum.

Yaegashi's patches are for this changes.

Thanks,
Fumitoshi UKAI

YAEGASHI Takeshi

unread,

Dec 31, 2001, 10:56:17 PM12/31/01

to

Hi and happy new year,

At Mon, 31 Dec 2001 23:27:17 +0900,
Fumitoshi UKAI wrote:

> At Mon, 31 Dec 2001 15:13:41 +0100,
> Wichert Akkerman wrote:
> >
> > Previously YAEGASHI Takeshi wrote:
> > > Here is another fix. Some dpkg scripts fail because newer md5sum adds
> > > trailing "-" to the output.
> >
> > Isn't that the same fix that was commited to CVS HEAD on June 20?
> > The ChangeLog seems to indicate that.

It seems that fix was partial. There are still md5sum calls that need
to be fixed, as you see.

> I don't think so.
>
> | 1.12 (doogie 28-Nov-01): printf("%s %c-\n", digest, bin_mode ? '*' : ' ');
>
> | revision 1.12
> | date: 2001/11/28 23:30:31; author: doogie; state: Exp; lines: +1 -1
> | When parsing stdin, md5sum now displays '-' as the filename, to match
> | textutils md5sum.
>
> Yaegashi's patches are for this changes.

That's right. We always need the treatment for trailing "-" even with
md5sum provided by dpkg.deb.

--
YAEGASHI Takeshi <t...@keshi.org> <tak...@yaegashi.jp> <yaeg...@dodes.org>

YAEGASHI Takeshi

unread,

Jan 1, 2002, 5:04:35 PM1/1/02

to

Hi,

At Tue, 01 Jan 2002 12:55:51 +0900,
YAEGASHI Takeshi wrote:

> > | 1.12 (doogie 28-Nov-01): printf("%s %c-\n", digest, bin_mode ? '*' : ' ');
> >
> > | revision 1.12
> > | date: 2001/11/28 23:30:31; author: doogie; state: Exp; lines: +1 -1
> > | When parsing stdin, md5sum now displays '-' as the filename, to match
> > | textutils md5sum.
> >
> > Yaegashi's patches are for this changes.
>
> That's right. We always need the treatment for trailing "-" even with
> md5sum provided by dpkg.deb.

Wichert, how do you think about this fix?

It's obvious that current CVS dpkg-dev scripts are useless for
packaging purposes. Some fixes would be needed anyway, whether my
patch was acceptable or not.

YAEGASHI Takeshi

unread,

Jan 3, 2002, 10:21:19 AM1/3/02

to

Hi,

At Thu, 3 Jan 2002 15:55:48 +0100,
Wichert Akkerman wrote:

> Previously YAEGASHI Takeshi wrote:
> > Here is another fix. Some dpkg scripts fail because newer md5sum adds
> > trailing "-" to the output.
>

> I just commited a different version of this to CVS.

Thanks. But unfortunately it seems insufficient...

% dpkg-buildpackage -rfakeroot -uc -us
dpkg-buildpackage: source package is dpkg
dpkg-buildpackage: source version is 1.10
dpkg-buildpackage: source maintainer is Wichert Akkerman <wakk...@debian.org>
dpkg-buildpackage: host architecture is i386
fakeroot debian/rules clean
test -f include/dpkg.h.in
rm -f debian/files debian/substvars
rm -f debian/dpkg.substvars debian/dpkg-static.substvars
rm -fr /home/yaegashi/cvs/cvs.debian.org/dpkg/dpkg/build /home/yaegashi/cvs/cvs.debian.org/dpkg/dpkg/build-static /home/yaegashi/cvs/cvs.debian.org/dpkg/dpkg/debian/tmp
rm -f po/{cat-id-tbl.c,stamp-cat-id,*.gmo}
rm -f stamp-build stamp-binary
dpkg-source -b dpkg
dpkg-source: warning: source directory `./dpkg' is not <sourcepackage>-<upstreamversion> `dpkg-1.10'
dpkg-source: building dpkg in dpkg_1.10.tar.gz
dpkg-source: failure: md5sum gave bogus output `'

Wichert Akkerman

unread,

Jan 3, 2002, 10:45:05 AM1/3/02

to

Previously YAEGASHI Takeshi wrote:
> Thanks. But unfortunately it seems insufficient...

Already fixed :)

Wichert.

--
_________________________________________________________________
/wic...@wiggy.net This space intentionally left occupied \
| wic...@deephackmode.org http://www.liacs.nl/~wichert/ |
| 1024D/2FA3BC2D 576E 100B 518D 2F16 36B0 2805 3CB8 9250 2FA3 BC2D |