[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: idempotency and modularity of tpm2deb-bin.pl



Norbert Preining <preining@logic.at> wrote:

> Hi Frank!
>
> On Don, 15 Mär 2007, Frank Küster wrote:
>> > - make_maintainer still uses global variables (@configfiles) which are
>> >   only set in make_deb; this could be worked around
>> 
>> I've a fixed version on my disk where @configfiles is handled
>> separately. 
>
> You disabled the clearing of @configfiles in tpm2deb-bin.pl:
> # 	# clean @configfiles
> # 	@configfiles = ();
> (I assume it was you). This is BAD, because the configfiles accumulate
> within a source package. I.e. if the first bin package has 5 config
> files, all the following bin packages have *ALSO* (at least) these 5
> config files in @configfiles. 
>
> Maybe I miss some other change you did, but I had the very same problem
> at one of the first test builds, and ended up with a lot of wrong calls
> to handle_config_files ...

You are right, that was a short-sighted change.  The aim was to be able
to call make_maintainer without make_deb, but that doesn't work,
anyway.  I'll revert that.

Hm, thinking about it, there's a reason why I did that, and that is the
fact that @configfiles is a variable with file-wide scope and no
indication that it is package specific.  I think it would be clearer if
we use a hash of array references: @{$configfiles{$package}}.

>> place again.  But again, this should give an error if the destination is
>> already present?
>
> Maybe a simple change to mycopy would suffice, as it is used to do the
> copy?

I still do not understand why it does not complain when (I think) the
destination already exists.  The symlink calls do fail, on the other
hand. 

Regards, Frank
-- 
Dr. Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)



Reply to: