#91 merge 2.1.0 with git diff doesn't apply cleanly

[Hans-Peter Jansen]

Hi Petri,

in an attempt to merge 2.1.0 with current git (on openSUSE OBS), I stumbled across a strange issue, where translation updates are rejected.

Here's what I did:
* download and uncompress vdr-xineliboutput-2.1.0.tgz from this site
* create patch in a current GIT tree: git diff 2.1.0..747f03ee > xineliboutput-2.1.0..747f03ee.diff
* cd xineliboutput-2.1.0
* patch -p1 < ../xineliboutput-2.1.0..747f03ee.diff

Result:

patching file Makefile
patching file config.c
patching file device.h
patching file frontend.h
patching file frontend_local.c
patching file frontend_svr.c
patching file media_player.c
patching file media_player.h
patching file menu.c
patching file menu.h
patching file osd.c
patching file po/cs_CZ.po
Hunk #1 FAILED at 6.
Hunk #2 FAILED at 17.
Hunk #3 succeeded at 248 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 311 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 326 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 458 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 516.
Hunk #8 succeeded at 744 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/cs_CZ.po.rej
patching file po/de_DE.po
Hunk #1 FAILED at 7.
Hunk #2 FAILED at 16.
Hunk #3 succeeded at 247 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 310 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 325 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 457 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 515.
Hunk #8 succeeded at 743 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/de_DE.po.rej
patching file po/et_EE.po
Hunk #1 FAILED at 7.
Hunk #2 FAILED at 16.
Hunk #3 succeeded at 247 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 310 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 325 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 457 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 515.
Hunk #8 succeeded at 743 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/et_EE.po.rej
patching file po/fi_FI.po
Hunk #1 FAILED at 8.
Hunk #2 FAILED at 17.
Hunk #3 succeeded at 248 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 311 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 326 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 458 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 516.
Hunk #8 succeeded at 744 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/fi_FI.po.rej
patching file po/it_IT.po
Hunk #1 FAILED at 6.
Hunk #2 FAILED at 18.
Hunk #3 succeeded at 249 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 312 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 327 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 459 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 517.
Hunk #8 succeeded at 745 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/it_IT.po.rej
patching file po/pl_PL.po
Hunk #1 FAILED at 7.
Hunk #2 FAILED at 17.
Hunk #3 succeeded at 249 with fuzz 2 (offset 69 lines).
Hunk #4 succeeded at 315 with fuzz 2 (offset 72 lines).
Hunk #5 succeeded at 330 with fuzz 2 (offset 75 lines).
Hunk #6 succeeded at 463 with fuzz 2 (offset 78 lines).
Hunk #7 FAILED at 520.
Hunk #8 succeeded at 749 with fuzz 2.
3 out of 8 hunks FAILED -- saving rejects to file po/pl_PL.po.rej
patching file po/ru_RU.po
Hunk #1 FAILED at 7.
Hunk #2 FAILED at 16.
Hunk #3 succeeded at 247 with fuzz 2 (offset 68 lines).
Hunk #4 succeeded at 310 with fuzz 2 (offset 71 lines).
Hunk #5 succeeded at 325 with fuzz 2 (offset 74 lines).
Hunk #6 succeeded at 457 with fuzz 2 (offset 77 lines).
Hunk #7 FAILED at 515.
[...]

OTOH, if I generate the tarball myself from git, patch applies fine.

git archive --format=tar.gz --prefix xineliboutput-2.1.0/ 2.1.0 > vdr-xineliboutput-2.1.0.tar.gz

I looks, like the tarball for 2.1.0, hosted here @sf.net is generated from a deviating git tree.

Puzzled,
Pete