[Pkg-clamav-devel] redo debian/rules with less dh for clamav

[Sebastian Andrzej Siewior]

On 2014-03-21 17:27:09 [+0100], Andreas Cadhalpun wrote:
> Hi,
Hi,

> I can't find a difference between the packages build with pkgdatadir to
> those without this. Can you point me to a difference?

Hmmm. Now that I was going to copy the pieces and make you see it, I
have trouble to find it. I am afraid, I've been wrong…
The reason to do it seems to be to specify "with-dbdir" before it was
the option was added to configure. Now it is obsolete and can be removed
as you have done :)

> So all override* targets should be phony to (marginally) improve
> performance?
> I'm not sure that's worth the additional lines in debian/rules.

you can have ".PHONY: a b c" so it just one line but yes, I think we can
live with out it.
I droped it.

> Because they are only created, when PO2DEBCONF is yes, which it normally is
> not. (I actually don't know, when this will be yes.)

Is there an error if it tries to remove files which aren't there?

> >>>>I think the result is now quite equivalent to the old rules file.
> >>>
> >>>boah. That makes my todo  list a little smaller. Thanks.
> >>
> >>Is there still something todo for the new debian/rules?
> >
> >Basically you are too fast for me now :) So let me look at the latest
> >thing we have now, test is somehow and then I get back to you.
> 
> The PDFs are excluded from compression now and I added a symbols file for
> libclamav.

Nice, the symbol file is good one. Thumb up :)

> I build with the old rules file (with added faketime) and compared the
> result to the packages build with the new rules file. I didn't find any
> fundamental difference, only:
>  * DEBIAN/control reads now 'libclamav6 (>= 0.98.1)' instead of 'libclamav6
> (>= 0.98.1+dfsg)', thanks to the symbols file.
>  * A few bytes in the binaries differ and I can't figure out why.
> 
> Thus I think we can switch to the new debian/rules, assuming it still builds
> in oldstable/stable.

Great. I'm going to build test it on stable/oldstable on a few
architectures and see how it goes.

> Best regards,
> Andreas

Sebastian