Sean Whitton writes ("Bug#863582: Allow configurable build directory"):
> On Sun, Jun 18, 2017 at 07:47:14PM -0700, Russ Allbery wrote:
> > Even with that flag, dgit writes some (but not all) of its output files in
> > the parent directory (pieces of the source package and the downloaded dsc
> > file of the previous release), and then aborts because it can't find the
> > right files.  It appears to be doing pieces of the build outside of
> > gbp-buildpackage, since I know gbp-buildpackage will always write all of
> > its files in my configured export-dir.
> I suspect that's exactly what's happening.  I think we now have two
> bugs.
> 1. --build-products-dir makes dgit /read/ build products from a new dir,
>    but it should probably also write the source package it generates to
>    that dir


Also, this codepath is far from well-tested.  I'm afraid that due to
my prevailing mental model (which doesn't contain something like gbp's
`export-dir') I have probably perpetrated many bugs where it's not

> 2. even if Ian wants to have dgit continue to put generated source
>    packages in .. no matter what, it needs to be able to read them from
>    .., even when --builds-products-dir points somewhere else.

This kind of hybrid approach would be asking for trouble.


Reply via email to