Re: Bug#43787: changed title, and remade the proposed change

[Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de>]

Hi,

On Wed, Sep 08, 1999 at 09:24:06PM -0400, Raul Miller wrote:
> On Thu, Sep 09, 1999 at 02:19:36AM +0200, Marcus Brinkmann wrote:
> > Raul, how hard do you want to make it for users to build with debugging
> > info? Activating a gcc wrapper, changing install and strip. This is
> > completely unreasonable.
> 
> I think I could build you a package which does this for you in a couple
> hours.  However it would be a hack, and it might (or might not) be a
> problem in an autobuilder environment that needed to support multiple
> compilers -- depending on exactly how that environment was architected.

Yes, of course. But as you said, it would be a hack. It would require
installation of a wrapper package although all the needed functionality is
there (just no well defined interface to activate it and support for the
interface in the package).
 
> > Indeed, I would prefer we had a way to provide a Debian system with
> > full debugging symbols included. I hope this is possible some day. But
> > this is not achievable currently, I know. Still, I think we should at
> > least try.
> 
> Ok.  If we want to have this as a goal then yeah: Ben's proposal doesn't
> meet this goal.

Yes, this is probably beyond Bens proposal, and needs to be developed
seperately. But in this case Ben should not make suggestions about the
interface in his proposal, IMO.
 
> > I hpoe my point is now more clear.
> 
> I think so:
> 
> To achieve this goal, you really want two things to happen:
> 
> (1) You want a way to guarantee that elf executables are built with
> debugging symbols.  [Depreciating the current possibility that they
> wouldn't be.]

Minor correction: Only if it is supported. I can think of some upstream
software which does not support this very well or at all. However, if it can
be supported, it should be, and if it is supported, it must be supported
through a well defined and extensible interface. Bug reports which implement
this feature should be accepted (if the patches are of acceptable quality of
course).
 
> (2) You want some way to prevent the executables from being stripped
> before they're installed on the target system.  [Depreciating the current
> unconditional stripping of debugging symbols from packages.]

Yes. Along the same lines as (1).

> Since Ben's proposal only touches on compilation -- not package building
> or installation -- you're only addressing (1) at the moment.
> 
> Do I have this correct?

Yes.
 
> [And, do you think there'd be a problem waiting on (1) until (2) is
> being addressed?]

Ah yes. Thanks for reminding me of (2). It is indeed true that it is not
adresses at all yet. Mmh. Okay, I take a step back now in this thread.
But Ben should probably take out the interface suggestion until we have
thought about it a bit longer. Raul, do you have any ideas about a clean
interface? What's better than a DEB_BUILD_OPTIONS variable in your view?

Thanks,
Marcus



-- 
`Rhubarb is no Egyptian god.' Debian http://www.debian.org  Check Key server 
Marcus Brinkmann              GNU    http://www.gnu.org    for public PGP Key 
Marcus.Brinkmann@ruhr-uni-bochum.de                        PGP Key ID 36E7CD09
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/