As a buildd admin, I want to congratulate the original policy on all the wasted
cpu cycles it has cost my system by forcing packages to compile with -g even
though those same binaries will be stripped later of this costly debugging
information.
Now, what I want to propose, is not a change so much
Raul == Raul Miller [EMAIL PROTECTED] writes:
Raul I guess this means that you didn't read
Raul
http://www.debian.org/Lists-Archives/debian-devel-announce-9908/msg5.html?
I read it, and I've waited over 20 days, but we're no closer to a
solution. Folling the new Policy would at least make
Joey == Joey Hess [EMAIL PROTECTED] writes:
Joey A hint: nobody ever claimed building packages that used
Joey /usr/share/doc would be a problem. In fact, as long ago as 2
Joey years, people were confident debhelper would handle that part of
Joey the transition very easily. That's not the
Johnie Ingram wrote:
The problem is (1) that Policy has made a big change without any kind
of transition plan, obsoleting all of potato and making it
incompatible with slink in a noticable way. (Bad for partial
upgrades.)
The problem is (2) that the issue is no longer under democratic
On Mon, Aug 30, 1999 at 06:25:52PM -0400, Ben Collins wrote:
The package can by default build without -g if it also provides a mechanism
to easily be rebuilt with debugging information. This can be done by providing
a build-debug make target, or allowing the user to specify BUILD_DEBUG=yes
in
Edward Betts [EMAIL PROTECTED] writes:
Gordon Matzigkeit [EMAIL PROTECTED] wrote:
Do you have any tangible examples of an architecture-specific example
file? Maybe I haven't been following this thread closely enough,
because I've only seen discussion of ``what-if'' scenarios.
$ ls -l
Shouldn´t packages use sensible-editor?
May the Source be with you.
Goswin
The package can by default build without -g if it also provides a mechanism
to easily be rebuilt with debugging information. This can be done by providing
a build-debug make target, or allowing the user to specify BUILD_DEBUG=yes
in the environment while compiling that package.
On Tue, Aug 31, 1999 at 11:11:47AM +0200, [EMAIL PROTECTED] wrote:
Shouldn't packages use sensible-editor?
Having VISUAL and EDITOR makes sense on most Unices; having
sensible-editor is Debian specific. Patches that can be usefully sent
upstream seem much better than patches that are likely to
On Mon, 30 Aug 1999, Ian Jackson wrote:
Package: debian-policy, emacs19, glibc-doc
Version: not known, 19.34-21, 2.1.1-12
I have:
Desired=Unknown/Install/Remove/Purge
| Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed
|/
On Tue, Aug 31, 1999 at 11:51:46AM +0200, Roman Hodek wrote:
And since the build targets of contain a lot
of commands, a second build-debug target often will mean to double
most of these commands.
No. Just set up the regular build target so that it honours the setting
of BUILD_DEBUG and add
build-debug: BUILD_DEBUG=y
Is that a GNU make feature that you can set vars at the place where a
dependency is expected?
Roman
On Tue, Aug 31, 1999 at 02:55:18PM +0300, Antti-Juhani Kaijanaho wrote:
On Tue, Aug 31, 1999 at 11:51:46AM +0200, Roman Hodek wrote:
And since the build targets of contain a lot
of commands, a second build-debug target often will mean to double
most of these commands.
No. Just set up
Previously Joseph Carter wrote:
sendmail does not support maildir, but it probably could if someone wrote
the rule to do it, or it's possible that a code patch could be (or has
been) written.
Sendmail rules have nothing to do with this, you need a delivery agent
that can do it. Procmail can,
Previously Raul Miller wrote:
First off, I'm not sure it's a good idea for policy to be a rapidly
changing entity.
It's not a good idea at all, but as Manoj pointed out it's now changing
rapidly.
Debian produces packages -- policy is a means to that end.
No, policy is a means of doing
Sorry to repeat the question,
but while my call for new static ids was shot down, nobody tried to answer
the question:
What should the package do, when it uses dynamic system ids (recognized by
name, not numeric value) and encounters a system, where the designed name
is already used by someone
On Tue, Aug 31, 1999 at 07:27:35AM -0400, Ben Collins wrote:
I think sticking with an env will make it much easier for some one to just
Of course. I just wanted to point out that it is possible to avoid code
duplication even in a Makefile :-)
--
%%% Antti-Juhani Kaijanaho % [EMAIL PROTECTED]
On Tue, 31 Aug 1999, Johnie Ingram wrote:
Raul == Raul Miller [EMAIL PROTECTED] writes:
Raul I guess this means that you didn't read
Raul
http://www.debian.org/Lists-Archives/debian-devel-announce-9908/msg5.html?
I read it, and I've waited over 20 days, but we're no closer to a
On Tue, Aug 31, 1999 at 03:03:50PM +0200, Jozef Hitzinger wrote:
I think we will need list of reserved names, just as now there's list
of reserved numbers _and_ names (0-99).
I agree.
FWIW, I'd also like to see a list of what the rc3.d start and stop
priorities actually mean, as per the
On Tue, Aug 31, 1999 at 02:36:37PM +0200, Roman Hodek wrote:
build-debug: BUILD_DEBUG=y
Is that a GNU make feature that you can set vars at the place where a
dependency is expected?
At least it works with GNU make, and it's documented in the node
Target-specific Variable Values of the
Hi,
Jozef == Jozef Hitzinger [EMAIL PROTECTED] writes:
Jozef I think we will need list of reserved names, just as now
Jozef there's list of reserved numbers _and_ names (0-99).
Since a list already exists, why not add the reserved names to
it, with no pre assigned IDs? Then we still
Hi,
Ben == Ben Collins [EMAIL PROTECTED] writes:
No. Just set up the regular build target so that it honours the setting
of BUILD_DEBUG and add this to debian/rules:
build-debug: BUILD_DEBUG=y
build-debug: build
You can use other make variables of course.
Ben I think sticking
Hi,
Ben == Ben Collins [EMAIL PROTECTED] writes:
Ben As a buildd admin, I want to congratulate the original policy on
Ben all the wasted cpu cycles it has cost my system by forcing
Ben packages to compile with -g even though those same binaries will
Ben be stripped later of this costly
Hi,
Roman == Roman Hodek [EMAIL PROTECTED] writes:
build-debug: BUILD_DEBUG=y
Roman Is that a GNU make feature that you can set vars at the place where a
Roman dependency is expected?
Yes.
File: make.info, Node: Target-specific, Next: Pattern-specific, Prev: Enviro\
nment, Up:
Last announcement I saw, we were *not* supposed to move things to
/usr/share until the transition was properly managed on the dpkg side;
I'm not going to move the emacs info files until I hear otherwise, at
least.
However, in this case, all you'd really want is for emacs to search a
path, so it
On Tue, Aug 31, 1999 at 10:17:28AM -0400, Dale Scheetz wrote:
There is currently a vote underway in the technical committee. Raul and
myself have voted, and are waiting for the others on the committee to
vote.
As has Manoj.
FYI,
--
Raul
-BEGIN PGP SIGNED MESSAGE-
I notice that mysql-server has the same situation; it creates or takes over
the 'mysql' group and user, in the mysql-server.preinst file.
(If I happened to have a user with that name before installing mysql-server,
I wouldn't be very happy.) I suspect that
Ben Collins [EMAIL PROTECTED] writes:
I think sticking with an env will make it much easier for some one to just
use dpkg-buildpackage (without modification) and call it like:
BUILD_DEBUG=y dpkg-buildpackage -B
Just a minor nit. That should be:
BUILD_DEBUG=y dpkg-buildpackage -B
Other
On Tue, Aug 31, 1999 at 03:44:07PM +0200, Wichert Akkerman wrote:
Previously Raul Miller wrote:
First off, I'm not sure it's a good idea for policy to be a rapidly
changing entity.
It's not a good idea at all, but as Manoj pointed out it's now changing
rapidly.
Debian produces
On Tue, Aug 31, 1999 at 06:09:26PM +0100, Philip Hands wrote:
Ben Collins [EMAIL PROTECTED] writes:
I think sticking with an env will make it much easier for some one to just
use dpkg-buildpackage (without modification) and call it like:
BUILD_DEBUG=y dpkg-buildpackage -B
Just a
In article [EMAIL PROTECTED],
Jason Gunthorpe [EMAIL PROTECTED] wrote:
On Thu, 26 Aug 1999, Roland Rosenfeld wrote:
The solution for this problem is to use fcntl(), because Linux 2.2.*
flushes the cache of a file in the moment when it is locked using
fcntl().
But only fcntl() locking is
31 matches
Mail list logo
