Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-13, 20:13:
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
Not sure why it was removed, it was not present when I took over source
from Micahel. After reading man page for *ld* I see that this option is
used to strip of debugging symbols of the binaries. By removing these
tools will contain the debugging symbols so I'm not sure how to fix
this? The policy 10.1 says symbols should be stripped off with -s to
install also section 4.9.1 says about nostrip option does this mean by
default debhelper strips of symbols?
Yes, dh_strip strips symbols by default (if DEB_BUILD_OPTIONS=nostrip is
not set).
Using -s unconditionally is a bad idea, because then symbols are
getting stripped even with nostrip.
Of course modifying upstream files in-place in debian/rules is probably
a bad idea. (If you wanted do that, you'd have to restore the original
ones in the clean target.)
Also note that this is slightly broken: ${LIBS} is expanded by make to
an empty string, so you actually end up calling:
find . -name config.mk | xargs sed -i 's,-s ,,'
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121115101858.ga5...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 11:18 Thu 15 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-13, 20:13:
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
Not sure why it was removed, it was not present when I took over
source from Micahel. After reading man page for *ld* I see that
this option is used to strip of debugging symbols of the binaries.
By removing these tools will contain the debugging symbols so I'm
not sure how to fix this? The policy 10.1 says symbols should be
stripped off with -s to install also section 4.9.1 says about
nostrip option does this mean by default debhelper strips of
symbols?
Yes, dh_strip strips symbols by default (if
DEB_BUILD_OPTIONS=nostrip is not set).
Using -s unconditionally is a bad idea, because then symbols are
getting stripped even with nostrip.
Of course modifying upstream files in-place in debian/rules is
probably a bad idea. (If you wanted do that, you'd have to restore
the original ones in the clean target.)
So I guess its okay that is dropped but I guess needs to be recorded in
changelog? But I don't know how to phrase this removal though :-/
Also note that this is slightly broken: ${LIBS} is expanded by make
to an empty string, so you actually end up calling:
find . -name config.mk | xargs sed -i 's,-s ,,'
True :-)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-15, 17:22:
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
Not sure why it was removed, it was not present when I took over
source from Micahel. After reading man page for *ld* I see that this
option is used to strip of debugging symbols of the binaries. By
removing these tools will contain the debugging symbols so I'm not
sure how to fix this? The policy 10.1 says symbols should be stripped
off with -s to install also section 4.9.1 says about nostrip option
does this mean by default debhelper strips of symbols?
Yes, dh_strip strips symbols by default (if DEB_BUILD_OPTIONS=nostrip
is not set).
Using -s unconditionally is a bad idea, because then symbols are
getting stripped even with nostrip.
Of course modifying upstream files in-place in debian/rules is
probably a bad idea. (If you wanted do that, you'd have to restore the
original ones in the clean target.)
So I guess its okay that is dropped but I guess needs to be recorded in
changelog? But I don't know how to phrase this removal though :-/
You have to get rid of -s one way or another, because it breaks
DEB_BUILD_OPTIONS=nostrip.
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121115160808.ga6...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 17:08 Thu 15 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-15, 17:22:
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
Not sure why it was removed, it was not present when I took
over source from Micahel. After reading man page for *ld* I
see that this option is used to strip of debugging symbols of
the binaries. By removing these tools will contain the
debugging symbols so I'm not sure how to fix this? The policy
10.1 says symbols should be stripped off with -s to install
also section 4.9.1 says about nostrip option does this mean
by default debhelper strips of symbols?
Yes, dh_strip strips symbols by default (if
DEB_BUILD_OPTIONS=nostrip is not set).
Using -s unconditionally is a bad idea, because then symbols are
getting stripped even with nostrip.
Of course modifying upstream files in-place in debian/rules is
probably a bad idea. (If you wanted do that, you'd have to
restore the original ones in the clean target.)
So I guess its okay that is dropped but I guess needs to be
recorded in changelog? But I don't know how to phrase this removal
though :-/
You have to get rid of -s one way or another, because it breaks
DEB_BUILD_OPTIONS=nostrip.
Oh sorry I guess I didn't properly read it that time ;) Fixed the
02_dpkg-buildflags.patch file also updated the header. I didn't record
this change in changelog as I updated patch header. Hope that is okay.
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-13, 20:13: Changelog now reads: Changed License to Expat instead of MIT/X Consortium License as DEP-5 doesn't allow spaces in License name. But in 38-2 there were no spaces in short license names. [...] Ah this is what happened in previous mail you said copyright information is wrong for some tools and while reading LICENSE file I noted it as MIT/X Consortium License instead of MIT license which was previously written in copyright. So I modified I did document this change I assume then as per your suggestion I changed it later to Expat!. I know what happened, but it's impossible for someone else to deduce it just by reading the changelog. The short history name was like this: A) MIT B) MIT/X Consortium License C) Expat A-B change is not documented, only B-C is. Also the fact that B existed at some point of VCS history (but not in any released version) is not really interesting for end users. So I propose only to document A-C switch; see the attached patch. Typo: soruce- source. Also, a space is needed between 3.0 and (quilt). -- Jakub Wilk diff --git a/debian/changelog b/debian/changelog --- a/debian/changelog +++ b/debian/changelog @@ -50,8 +50,7 @@ + This file is added giving instruction on how to use the slock command. * debian/copyright: + Fixed the Copyright fields for lsw, dmenu, tabbed. -+ Changed License to Expat instead of MIT/X Consortium License as DEP-5 - doesn't allow spaces in License name. ++ Changed short license name from MIT to Expat. * Switched to soruce format 3.0(quilt) -- Vasudev Kamath kamathvasu...@gmail.com Tue, 13 Nov 2012 19:52:04 +0530
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 18:37 Wed 14 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-13, 20:13:
Changelog now reads: Changed License to Expat instead of MIT/X
Consortium License as DEP-5 doesn't allow spaces in License
name. But in 38-2 there were no spaces in short license names.
[...]
Ah this is what happened in previous mail you said copyright
information is wrong for some tools and while reading LICENSE file
I noted it as MIT/X Consortium License instead of MIT license
which was previously written in copyright. So I modified I did
document this change I assume then as per your suggestion I
changed it later to Expat!.
I know what happened, but it's impossible for someone else to deduce
it just by reading the changelog. The short history name was like
this:
A) MIT
B) MIT/X Consortium License
C) Expat
A-B change is not documented, only B-C is. Also the fact that B
existed at some point of VCS history (but not in any released
version) is not really interesting for end users. So I propose only
to document A-C switch; see the attached patch.
Thanks for the patch I applied it
Typo: soruce- source. Also, a space is needed between 3.0 and
(quilt).
Fixed now :-)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 18:49 Mon 12 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-11, 12:55:
Done changed to Expat License.
Changelog now reads: Changed License to Expat instead of MIT/X
Consortium License as DEP-5 doesn't allow spaces in License name.
But in 38-2 there were no spaces in short license names.
Please double-check that there are no undocumented changes to
d/copyright.
Ah this is what happened in previous mail you said copyright information
is wrong for some tools and while reading LICENSE file I noted it as
MIT/X Consortium License instead of MIT license which was previously
written in copyright. So I modified I did document this change I assume
then as per your suggestion I changed it later to Expat!.
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
Not sure why it was removed, it was not present when I took over source
from Micahel. After reading man page for *ld* I see that this option is
used to strip of debugging symbols of the binaries. By removing these
tools will contain the debugging symbols so I'm not sure how to fix
this? The policy 10.1 says symbols should be stripped off with -s to
install also section 4.9.1 says about nostrip option does this mean by
default debhelper strips of symbols?
As I understand it Added VERSION for each involved tool to contain
current version is a change to debian/create_orig_source, not to
debian/rules. (But then, this file didn't exist in 38-2, so
mentioning it would be rather confusing...)
Removed the line from changelog
Added hardening flags options is also listed as a change to
debian/rules, but did debian/rules need actually any modification to
enable hardening? I believe that with compat 9, dh took care of
this.
Yes removed the line now
Typo: manpage-hyphe-fix - manpage-hyphen-fix.
Fixed
Switching to source format 3.0 (quilt) is not documented in the
changelog.
Now added
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-11, 12:55:
Done changed to Expat License.
Changelog now reads: Changed License to Expat instead of MIT/X
Consortium License as DEP-5 doesn't allow spaces in License name. But
in 38-2 there were no spaces in short license names.
Please double-check that there are no undocumented changes to
d/copyright.
What did this line do, and why it was removed?
find . -name config.mk | xargs sed -i 's,-s ${LIBS},${LIBS},'
As I understand it Added VERSION for each involved tool to contain
current version is a change to debian/create_orig_source, not to
debian/rules. (But then, this file didn't exist in 38-2, so mentioning
it would be rather confusing...)
Added hardening flags options is also listed as a change to
debian/rules, but did debian/rules need actually any modification to
enable hardening? I believe that with compat 9, dh took care of this.
Typo: manpage-hyphe-fix - manpage-hyphen-fix.
Switching to source format 3.0 (quilt) is not documented in the
changelog.
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121112174932.ga3...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Jakub Wilk wrote (11 Nov 2012 00:38:28 GMT) : * Vasudev Kamath kamathvasu...@gmail.com, 2012-11-10, 11:28: debian/watch contains only a single line version=3? Is that intentional? As far as I can see, this change is not documented in the changelog. Yes file is introduced to suppress the lintian warning. A Lintian override would be less hackish, and would express the intent a bit more clearly, wouldn't it? Cheers! -- intrigeri | GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc | OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/85d2zk1rfr@boum.org
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 09:20 Sun 11 Nov , intrigeri wrote:
Jakub Wilk wrote (11 Nov 2012 00:38:28 GMT) :
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-10, 11:28:
debian/watch contains only a single line version=3? Is that
intentional? As far
as I can see, this change is not documented in the changelog.
Yes file is introduced to suppress the lintian warning.
A Lintian override would be less hackish, and would express the intent
a bit more clearly, wouldn't it?
Anything is fine with me till date I was planting dummy watch file in
all my packages to avoid lintian warning but I can even add lintian override
Warm Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-10, 11:28: debian/watch contains only a single line version=3? Is that intentional? As far as I can see, this change is not documented in the changelog. Yes file is introduced to suppress the lintian warning. Do you think I need to remove that file? Nah, keep it. A comment inside the watch file explaining that it's intentionally defunct would be helpful. But now that the change is included in the changelog, I can live without it. At least for lsw, dmenu and tabbed, copyright statement in d/copyright don't match those in LICENSE files. Please update the copyright file. Fixed It's getting better. :) tabbed/LICENSE has 2009-2011 Enno Boland, but it's only 2009-2010 in d/copyright. DEP-5 says License names are case-insensitive, and may not contain spaces and There are many versions of the MIT license. Please use Expat instead, when it matches. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2012003828.ga4...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 01:38 Sun 11 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-11-10, 11:28:
debian/watch contains only a single line version=3? Is that
intentional? As far as I can see, this change is not documented
in the changelog.
Yes file is introduced to suppress the lintian warning. Do you
think I need to remove that file?
Nah, keep it. A comment inside the watch file explaining that it's
intentionally defunct would be helpful. But now that the change is
included in the changelog, I can live without it.
OK :)
At least for lsw, dmenu and tabbed, copyright statement in
d/copyright don't match those in LICENSE files. Please update
the copyright file.
Fixed
It's getting better. :)
tabbed/LICENSE has 2009-2011 Enno Boland, but it's only
2009-2010 in d/copyright.
Eh I missed it writing copyright is really boring ;-)
DEP-5 says License names are case-insensitive, and may not contain
spaces and There are many versions of the MIT license. Please use
Expat instead, when it matches.
Done changed to Expat License.
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 17:55 Wed 07 Nov , Jakub Wilk wrote:
Addition of debian/README.slock.Debian doesn't seem to be documented
in the changelog.
Done now.. Sorry I missed that file also I updated the file now to
reflect changes in slock 1.1
debian/watch contains only a single line version=3? Is that
intentional? As far as I can see, this change is not documented in
the changelog.
Yes file is introduced to suppress the lintian warning. Do you think I
need to remove that file? Also now documented it in changelog
* Merge new version 39 - sorry, what does it mean? AFAIUI the
package version number is completely Debian-specific, so I don't
understand _what_ has been merged here.
Yeah I didn't think of that when writing it :-). Now rephrased it as
created version 39 with bit more info on what is introduced as new.
* Vcs-Git and Vcs-Browser now points to collab-maint repository -
this has been already fixed in 38-2. Is the switch from
http://anonscm.debian.org/ to http://git.debian.org/ intentional?
It was introduced by Michael I didn't notice it now fixed.
Typo in the changelog: ssidd - ssid.
Fixed
At least for lsw, dmenu and tabbed, copyright statement in
d/copyright don't match those in LICENSE files. Please update the
copyright file.
Fixed
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Addition of debian/README.slock.Debian doesn't seem to be documented in the changelog. debian/watch contains only a single line version=3? Is that intentional? As far as I can see, this change is not documented in the changelog. * Merge new version 39 - sorry, what does it mean? AFAIUI the package version number is completely Debian-specific, so I don't understand _what_ has been merged here. * Vcs-Git and Vcs-Browser now points to collab-maint repository - this has been already fixed in 38-2. Is the switch from http://anonscm.debian.org/ to http://git.debian.org/ intentional? Typo in the changelog: ssidd - ssid. At least for lsw, dmenu and tabbed, copyright statement in d/copyright don't match those in LICENSE files. Please update the copyright file. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121107165520.ga6...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 20:33 Sun 04 Nov , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-30, 20:32:
.hg_archival.txt is no longer in sprop tarball, so it should be
removed from the repository, too.
Done and changes back in the git.
I don't see any relevant changes in the repository…
Sorry forgot to push ;-). Now it is really in git :-D
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-30, 20:32: .hg_archival.txt is no longer in sprop tarball, so it should be removed from the repository, too. Done and changes back in the git. I don't see any relevant changes in the repository… -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121104193305.ga3...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 17:59 Mon 29 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-28, 23:06:
Anselm got sprop a place in dl.suckless.org so I reverted
copyright to its original format and modified create_orig_source
to refer dl.suckless.org.
I can't build source package anymore:
| dpkg-source: info: local changes detected, the modified files are:
| git/sprop/.hg_archival.txt
| dpkg-source: info: you can integrate the local changes with dpkg-source
--commit
.hg_archival.txt is no longer in sprop tarball, so it should be
removed from the repository, too.
Done and changes back in the git.
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-28, 23:06: Anselm got sprop a place in dl.suckless.org so I reverted copyright to its original format and modified create_orig_source to refer dl.suckless.org. I can't build source package anymore: | dpkg-source: info: local changes detected, the modified files are: | git/sprop/.hg_archival.txt | dpkg-source: info: you can integrate the local changes with dpkg-source --commit .hg_archival.txt is no longer in sprop tarball, so it should be removed from the repository, too. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121029165946.ga7...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Mon, Oct 29, 2012 at 10:29 PM, Jakub Wilk jw...@debian.org wrote:
I can't build source package anymore:
| dpkg-source: info: local changes detected, the modified files are:
| git/sprop/.hg_archival.txt
| dpkg-source: info: you can integrate the local changes with dpkg-source
--commit
.hg_archival.txt is no longer in sprop tarball, so it should be removed from
the repository, too.
Ah yes! I forgot to delete it will do it tonight
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPWm9MKyd2N8tc998HRBCoxnKtXJqZTC90b+Wra=qzv...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-28, 10:45: I'd prefer if create_get_orig source downloaded a specific revision of prop (currently ecfe2752b310) rather than tip. That would make the script more deterministic. Done. Okay, how about this patch? (see attachment) I wonder what is the purpose of ssid. The manpage says it's an extremly[0] simple setsid replacement. But setsid itself _is_ extremely simple. In fact, if you compare ssid.c with setsid.c (from util-linux), you'll see they're almost identical. [0] Typo! -- Jakub Wilk diff --git a/debian/create_orig_source b/debian/create_orig_source --- a/debian/create_orig_source +++ b/debian/create_orig_source @@ -9,6 +9,7 @@ ssid_VERSION=0.1 tabbed_VERSION=0.4 wmname_VERSION=0.1 +sprop_REVISION=ecfe2752b310 CURRENT_VERSION=39 @@ -43,8 +44,8 @@ echo Downloading wmname-$wmname_VERSION... wget -q http://dl.suckless.org/tools/wmname-$wmname_VERSION.tar.gz -O $CURDIR/suckless-tools_$CURRENT_VERSION.orig-wmname.tar.gz -echo Downloading sprop from hg repository tip -wget -q http://hg.suckless.org/sprop/archive/ecfe2752b310.tar.gz -O $CURDIR/suckless-tools_$CURRENT_VERSION.orig-sprop.tar.gz +echo Downloading sprop from hg repository, revision $sprop_REVISION +wget -q http://hg.suckless.org/sprop/archive/$sprop_REVISION.tar.gz -O $CURDIR/suckless-tools_$CURRENT_VERSION.orig-sprop.tar.gz echo Creating empty suckless-tools_$CURRENT_VERSION.orig.tar.gz mkdir -p $TMPDIR/suckless-tools_$CURRENT_VERSION
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 13:40 Sun 28 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-28, 10:45:
I'd prefer if create_get_orig source downloaded a specific
revision of prop (currently ecfe2752b310) rather than tip.
That would make the script more deterministic.
Done.
Okay, how about this patch? (see attachment)
Applied, thanks for the patch :-)
I wonder what is the purpose of ssid. The manpage says it's an
extremly[0] simple setsid replacement. But setsid itself _is_
extremely simple. In fact, if you compare ssid.c with setsid.c (from
util-linux), you'll see they're almost identical.
Yes they both are identical even I don't know why this tool even
created! I will contact Anselm.
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 19:27 Sun 28 Oct , Vasudev Kamath wrote:
On 13:40 Sun 28 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-28, 10:45:
I'd prefer if create_get_orig source downloaded a specific
revision of prop (currently ecfe2752b310) rather than tip.
That would make the script more deterministic.
Done.
Okay, how about this patch? (see attachment)
Applied, thanks for the patch :-)
I wonder what is the purpose of ssid. The manpage says it's an
extremly[0] simple setsid replacement. But setsid itself _is_
extremely simple. In fact, if you compare ssid.c with setsid.c (from
util-linux), you'll see they're almost identical.
Yes they both are identical even I don't know why this tool even
created! I will contact Anselm.
Updated slock to 1.1 version changes pushed back to git
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 13:40 Sun 28 Oct , Jakub Wilk wrote:
snipped out as this is not relevant for this reply
Anselm got sprop a place in dl.suckless.org so I reverted copyright to
its original format and modified create_orig_source to refer
dl.suckless.org.
Changes are back in the git
Best Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 16:09 Thu 25 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-25, 11:54:
Shouldn't the copyright file also mention download location for
sprop (i.e. hg.suckless.org) too?
I'd prefer if create_get_orig source downloaded a specific
revision of prop (currently ecfe2752b310) rather than tip. That
would make the script more deterministic.
Done. I've contacted upstream author to get it in same download
location but since he is out of the development for a while I need
to contact Anselm from suckless community to get it there.
Besides this he was suggesting to get individual binaries involved
in suckless-tools as separate binaries. (i.e. multiple binary
single source package) What do you suggest?
The binaries are very small, and their dependencies aren't big
either, so I don't really see a point.
Ok so I will leave them in single package but there are some scattering
like ii and sic which are out there as independent packages all these
are part of tools in suckless community.
Besides this I updated changelog to merge changelog of 38-2
version and then update 39-1 changelog to remove already closed
bug references. I added closes to 658386 in changelog as new dmenu
fixes it. And I updated the Standards-version to 3.9.4
Copyright file now adheres to Debian copyright-format 1.0 should
be removed from 39-1 entry, as it was fixed in 38-2.
Done
When I run bhlc on build log of new package I get this
http://sprunge.us/CiYR But as far as I know this should be
automatically introduced when using compat level 9 with debhelper
= 9 right? Or am I missing something?
Perhaps 02_dpkg-buildflags.patch needs to be improved? :)
Fixed :-)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 22:23 Tue 09 Oct , Jakub Wilk wrote:
Shouldn't the copyright file also mention download location for
sprop (i.e. hg.suckless.org) too?
I'd prefer if create_get_orig source downloaded a specific revision
of prop (currently ecfe2752b310) rather than tip. That would make
the script more deterministic.
Done. I've contacted upstream author to get it in same download
location but since he is out of the development for a while I need to
contact Anselm from suckless community to get it there.
Besides this he was suggesting to get individual binaries involved in
suckless-tools as separate binaries. (i.e. multiple binary single
source package) What do you suggest?
Besides this I updated changelog to merge changelog of 38-2 version
and then update 39-1 changelog to remove already closed bug
references. I added closes to 658386 in changelog as new dmenu fixes
it. And I updated the Standards-version to 3.9.4
When I run bhlc on build log of new package I get this
http://sprunge.us/CiYR But as far as I know this should be
automatically introduced when using compat level 9 with debhelper = 9
right? Or am I missing something?
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-25, 11:54: Shouldn't the copyright file also mention download location for sprop (i.e. hg.suckless.org) too? I'd prefer if create_get_orig source downloaded a specific revision of prop (currently ecfe2752b310) rather than tip. That would make the script more deterministic. Done. I've contacted upstream author to get it in same download location but since he is out of the development for a while I need to contact Anselm from suckless community to get it there. Besides this he was suggesting to get individual binaries involved in suckless-tools as separate binaries. (i.e. multiple binary single source package) What do you suggest? The binaries are very small, and their dependencies aren't big either, so I don't really see a point. Besides this I updated changelog to merge changelog of 38-2 version and then update 39-1 changelog to remove already closed bug references. I added closes to 658386 in changelog as new dmenu fixes it. And I updated the Standards-version to 3.9.4 Copyright file now adheres to Debian copyright-format 1.0 should be removed from 39-1 entry, as it was fixed in 38-2. When I run bhlc on build log of new package I get this http://sprunge.us/CiYR But as far as I know this should be automatically introduced when using compat level 9 with debhelper = 9 right? Or am I missing something? Perhaps 02_dpkg-buildflags.patch needs to be improved? :) -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121025140945.ga6...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Vasudev Kamath wrote (15 Oct 2012 17:38:41 GMT) : On 19:08 Mon 15 Oct , intrigeri wrote: I think you should read the documentation about -s ours, before concluding you can't merge it back to master. Tried that but what here happens is wheezy branch is based on master which doesn't have any 38 related history at all. So merging will overwrite my 39 work. (I tried this) It's unclear to me how git merge -s ours, if done in the correct direction, could overwrite any work in the current branch, but I may have misunderstood the problem entirely, and I don't care that much, so let's leave it at that. Cheers! -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/85ehks57z7@boum.org
Processed (with 2 errors): Re: Bug#683184: RFS: suckless-tools/39-1 [ITA]
Processing control commands:
clone -1 -2
close -2
Failed to mark -2 as done: The 'bug' parameter (-2) to
Debbugs::Control::set_done did not pass regex check
Debbugs::Control::set_done('transcript', 'GLOB(0x1afcca0)',
'requester', 'Jakub Wilk jw...@debian.org', 'request_addr',
'683184-sub...@bugs.debian.org', 'request_msgid',
'20121017120810.ga5...@jwilk.net', 'request_subject', ...) called at
/usr/local/lib/site_perl/Debbugs/Control/Service.pm line 231
eval {...} called at
/usr/local/lib/site_perl/Debbugs/Control/Service.pm line 230
Debbugs::Control::Service::control_line('line', 'close -2',
'clonebugs', 'HASH(0x1b19e50)', 'limit', 'HASH(0x1b21ff0)',
'common_control_options', 'ARRAY(0x1b220c8)', 'errors', ...) called at
/usr/lib/debbugs/process line 1035
retitle -1 RFS: suckless-tools/39-1
Bug #683184 [sponsorship-requests] RFS: suckless-tools/38-2 [ITA]
Changed Bug title to 'RFS: suckless-tools/39-1' from 'RFS: suckless-tools/38-2
[ITA]'
--
683184: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=683184
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/handler.s.b683184.135047569517264.transcr...@bugs.debian.org
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Control: clone -1 -2 Control: close -2 Control: retitle -1 RFS: suckless-tools/39-1 I uploaded suckless-tools 38-2. Release Team normally notices and unblocks uploads fixing RC bugs themselves, but if this doesn't happen in a few days, please request an unblock (reportbug release.debian.org). -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121017120810.ga5...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 23:28 Mon 15 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-15, 20:13:
Now understood the issue correctly. I was using pdebuild and I
don't know how it started considering 38-2 as the orig tarball!
Well now I used the git-buildpackage and it looks fine.
Indeed, it looks okay now.
Why are there 2 newlines between the paragraphs in debian/control?
It's not clear to me whether this is compliant with Policy §5.1
(though admittedly both dpkg and debhelper parsers are happy about
it). Could you remove the extra newline?
Done updated package in repo and in git
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 22:28 Sun 14 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-14, 16:09:
W: suckless-tools source: native-package-with-dash-version
Yeah I saw these but I think changing the version number will
cause lot of problem right?
It's not the version that is wrong. -1 was a non-native package, and
so should be -2. Quoting the tag description:
Native source packages are sometimes created by accident. In most
cases the reason is the location of the original source tarball. For
version 1.0 source packages, dpkg-source determines whether they're
non-native by looking for a file named
package_upversion.orig.tar.gz in the parent directory, where
upversion is the upstream version from the most recent
debian/changelog entry.
Now understood the issue correctly. I was using pdebuild and I don't
know how it started considering 38-2 as the orig tarball! Well now I
used the git-buildpackage and it looks fine.
Also now as intregeri said I've moved the package on separate branch
called in Wheezy. Master branch tracks only 39 version. Please review
and let me know if any more things needs to be fixed
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 14:53 Sun 14 Oct , intrigeri wrote:
Hi,
(meta: I'm Vasudev's AM ;)
Vasudev Kamath wrote (12 Oct 2012 03:31:39 GMT) :
On Fri, Oct 12, 2012 at 12:28 AM, Jakub Wilk jw...@debian.org wrote:
New repository for the squeeze branch feels wrong to me
FWIW it feels wrong to me to. That's what branches are for.
Note that branches in the same repository don't necessarily have to
share common ancestors: e.g. the pristine-tar branch, when using gbp +
pristine-tar, does not share its history with the upstream and
packaging branches.
Well it is actually wrong but I can't see other altenative but
I think I will try it on -mentors and see if others have any
good idea.
One thing I can think is just remove existing suckless-tools
repository (I any how have my local copy) then rename
suckless-tools-38 to suckless-tools and on top of this import my 39
work so 38 history and 39 both can leave together. Let me see
Rewriting already published branches' history is a no-go.
What I would suggest is:
* create a squeeze branch in the existing repository (from scratch,
no shared ancestors)
* import the needed and missing (older) version into the squeeze
branch
Done the package now resides under wheezy branch. To be frank every
problem has simple solution :) I was overly complicating it thanks for
heads up ;)
* git checkout master git merge -s ours squeeze
Well *wheezy* branch will go independent of master. If you think it
can be merged please let me know your thoughts :-)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Hi, Vasudev Kamath wrote (15 Oct 2012 03:43:55 GMT) : * git checkout master git merge -s ours squeeze Well I don't really think I can merge it back to master! I think you should read the documentation about -s ours, before concluding you can't merge it back to master. Well unfortunately that repository no more exists! I don't know what happened but I guess either domain moved or something else happened. Have you tried asking the previous maintainer to provide you with their old repository's history? Cheers, -- intrigeri | GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc | OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/85y5j7d5na@boum.org
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 19:08 Mon 15 Oct , intrigeri wrote:
Hi,
Vasudev Kamath wrote (15 Oct 2012 03:43:55 GMT) :
* git checkout master git merge -s ours squeeze
Well I don't really think I can merge it back to master!
I think you should read the documentation about -s ours,
before concluding you can't merge it back to master.
Tried that but what here happens is wheezy branch is based on master
which doesn't have any 38 related history at all. So merging will
overwrite my 39 work. (I tried this)
Well unfortunately that repository no more exists! I don't know what
happened but I guess either domain moved or something else happened.
Have you tried asking the previous maintainer to provide you with
their old repository's history?
No I didn't take over directly, it was a bit of mess again one more
guy wanted to take over but since he didn't had much time I took over
from his work.
Best Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-15, 20:13: Now understood the issue correctly. I was using pdebuild and I don't know how it started considering 38-2 as the orig tarball! Well now I used the git-buildpackage and it looks fine. Indeed, it looks okay now. Why are there 2 newlines between the paragraphs in debian/control? It's not clear to me whether this is compliant with Policy §5.1 (though admittedly both dpkg and debhelper parsers are happy about it). Could you remove the extra newline? -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121015212825.ga...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Tue, Oct 16, 2012 at 2:58 AM, Jakub Wilk jw...@debian.org wrote:
Indeed, it looks okay now.
Why are there 2 newlines between the paragraphs in debian/control? It's not
clear to me whether this is compliant with Policy §5.1 (though admittedly
both dpkg and debhelper parsers are happy about it). Could you remove the
extra newline?
Well I guess I introduced it while patching from my already finished
work! Let me fix it tonight
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/cak+nopumyhhdkyagi-qbw75xrcjws7p48h2autdnatkmboa...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-13, 22:10: + Converted to copyright-format 1.0. (Closes: #685611) Converted to... doesn't usually play well with the freeze policy. But in this case the required changes are small enough that hopefully the Release Team won't mind. Just changed the wordings a bit. I hope the same :) license headers - what's that? Fixed it :-) Another thing you changed (and which is not currently covered by the changelog) is removal of commas from Files fields. Something went wrong when building source package; lintian emits: W: suckless-tools source: native-package-with-dash-version W: suckless-tools source: diff-contains-git-control-dir .git -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121014094546.ga...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 11:45 Sun 14 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-13, 22:10:
+ Converted to copyright-format 1.0. (Closes: #685611)
Converted to... doesn't usually play well with the freeze
policy. But in this case the required changes are small
enough that hopefully the Release Team won't mind.
Just changed the wordings a bit. I hope the same :)
license headers - what's that?
Fixed it :-)
Another thing you changed (and which is not currently covered by the
changelog) is removal of commas from Files fields.
I see previous changelog had converted to copyright-format 1.0 which
will cover removal of , but then I changed the phrase so yes I will
add this line to changelog
Something went wrong when building source package; lintian emits:
W: suckless-tools source: native-package-with-dash-version
Yeah I saw these but I think changing the version number will cause
lot of problem right? And possibly will not be accepted by
release-team so I didn't touch it. And 39 version will be quilt format
so this error will go.
I think proper way to version native package is 1:38.1 1:38.2 etc right?
W: suckless-tools source: diff-contains-git-control-dir .git
Ehm how did this happen! I just ran pdebuild probably I should have
run it using -I.git option. (And after that it looks fine).
Okay updated package is uploaded to mentors.
Best Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Hi, (meta: I'm Vasudev's AM ;) Vasudev Kamath wrote (12 Oct 2012 03:31:39 GMT) : On Fri, Oct 12, 2012 at 12:28 AM, Jakub Wilk jw...@debian.org wrote: New repository for the squeeze branch feels wrong to me FWIW it feels wrong to me to. That's what branches are for. Note that branches in the same repository don't necessarily have to share common ancestors: e.g. the pristine-tar branch, when using gbp + pristine-tar, does not share its history with the upstream and packaging branches. Well it is actually wrong but I can't see other altenative but I think I will try it on -mentors and see if others have any good idea. One thing I can think is just remove existing suckless-tools repository (I any how have my local copy) then rename suckless-tools-38 to suckless-tools and on top of this import my 39 work so 38 history and 39 both can leave together. Let me see Rewriting already published branches' history is a no-go. What I would suggest is: * create a squeeze branch in the existing repository (from scratch, no shared ancestors) * import the needed and missing (older) version into the squeeze branch * git checkout master git merge -s ours squeeze Also, given previous maintainer had his own repository in his own domain, importing their history, merged with ours strategy, may be an option too. Cheers! -- intrigeri | GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc | OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/85wqytnrjd@boum.org
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-14, 16:09: W: suckless-tools source: native-package-with-dash-version Yeah I saw these but I think changing the version number will cause lot of problem right? It's not the version that is wrong. -1 was a non-native package, and so should be -2. Quoting the tag description: Native source packages are sometimes created by accident. In most cases the reason is the location of the original source tarball. For version 1.0 source packages, dpkg-source determines whether they're non-native by looking for a file named package_upversion.orig.tar.gz in the parent directory, where upversion is the upstream version from the most recent debian/changelog entry. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121014202838.ga1...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Sun, Oct 14, 2012 at 6:23 PM, intrigeri intrig...@debian.org wrote:
Hi,
(meta: I'm Vasudev's AM ;)
Vasudev Kamath wrote (12 Oct 2012 03:31:39 GMT) :
On Fri, Oct 12, 2012 at 12:28 AM, Jakub Wilk jw...@debian.org wrote:
New repository for the squeeze branch feels wrong to me
FWIW it feels wrong to me to. That's what branches are for.
Note that branches in the same repository don't necessarily have to
share common ancestors: e.g. the pristine-tar branch, when using gbp +
pristine-tar, does not share its history with the upstream and
packaging branches.
Agreed!
Well it is actually wrong but I can't see other altenative but
I think I will try it on -mentors and see if others have any
good idea.
One thing I can think is just remove existing suckless-tools
repository (I any how have my local copy) then rename
suckless-tools-38 to suckless-tools and on top of this import my 39
work so 38 history and 39 both can leave together. Let me see
Rewriting already published branches' history is a no-go.
Yes but 39 version isn't uploaded yet so do you think its published!
What I would suggest is:
* create a squeeze branch in the existing repository (from scratch,
no shared ancestors)
* import the needed and missing (older) version into the squeeze
branch
OK I will try that!
* git checkout master git merge -s ours squeeze
Well I don't really think I can merge it back to master! Reason master
has gone in a different path 38 version is a native package and 39 is
3.0 quilt format with multi orig tarballs! To give more idea on what
I'm talking please have a look at [1]
[1]
http://anonscm.debian.org/gitweb/?p=collab-maint/suckless-tools-39.git;a=summary
So what really I was doing is prepared a new repository importing
suckless-tools 38-1 version and once 38-2 is uploaded I was planning
to get the patches of 39 repo and apply it to suckless-tools-38
repository. I agree it is rewriting of history but I thought that is
better alternative I had.
Also, given previous maintainer had his own repository in his own
domain, importing their history, merged with ours strategy, may be
an option too.
Well unfortunately that repository no more exists! I don't know what
happened but I guess either domain moved or something else happened.
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/cak+nopwilr5zqszypzr9500+zxnosnfouorohl451vca7xf...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Mon, Oct 15, 2012 at 1:58 AM, Jakub Wilk jw...@debian.org wrote:
It's not the version that is wrong. -1 was a non-native package, and so
should be -2. Quoting the tag description:
Native source packages are sometimes created by accident. In most cases the
reason is the location of the original source tarball. For version 1.0
source packages, dpkg-source determines whether they're non-native by
looking for a file named package_upversion.orig.tar.gz in the parent
directory, where upversion is the upstream version from the most recent
debian/changelog entry.
Still not sure if I understood correctly looking at changelog [1] I
see upstream version is still 38. Well let me try again with current
version in Debian and 38-2 to see what is the difference
[1]
http://anonscm.debian.org/gitweb/?p=collab-maint/suckless-tools.git;a=blob;f=debian/changelog;h=b8dc752fc98de1894415f041b3c52de782dd408b;hb=HEAD
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPUp=vzdsp4zkw-7zggu1xnsoe6fahmbgu-r76ogqgz...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 20:53 Thu 11 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-11, 22:34:
suckless-tools (38-2) unstable; urgency=low
* debian/control:
+ Added myself as maintainer. (Closes: #647090)
+ Vcs-* fields now points to repository on collab-maint
* debian/copyright:
+ Converted to copyright-format 1.0. (Closes: #685611)
Converted to... doesn't usually play well with the freeze
policy. But in this case the required changes are small enough
that hopefully the Release Team won't mind.
Just changed the wordings a bit. I hope the same :)
license headers - what's that?
Fixed it :-)
I noticed another flaw of the current copyright file: the
license for st/* is wrong. Please fix this, too.
Fixed!
You fixed the license short name (well, almost; it should be:
BSD-3-clause), but the actual license text (which is more
important) is still incorrect.
Fixed it too...
Ah and now it points to suckless-tools.git once you upload this
version I will bring in the 39 from my local repository. Hope it works
fine :)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 22:10 Wed 10 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-10, 21:55:
suckless-tools (38-2) unstable; urgency=low
* debian/control:
+ Added myself as maintainer. (Closes: #647090)
+ Vcs-* fields now points to repository on collab-maint
* debian/copyright:
+ Converted to copyright-format 1.0. (Closes: #685611)
Converted to... doesn't usually play well with the freeze policy.
But in this case the required changes are small enough that
hopefully the Release Team won't mind.
Just changed the wordings a bit. I hope the same :)
I noticed another flaw of the current copyright file: the license
for st/* is wrong. Please fix this, too.
Fixed! While fixing that I noticed stterm is also putting wrong
license for st (st is actually duplicated in 2 packages but will be
removed from suckless-tools from 39)
Along with closing the bug I added myself as maintainer
That should be okay.
but I didn't change standards version.
Very well.
Additionally I gave new repository name as
suckless-tools-wheezy.git I've not created it yet on collab just
would like to know if name is okay or you can suggest me better
name ;)
I'm not a git expert, but isn't it usual to keep multiple branches
in a single repository?
I placed source here and mentors package is updated too!
http://anonscm.debian.org/gitweb/?p=collab-maint/suckless-tools-38.git;a=summary
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-11, 22:34: suckless-tools (38-2) unstable; urgency=low * debian/control: + Added myself as maintainer. (Closes: #647090) + Vcs-* fields now points to repository on collab-maint * debian/copyright: + Converted to copyright-format 1.0. (Closes: #685611) Converted to... doesn't usually play well with the freeze policy. But in this case the required changes are small enough that hopefully the Release Team won't mind. Just changed the wordings a bit. I hope the same :) license headers - what's that? I noticed another flaw of the current copyright file: the license for st/* is wrong. Please fix this, too. Fixed! You fixed the license short name (well, almost; it should be: BSD-3-clause), but the actual license text (which is more important) is still incorrect. While fixing that I noticed stterm is also putting wrong license for st sterm, on the other hand, has the licence text copied verbatim, but only the short license name wrong. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121011185336.ga...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-11, 09:17: Additionally I gave new repository name as suckless-tools-wheezy.git I've not created it yet on collab just would like to know if name is okay or you can suggest me better name ;) I'm not a git expert, but isn't it usual to keep multiple branches in a single repository? We can keep if we had the previous history in Git but in my case the suckless-tools is a fresh repository from 39-1 and previous maintainer had his own repository in his own domain. So its kinda messy situation that I can't hold this package history in existing suckless-tools.git. Or better I host it some where else than collab like github or gitorious with same name as suckless-tools.git what do you think? New repository for the squeeze branch feels wrong to me, but I'm not really qualified to give you any advice. Maybe other -mentors@ readers could chime in? -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121011185854.ga...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Fri, Oct 12, 2012 at 12:23 AM, Jakub Wilk jw...@debian.org wrote:
license headers - what's that?
Err..
I noticed another flaw of the current copyright file: the license for
st/* is wrong. Please fix this, too.
Fixed!
You fixed the license short name (well, almost; it should be:
BSD-3-clause), but the actual license text (which is more important) is
still incorrect.
Err
While fixing that I noticed stterm is also putting wrong license for st
sterm, on the other hand, has the licence text copied verbatim, but only the
short license name wrong.
yes :)
So much for doing the work when I was sleepy :(. I will fix it tonight
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPXd_AW4jPtqwhjB_DEPOvZ8HDj+BzdY_6Gx9GcHma=m...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Fri, Oct 12, 2012 at 12:28 AM, Jakub Wilk jw...@debian.org wrote:
New repository for the squeeze branch feels wrong to me, but I'm not really
qualified to give you any advice. Maybe other -mentors@ readers could chime
in?
Well it is actually wrong but I can't see other altenative but I think
I will try it on -mentors and see if others have any good idea.
One thing I can think is just remove existing suckless-tools
repository (I any how have my local copy) then rename
suckless-tools-38 to suckless-tools and on top of this import my 39
work so 38 history and 39 both can leave together. Let me see
Best Regards
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPUG6X4R9=A18=3wn-hz8iygo8h0brs_gftn9-n6xdu...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-10, 09:16: Shouldn't the copyright file also mention download location for sprop (i.e. hg.suckless.org) too? Does source: field allow multiple URL's? According to Debian copyright-format [0] this field can be free form text which means I can put with a note that sprop was got from the different URL like below Source: http://dl.suckless.org/tools except sprop which was downloaded from hg tip at http://hg.suckless.org Yes, that would be all right. Would you mind preparing also a package for Wheezy, with *minimal* fixes for #685611 (and perhaps other bugs that would fit the freeze criteria[0])? Yes I can do that I was just not sure whether to do it or not :-). Only problem will be I can track this wheezy source in my current git reposiotory but I guess the old repository still exist and only just renamed. I will prepare this and ping back to you tonight! Just a note is it okay to change Vcs-* field in this version even though its not mentioned in bug report? I'd say that updating Vcs-* falls under 4. translation updates and documentation fixes. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121010090338.ga5...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 11:03 Wed 10 Oct , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-10, 09:16:
Shouldn't the copyright file also mention download location for
sprop (i.e. hg.suckless.org) too?
Does source: field allow multiple URL's? According to Debian
copyright-format [0] this field can be free form text which means
I can put with a note that sprop was got from the different URL
like below
Source: http://dl.suckless.org/tools except sprop which was
downloaded from hg tip at http://hg.suckless.org
Yes, that would be all right.
I will prepare this later.
Would you mind preparing also a package for Wheezy, with
*minimal* fixes for #685611 (and perhaps other bugs that would
fit the freeze criteria[0])?
Yes I can do that I was just not sure whether to do it or not :-).
Only problem will be I can track this wheezy source in my current
git reposiotory but I guess the old repository still exist and
only just renamed. I will prepare this and ping back to you
tonight!
Just a note is it okay to change Vcs-* field in this version even
though its not mentioned in bug report?
I'd say that updating Vcs-* falls under 4. translation updates and
documentation fixes.
I prepared a version with few changes and is at [0]
Changelog
suckless-tools (38-2) unstable; urgency=low
* debian/control:
+ Added myself as maintainer. (Closes: #647090)
+ Vcs-* fields now points to repository on collab-maint
* debian/copyright:
+ Converted to copyright-format 1.0. (Closes: #685611)
-- Vasudev Kamath kamathvasu...@gmail.com Wed, 10 Oct 2012 20:48:05 +0530
Along with closing the bug I added myself as maintainer but I didn't
change standards version. I'm not sure if it will be considered as
major change so please let me know If I need to change it.
Additionally I gave new repository name as suckless-tools-wheezy.git
I've not created it yet on collab just would like to know if name is
okay or you can suggest me better name ;)
There is also a patch for removing FTBFS on Gnu/hurd but this is
fixed in new version of dmenu do you think it is good to add this
patch [1]
[0] https://mentors.debian.net/package/suckless-tools
[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=658386
Best Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-10-10, 21:55: suckless-tools (38-2) unstable; urgency=low * debian/control: + Added myself as maintainer. (Closes: #647090) + Vcs-* fields now points to repository on collab-maint * debian/copyright: + Converted to copyright-format 1.0. (Closes: #685611) Converted to... doesn't usually play well with the freeze policy. But in this case the required changes are small enough that hopefully the Release Team won't mind. I noticed another flaw of the current copyright file: the license for st/* is wrong. Please fix this, too. Along with closing the bug I added myself as maintainer That should be okay. but I didn't change standards version. Very well. Additionally I gave new repository name as suckless-tools-wheezy.git I've not created it yet on collab just would like to know if name is okay or you can suggest me better name ;) I'm not a git expert, but isn't it usual to keep multiple branches in a single repository? There is also a patch for removing FTBFS on Gnu/hurd but this is fixed in new version of dmenu do you think it is good to add this patch [1] Hurd is not a release architecture, so I'm pretty sure Release Team won't be happy about such a change. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121010201026.ga6...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Thu, Oct 11, 2012 at 1:40 AM, Jakub Wilk jw...@debian.org wrote:
Converted to... doesn't usually play well with the freeze policy. But in
this case the required changes are small enough that hopefully the Release
Team won't mind.
Do you think I should change the Phrase like added the Format:
Upstream-Name: and Source field instead?
I noticed another flaw of the current copyright file: the license for st/*
is wrong. Please fix this, too.
Ok I will fix this tonight!.
Along with closing the bug I added myself as maintainer
That should be okay.
but I didn't change standards version.
Very well.
Additionally I gave new repository name as suckless-tools-wheezy.git I've
not created it yet on collab just would like to know if name is okay or you
can suggest me better name ;)
I'm not a git expert, but isn't it usual to keep multiple branches in a
single repository?
We can keep if we had the previous history in Git but in my case the
suckless-tools is a fresh repository from 39-1 and previous maintainer
had his own repository in his own domain. So its kinda messy situation
that I can't hold this package history in existing suckless-tools.git.
Or better I host it some where else than collab like github or
gitorious with same name as suckless-tools.git what do you think?
There is also a patch for removing FTBFS on Gnu/hurd but this is fixed in
new version of dmenu do you think it is good to add this patch [1]
Hurd is not a release architecture, so I'm pretty sure Release Team won't be
happy about such a change.
OK I won't apply it since its any how fixed in new version of dmenu
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPUrSg66CoFLc_b0AVfrSt6JbY+htBs_gEPcFctTP1ew=g...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Shouldn't the copyright file also mention download location for sprop (i.e. hg.suckless.org) too? I'd prefer if create_get_orig source downloaded a specific revision of prop (currently ecfe2752b310) rather than tip. That would make the script more deterministic. Would you mind preparing also a package for Wheezy, with *minimal* fixes for #685611 (and perhaps other bugs that would fit the freeze criteria[0])? [0] http://release.debian.org/wheezy/freeze_policy.html -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121009202359.ga8...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Wed, Oct 10, 2012 at 1:53 AM, Jakub Wilk jw...@debian.org wrote:
Shouldn't the copyright file also mention download location for sprop (i.e.
hg.suckless.org) too?
Does source: field allow multiple URL's? According to Debian
copyright-format [0]
this field can be free form text which means I can put with a note
that sprop was
got from the different URL like below
Source: http://dl.suckless.org/tools except sprop which was downloaded
from hg tip at http://hg.suckless.org
I'd prefer if create_get_orig source downloaded a specific revision of prop
(currently ecfe2752b310) rather than tip. That would make the script more
deterministic.
Well yes I will modify it and also I will try to persuade conor lane
smith to put it in dl.suckless.org ;-)
that way it will be more clean for both above source: field and get-orig-source
Would you mind preparing also a package for Wheezy, with *minimal* fixes for
#685611 (and perhaps other bugs that would fit the freeze criteria[0])?
Yes I can do that I was just not sure whether to do it or not :-).
Only problem will be I can track this wheezy source
in my current git reposiotory but I guess the old repository still
exist and only just renamed. I will prepare this and
ping back to you tonight!
Just a note is it okay to change Vcs-* field in this version even
though its not mentioned in bug report?
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/CAK+NOPUbuUXN_ny8o-LW1ppeSH=vkpuujs7t7y5u_qzrkfj...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On Thu, Aug 30, 2012 at 10:04 PM, Jakub Wilk jw...@debian.org wrote:
Now I realized that the other tarballs were created in a wrong directory,
too. (“This target […] leaves [the tarball] in the current directory.” —
Policy §4.9)
Hmm.. So I need to leave the tarballs in the directory from where
rules is executed?
eg debian/rules get-orig-source should leave them in package root
directory and if I execute make -f rules get-orig-source it should
execute in the debian/ directory?
The word current directory actually confuses me :-)
I was not aware of this policy (I should read policy document
properly). Any way will fix this tonight
Regards
--
Vasudev Kamath
http://copyninja.info
copyninja@{frndk.de|vasudev.homelinux.net}
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.org/cak+nopxgkgkporkguwvhxhp5wy_q7e6sdyznhm9clkzr0e9...@mail.gmail.com
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-08-31, 16:40: Now I realized that the other tarballs were created in a wrong directory, too. (“This target […] leaves [the tarball] in the current directory.” — Policy §4.9) Hmm.. So I need to leave the tarballs in the directory from where rules is executed? eg debian/rules get-orig-source should leave them in package root directory and if I execute make -f rules get-orig-source it should execute in the debian/ directory? Yes. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2012083601.ga2...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 13:16 Fri 31 Aug , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-08-31, 16:40:
Now I realized that the other tarballs were created in a wrong
directory, too. (“This target […] leaves [the tarball] in the
current directory.” — Policy §4.9)
Hmm.. So I need to leave the tarballs in the directory from where
rules is executed? eg debian/rules get-orig-source should leave
them in package root directory and if I execute make -f rules
get-orig-source it should execute in the debian/ directory?
Yes.
Fixed and pushed to the repo
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-08-25, 13:02:
You create the empty suckless-tools_${version}.orig.tar.gz tarball in
the wrong directory.
Fixed
Now I realized that the other tarballs were created in a wrong
directory, too. (“This target […] leaves [the tarball] in the current
directory.” — Policy §4.9)
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120830163434.ga9...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 20:32 Fri 24 Aug , Jakub Wilk wrote:
You create the empty suckless-tools_${version}.orig.tar.gz tarball
in the wrong directory.
Ok I messed up. It was fine till I didn't introduced the mktemp to
create secure tmp directory. But when I did I forgot to check where on
earth the .orig.tar.gz was created. I will fix this
Why do you redirect tar's stderr to /dev/null?
Just to keep the output clean. Is that not recommended? If so I will
remove it
I recommend passing these options to tar, so that the user's
name/group/umask are not leaked:
--owner root --group root --mode a+rX
Ok I will fix it
Have you seen #685611?
Just now noticed it. It went to original maintainer who orphaned
thanks for pointing out. I will reply for the bug on the bug itself.
Will prepare a new source and ping you again
Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 20:32 Fri 24 Aug , Jakub Wilk wrote:
You create the empty suckless-tools_${version}.orig.tar.gz tarball
in the wrong directory.
Fixed
Why do you redirect tar's stderr to /dev/null?
Removed this too
I recommend passing these options to tar, so that the user's
name/group/umask are not leaked:
--owner root --group root --mode a+rX
Done
snip
Warm Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
You create the empty suckless-tools_${version}.orig.tar.gz tarball in
the wrong directory.
Why do you redirect tar's stderr to /dev/null?
I recommend passing these options to tar, so that the user's
name/group/umask are not leaked:
--owner root --group root --mode a+rX
Have you seen #685611?
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120824183245.ga4...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 15:49 Mon 20 Aug , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-08-18, 22:07:
get-orig-source: TMPDIR :=$(shell (mktemp --tmpdir -d suckless-tools.))
It looks like a nice hack, but... it will create a temporary
directory every time debian/rules is run (not only for the
get-orig-source target).
Since it was with get-orig-source: target I thought like all
Makefiles it should be called only when we do debian/rules
get-orig-source but looks like it creates directory even when I
call debian/rules clean which I don't understand why!
Apparently := assignments are evaluated always exactly once,
regardless of whether the variable is global or target-specific.
I see I didn't knew that. Thanks!
Do you have any other alternatives for this?
The alternative is to use shell variables for this purpose. I find
it convenient to have a separate script to create .orig.tar, and
make get-orig-source just call it.
Done :-). (well you already know from d-m discussion)
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-08-18, 22:07: get-orig-source: TMPDIR :=$(shell (mktemp --tmpdir -d suckless-tools.)) It looks like a nice hack, but... it will create a temporary directory every time debian/rules is run (not only for the get-orig-source target). Since it was with get-orig-source: target I thought like all Makefiles it should be called only when we do debian/rules get-orig-source but looks like it creates directory even when I call debian/rules clean which I don't understand why! Apparently := assignments are evaluated always exactly once, regardless of whether the variable is global or target-specific. Do you have any other alternatives for this? The alternative is to use shell variables for this purpose. I find it convenient to have a separate script to create .orig.tar, and make get-orig-source just call it. -- Jakub Wilk -- To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120820134906.ga4...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 20:46 Thu 16 Aug , Jakub Wilk wrote:
* Vasudev Kamath kamathvasu...@gmail.com, 2012-07-30, 21:56:
It doesn't look like it's suitable for wheezy, so please make it
s/unstable/experimental/.
Done! When it should be moved to unstable? After wheezy release?
Yes, after wheezy is released.
Okay
+-$ $(tabbed -d /tmp/tabbed.xid); urxvt -embed $(/tmp/tabbed.xid);
++$ $(tabbed \-d /tmp/tabbed.xid); urxvt \-embed $(/tmp/tabbed.xid);
If you're fixing this, please also fix the security hole
(insecure use of temporary files).
Done too
So it's now:
TMPFILE=$(mktemp \-\-tmpdir tabbedxid.XXX);$(tabbed \-d ${TMPFILE});urxvt
\-embed $(${TMPFILE})
which is certainly better from security perspective, but not
necessarily from usability one. I think the reason for storing xid
in a file with predictable name was that it allowed you to connect
more than one urxvt to the same tabbed. So I think that:
1) TMPFILE is a bad variable name for this purpose;
2) you could have saved one level of indirection by storing tabbed
-d output directly into the variable (rather than into temporary
file);
3) perhaps it would make more sense to use a file somewhere in
$HOME.
Well yes that line looks most ugly :-). Now I'm choosing 2nd option
given by you and patch is cleaned up.
+ @cd /tmp
+ @tar -cvf - suckless-tools_$(CURRENT_VERSION) 2 /dev/null | gzip -9
../suckless-tools_$(CURRENT_VERSION).orig.tar.gz
+ @rm -rf /tmp/suckless-tools_$(CURRENT_VERSION)
This creates temporary files insecurely.
Fixed.
I see you did this:
get-orig-source: TMPDIR :=$(shell (mktemp --tmpdir -d suckless-tools.))
It looks like a nice hack, but... it will create a temporary
directory every time debian/rules is run (not only for the
get-orig-source target).
Since it was with get-orig-source: target I thought like all Makefiles
it should be called only when we do debian/rules get-orig-source but
looks like it creates directory even when I call debian/rules clean
which I don't understand why! Do you have any other alternatives for
this? and possibly reason why its being called on targets other than
get-orig-source?
I actually used this method from git-flow package of course its not
creating any TMPDIR but it does same for setting one variable. I asked
the package owner and he told that is only way to set the variable (or
at least that is what I interpreted )
Note: I referred git-flow because its also multi source package
I've pushed the changes and waiting for your comments
With Regards
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
* Vasudev Kamath kamathvasu...@gmail.com, 2012-07-30, 21:56:
It doesn't look like it's suitable for wheezy, so please make it
s/unstable/experimental/.
Done! When it should be moved to unstable? After wheezy release?
Yes, after wheezy is released.
+-$ $(tabbed -d /tmp/tabbed.xid); urxvt -embed $(/tmp/tabbed.xid);
++$ $(tabbed \-d /tmp/tabbed.xid); urxvt \-embed $(/tmp/tabbed.xid);
If you're fixing this, please also fix the security hole (insecure use
of temporary files).
Done too
So it's now:
TMPFILE=$(mktemp \-\-tmpdir tabbedxid.XXX);$(tabbed \-d ${TMPFILE});urxvt
\-embed $(${TMPFILE})
which is certainly better from security perspective, but not necessarily
from usability one. I think the reason for storing xid in a file with
predictable name was that it allowed you to connect more than one urxvt
to the same tabbed. So I think that:
1) TMPFILE is a bad variable name for this purpose;
2) you could have saved one level of indirection by storing tabbed -d
output directly into the variable (rather than into temporary file);
3) perhaps it would make more sense to use a file somewhere in $HOME.
+ @cd /tmp
+ @tar -cvf - suckless-tools_$(CURRENT_VERSION) 2 /dev/null | gzip -9
../suckless-tools_$(CURRENT_VERSION).orig.tar.gz
+ @rm -rf /tmp/suckless-tools_$(CURRENT_VERSION)
This creates temporary files insecurely.
Fixed.
I see you did this:
get-orig-source: TMPDIR :=$(shell (mktemp --tmpdir -d suckless-tools.))
It looks like a nice hack, but... it will create a temporary directory
every time debian/rules is run (not only for the get-orig-source
target).
Instead of pushing new package to mentors I've pushed my changes to
collab-maint repository [1] Hope that is fine with you if not let me
know
[1] git.debian.org:/git/collab-maint/suckless-tools.git
That's okay.
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120816184623.ga4...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
This is only a very rudimentary review. I don't have time to review this
properly for the time being. Anybody else is welcome to do it for me. :)
* Vasudev Kamath kamathvasu...@gmail.com, 2012-07-29, 22:27:
dget -x
http://mentors.debian.net/debian/pool/main/s/suckless-tools/suckless-tools_39-1.dsc
More information about hello can be obtained from http://www.example.com.
Changes since the last upload:
suckless-tools (39-1) unstable; urgency=low
It doesn't look like it's suitable for wheezy, so please make it
s/unstable/experimental/.
[ Michael Stummvoll ]
* New Maintainer (Closes: #647090)
[0]
[ Vasudev Kamath ]
* Imported new version of slock (Closes: #667796)
This fixes a security issue, so please mention CVE number in the
changelog.
+ Added myself as maintainer and Michael Stummvoll as Uploader
I'd merge this item with [0].
+ Added dependency on dpkg-dev = 1.16.1.1
It'd nice to mention why it's needed.
+This package contains a set of tools from suckless community as
+single package. To build the package you need to create source
+tarballs of individual tool component involved. This can be done
+by running following command from suckless-tools folder
+
+ fakeroot debian/rules get-orig-source
Why fakeroot?
+Forwarded: no|not-needed|url proving that it has been forwarded
Please choose one. :)
+-$ $(tabbed -d /tmp/tabbed.xid); urxvt -embed $(/tmp/tabbed.xid);
++$ $(tabbed \-d /tmp/tabbed.xid); urxvt \-embed $(/tmp/tabbed.xid);
If you're fixing this, please also fix the security hole (insecure use
of temporary files).
+override_dh_installdocs:
+ dh_installdocs
+ for TOOL in $(TOOLS); \
+ do \
+ cp $${TOOL}/README
$(D)/usr/share/doc/suckless-tools/README.$${TOOL}; \
+ done
This for loop needs a set -e; see Policy §4.6. I see other parts of
debian/rules has the same problem.
+ @cd /tmp
+ @tar -cvf - suckless-tools_$(CURRENT_VERSION) 2 /dev/null | gzip -9
../suckless-tools_$(CURRENT_VERSION).orig.tar.gz
+ @rm -rf /tmp/suckless-tools_$(CURRENT_VERSION)
This creates temporary files insecurely.
--
Jakub Wilk
--
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20120730084811.ga6...@jwilk.net
Bug#683184: RFS: suckless-tools/39-1 [ITA]
On 10:48 Mon 30 Jul , Jakub Wilk wrote:
This is only a very rudimentary review. I don't have time to review
this properly for the time being. Anybody else is welcome to do it
for me. :)
Thanks.. Wonder how many more silly stuffs show up on actual review
;-)
* Vasudev Kamath kamathvasu...@gmail.com, 2012-07-29, 22:27:
dget -x
http://mentors.debian.net/debian/pool/main/s/suckless-tools/suckless-tools_39-1.dsc
More information about hello can be obtained from http://www.example.com.
Changes since the last upload:
suckless-tools (39-1) unstable; urgency=low
It doesn't look like it's suitable for wheezy, so please make it
s/unstable/experimental/.
Done! When it should be moved to unstable? After wheezy release? It
does contains some new version of tools so asking.
[ Michael Stummvoll ]
* New Maintainer (Closes: #647090)
[0]
[ Vasudev Kamath ]
* Imported new version of slock (Closes: #667796)
This fixes a security issue, so please mention CVE number in the
changelog.
Done
+ Added myself as maintainer and Michael Stummvoll as Uploader
I'd merge this item with [0].
Done
+ Added dependency on dpkg-dev = 1.16.1.1
It'd nice to mention why it's needed.
Did it :-)
+This package contains a set of tools from suckless community as
+single package. To build the package you need to create source
+tarballs of individual tool component involved. This can be done
+by running following command from suckless-tools folder
+
+ fakeroot debian/rules get-orig-source
Why fakeroot?
Well by habit wrote it :-).. Now fixed
+Forwarded: no|not-needed|url proving that it has been forwarded
Please choose one. :)
Ouch.. I will end up doing one or other copy paste error. Fixed
+-$ $(tabbed -d /tmp/tabbed.xid); urxvt -embed $(/tmp/tabbed.xid);
++$ $(tabbed \-d /tmp/tabbed.xid); urxvt \-embed $(/tmp/tabbed.xid);
If you're fixing this, please also fix the security hole (insecure
use of temporary files).
Done too
+override_dh_installdocs:
+dh_installdocs
+for TOOL in $(TOOLS); \
+do \
+cp $${TOOL}/README
$(D)/usr/share/doc/suckless-tools/README.$${TOOL}; \
+done
This for loop needs a set -e; see Policy §4.6. I see other parts
of debian/rules has the same problem.
As discussed in IRC the SHELL := sh -e on top of rule file should
handle this.
+@cd /tmp
+@tar -cvf - suckless-tools_$(CURRENT_VERSION) 2 /dev/null | gzip -9
../suckless-tools_$(CURRENT_VERSION).orig.tar.gz
+@rm -rf /tmp/suckless-tools_$(CURRENT_VERSION)
This creates temporary files insecurely.
Fixed.
Instead of pushing new package to mentors I've pushed my changes to
collab-maint repository [1] Hope that is fine with you if not let me know
[1] git.debian.org:/git/collab-maint/suckless-tools.git
Thanks for the review.
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
Bug#683184: RFS: suckless-tools/39-1 [ITA]
Package: sponsorship-requests
Severity: normal
Dear mentors,
I am looking for a sponsor for my package suckless-tools
* Package name: suckless-tools
Version : 39-1
Upstream Author : Suckless community (multiple authors)
* URL : http://suckless.org
* License : MIT
Section : x11
It builds those binary packages:
suckless-tools - simple commands for minimalistic window managers
To access further information about this package, please visit the following
URL:
http://mentors.debian.net/package/suckless-tools
Alternatively, one can download the package with dget using this command:
dget -x
http://mentors.debian.net/debian/pool/main/s/suckless-tools/suckless-tools_39-1.dsc
More information about hello can be obtained from http://www.example.com.
Changes since the last upload:
suckless-tools (39-1) unstable; urgency=low
[ Michael Stummvoll ]
* New Maintainer (Closes: #647090)
* Updated lsw (Closes: #650573)
* Updated dmenu (Closes: #650574)
* Added sprop and lsx (Closes: #627662)
* Added manpage for dmenu_run (Closes: #610478)
* Updated manpages for slock, ssidd, swarp and wmname (Closes: #636030)
* Updated manpage for sselp (Closes: #378175)
* Vcs-Git and Vcs-Browser now points to collab-maint repository
(Closes: #636012)
* Removed st from the package. Suggests stterm now. Since st is
no longer part of this package it doesn't install st.256color.
(Closes: #642307,#665884)
* Fixed a typo in the wmname manpage
* Added some docs
[ Vasudev Kamath ]
* Imported new version of slock (Closes: #667796)
* Imported new version of tabbed
* debian/control:
+ Increased minimum debhelper required to 9
+ Bumped Standards-Version to 3.9.3. This did not require any change
to package.
+ Added myself as maintainer and Michael Stummvoll as Uploader
+ Added dependency on dpkg-dev = 1.16.1.1
* Set debian/compat to 9
* Merge new version 39
* debian/rules:
+ Added hardening flags options
+ Added get-orig-source target to get upstream source tarball for included
package
+ Added VERSION for each involved tool to contain current version
* debian/patches:
+ Added patch to do setgid shadow instead of setuid root on slock Makefile
(01_fix_setuid_slock.patch).
+ Added patch to introduce hardening flags and allow DEB_BUILD_OPTIONS=noopt
(02_dpkg-buildflags.patch).
+ Added patch to make command execution visible in Makefiles
(03_transparent-makefiles.patch).
+ Added patch to escape '-' symbol in manpage for tabbed
(04_tabbed-manpage-hyphe-fix.patch).
* debian/copyright:
+ Copyright file now adheres to Debian copyright-format 1.0
-- Vasudev Kamath kamathvasu...@gmail.com Sun, 29 Jul 2012 22:18:03 +0530
-- System Information:
Debian Release: wheezy/sid
APT prefers testing
APT policy: (500, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 3.2.0-3-amd64 (SMP w/2 CPU cores)
Locale: LANG=en_IN, LC_CTYPE=en_IN (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Regards,
--
Vasudev Kamath
http://copyninja.info
Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net}
IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net}
GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E
signature.asc
Description: Digital signature
