Re: RFC/RFR: dh-exec -- Scripts to help with executable debhelper files

[Gergely Nagy <algernon@balabit.hu>]

Goswin von Brederlow <goswin-v-b@web.de> writes:

>> And for the record, I'm seeking input on the dh-exec implementation
>> here. Keep the anti-executable-debhelper-foo thing in the other thread,
>> please.
>
> But that is the actualy problem. The executable-debhelper-foo thing is
> in many peoples optinion bad.
>
> Your implementation and the difficulties to handle the various debhelper
> options like --sourcedir correctly show that it might not have been a
> good solution afterall. Or at least not yet quite thought out and
> finished.

It's a good solution, in my opinion, just one that could use a few
little tweaks, like passing down the options to programs ran.

> A general dh-exec that does tons of things could be nice. But if it can
> only do variable susbtituion without breaking other debhelper options
> then why not do the variable substitution in dh_install directly? All
> the promised flexibility is of no use if it can't be used safely.

Because it's easier to extend outside of dh, and can be first done on a
per-package basis to experiment with it. (Yes, it would be bad if
everyone started to do their own experiments, but one or two is fine,
until the use-case gets a generic solution)

> This doesn't mean it isn't possible to do more. Just so far no one has
> succeeded to do e.g. rename on install without breaking stuff. And that
> in itself is a big point against the feature.

To be honest, I'm fairly happy with the current implementation. Even if
it breaks --sourcedir and --that-other-option-I-never-heard-of-before,
that can be documented, and all will be well.

And if and when debhelper will start passing its options downwards, we
can fix these two, too.

-- 
|8]