Hi Ian,

On Jan 04 2017, Ian Jackson <ijack...@chiark.greenend.org.uk> wrote:
> Control: clone -1 -2
> Control: retitle -2 split brain o+d/p tree should copy HEAD debian for clarity
> Control: severity -2 minor
>
> Nikolaus Rath writes ("Bug#850005: dgit push without dgit build-source"):
>> On Jan 03 2017, Ian Jackson <ijack...@chiark.greenend.org.uk> wrote:
>> > I think something like
>> >   git diff --stat 9e2aab849fc3a861ab5a..HEAD
>> > may be informative.
>> 
>> Uh, that does not look good. It contains ... a ton of files and changes
>> that I won't even bother to paste here.
>
> OK, I've looked at this now.  Most of the output there is in debian/
> and is a red herring.  dgit only cares about upstream files.  (I have
> cloned this bug to request an improvement here.)

Ah, ok.

> The real problem is that you have files in your orig tarball which are
> not in your git branch, but also not deleted by patches.
>
> I guess git-dpm doesn't like to match patches which delete files ?

I don't know what you mean with "matching patches", but generally
git-dpm should handle removal of files in debian patches just fine:

$ git-dpm checkout-patched
$ git rm setup.cfg
$ git commit -m "Removal test"
$ git-dpm dch -- -i "Testing removal"
$ head -15 debian/patches/0009-Removal-test.patch 
>From d97af66bec2a27d567f010ee0b7b3cec7641cf2a Mon Sep 17 00:00:00 2001
From: Nikolaus Rath <nikol...@rath.org>
Date: Wed, 4 Jan 2017 08:37:02 -0800
Subject: Removal test

---
 setup.cfg | 12 ------------
 1 file changed, 12 deletions(-)
 delete mode 100644 setup.cfg

diff --git a/setup.cfg b/setup.cfg
deleted file mode 100644
index eea4f..00000
--- a/setup.cfg
+++ /dev/null


> git-dpm has --no-ignore-deletions but AFAICT from the FM that seems to
> do something subtly different.  Is there a way to do the equivalent of
> dpkg-source --commit --ignore-removal, do you know ?

I don't quite follow you, but I suspect given the above this question is
answered?


It seems to me that something is wrong with the tree rather than with
git-dpm or dgit. I think this may be a consequence of the tree having
been imported from SVN. I will try to manually restore the missing files
in the master branch (bypassing git-dpm).

If this is indeed the problem, then this potentially affects all
packages from the debian Python modules team (they've all been batch
converted from svn-buildpackage).


Best,
-Nikolaus

-- 
GPG encrypted emails preferred. Key id: 0xD113FCAC3C4E599F
Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F

             »Time flies like an arrow, fruit flies like a Banana.«

Reply via email to