[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Gnu-arch-users] Patching patchlogs
From: |
David Allouche |
Subject: |
[Gnu-arch-users] Patching patchlogs |
Date: |
Wed, 14 Jan 2004 19:38:47 +0100 |
User-agent: |
Mutt/1.5.4i |
Is it considered acceptable to make a revision which modify the body of
existing patchlog files.
The changeset for this revision would contain a patch for a patchlog
file under ./{arch}. That would be equivalent to editing the
automatically generated changelog in a persistent manner.
I do not see anything it Arch design which forbids it, and I know that
will work, but what I am wondering is wether that is breaking an untold
convention.
In my specific case, I created a revision which cherrypicks several
existing patches. The cherrypicking is done by a script which extract
and copy the picked patch log body. I had modified the script to also
append the output of log-for-merge and committed a revision with several
applied-patches blocks with incrementally more entries, because later
log-for-merge runs included the output of former runs.
Another, more usual use would be fixing typos in the changelog.
The answer I had on #arch were roughly:
asuffield: the patchlogs are in the archive, there is only on correct
value for a patchlog.
I answered that in the same way that "tla logs" may give a different
output than "tla revisions", because changesets may add or remove
patchlogs, "tla cat-log" may be allowed to give a different output than
"tla cat-archive-log".
jblack: the changelog is history, it shall not be modified
I answered that the changelog is documentation, so it at least needs to
be fixed for typos or thrashing like what I put in my revision log.
Since I am quite sure our Benevolent Dictator has a well founded
opinion, I am explicitely asking him here what he thinks of it.
--
-- ddaa
- [Gnu-arch-users] Patching patchlogs,
David Allouche <=