Christian PERRIER <bubulle@debian.org> (2014-09-28):
> Quoting Cyril Brulebois (kibi@debian.org):
>
> > dh-autoreconf usage has been introduced a while ago, specifically to
> > avoid such crazy diffs. Can you please tell me what happened on your
> > side and how we could improve things so that it doesn't happen again?
>
> What happened is basically what I do for every build..:-)
>
> pdebuild --pbuilder cowbuilder -- --basepath /var/cache/pbuilder/base-sid.cow --distribution sid
Another example, between partman-crypto/76 and partman-crypto/77:
/tmp/z3cjgPs0hl/partman-crypto-77/base64/tests/case5 |binary
/tmp/z3cjgPs0hl/partman-crypto-77/base64/tests/case6 |binary
/tmp/z3cjgPs0hl/partman-crypto-77/base64/tests/case8 |binary
/tmp/z3cjgPs0hl/partman-crypto-77/base64/tests/case9 |binary
partman-crypto-77/base64/Makefile | 23 ---
partman-crypto-77/base64/base64.c | 101 ---------------
partman-crypto-77/base64/tests/TESTS | 13 -
partman-crypto-77/base64/tests/case1 | 1
partman-crypto-77/base64/tests/case2 | 1
partman-crypto-77/base64/tests/case3 | 1
partman-crypto-77/base64/tests/case4 | 1
partman-crypto-77/base64/tests/case7 | 1
partman-crypto-77/base64/tests/want1 | 1
partman-crypto-77/base64/tests/want2 | 1
partman-crypto-77/base64/tests/want3 | 1
partman-crypto-77/base64/tests/want4 | 1
partman-crypto-77/base64/tests/want5 | 6
partman-crypto-77/base64/tests/want6 | 28 ----
partman-crypto-77/base64/tests/want7 | 1
partman-crypto-77/base64/tests/want8 | 64 ---------
partman-crypto-77/base64/tests/want9 | 65 ---------
partman-crypto-77/blockdev-keygen | 21 ++-
partman-crypto-77/ciphers/loop-AES/AES128/desc | 1
partman-crypto-77/ciphers/loop-AES/AES128/module | 1
partman-crypto-77/ciphers/loop-AES/AES192/desc | 1
partman-crypto-77/ciphers/loop-AES/AES192/module | 1
partman-crypto-77/ciphers/loop-AES/AES256/desc | 1
partman-crypto-77/ciphers/loop-AES/AES256/module | 1
partman-crypto-77/ciphers/loop-AES/keytype | 1
partman-crypto-77/ciphers/loop-AES/serpent128/desc | 1
partman-crypto-77/ciphers/loop-AES/serpent128/module | 1
partman-crypto-77/ciphers/loop-AES/serpent192/desc | 1
partman-crypto-77/ciphers/loop-AES/serpent192/module | 1
partman-crypto-77/ciphers/loop-AES/serpent256/desc | 1
partman-crypto-77/ciphers/loop-AES/serpent256/module | 1
partman-crypto-77/ciphers/loop-AES/twofish128/desc | 1
partman-crypto-77/ciphers/loop-AES/twofish128/module | 1
partman-crypto-77/ciphers/loop-AES/twofish192/desc | 1
partman-crypto-77/ciphers/loop-AES/twofish192/module | 1
partman-crypto-77/ciphers/loop-AES/twofish256/desc | 1
partman-crypto-77/ciphers/loop-AES/twofish256/module | 1
partman-crypto-77/commit.d/_numbers | 1
partman-crypto-77/commit.d/unsafe_swap | 56 --------
partman-crypto-77/debian/changelog | 20 ++
partman-crypto-77/debian/install | 1
partman-crypto-77/debian/partman-crypto-loop.install | 1
partman-crypto-77/debian/partman-crypto.templates | 10 +
partman-crypto-77/finish-install.d/05crypto | 71 ----------
partman-crypto-77/post-base-installer.d/61partman-crypto | 11 -
49 files changed, 45 insertions(+), 476 deletions(-)
which is basically the reversal of what happened between
partman-crypto/75 and partman-crypto/76 (plenty of noise got added).
I originally wrote this:
It would be nice if someone knowing about pdebuild/cowbuilder and
friends could figure out why we get crazy things like this. Since
we're about to freeze, it's quite doubtful that release team members
will enjoy getting such diffs to review for just some .po file
updates.
but looking at base64 and ciphers/loop-AES, that matches deletions in
the git history. Christian, I suspect git clean -ndf / git clean -xdf
would list then remove contents on your side?
In which case, the same probably happened for the other package, since
the introduction of dh-autoreconf meant we would remove old, generated
files.
Mraw,
KiBi.
Attachment:
signature.asc
Description: Digital signature