Date: Tue, 12 Jun 2001 18:46:31 +0100 From: Nik Clayton <nik@FreeBSD.org> To: doc@Freebsd.org, cvs-committers@freebsd.org Subject: On the recent -i18n changes Message-ID: <01061218463104.37769@clan.nothing-going-on.org>
next in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi folks,
Right, I'm back now. Somewhat shattered from the driving (~450 miles in 12
hours), but we did a good presentation at the West Yorkshire Linux User
Group, so that's another deposit in my "Spreading the FreeBSD word" karma
account.
In re the recent commits to the src/ and (in particular doc/) trees as a
result of discussions on the -i18n mailing list, I've got a few points to
make. Hopefully people won't take these personally, but we can use them to
try and ensure that communications breakdowns like this occur less
frequently in the future.
1. I should apologise for suggesting that the changes were carried out
without a repo copy. A combination of bogons in my local /home/ncvs
(note: not the checked out copy, but the repo), plus no advance warning
from any of the repo-meisters meant I jumped to conclusions that could
have been avoided had I checked on freefall.
2. I don't have a problem with the nature of the changes. I like the
CVS trees to remain consistent with one another, and if that entails a
bit of renaming now and then, so be it. Those with long memories
should recall I spearheaded the original renaming in doc/ a few years
ago to keep consistent with /usr/share/locale/.
3. Freezing the doc/ tree was not something I did to try and prevent the
ongoing changes. Rather, here were a collection of changes to doc/
which had not been discussed on the -doc list, where the last time
changes like this were made (see point #2) we talked about them first
for some weeks (May 13th to June 25th, if my mail archives are
reliable).
I did not want to have the situation where perhaps a mistake had been
made, and then commits had happened to the tree -- that would make any
fixing that would need to be done unnecessarily complex, as changes
would need to be backed out, and then rolled forward as necessary.
4. Communication about this change was significantly sub-optimal.
The initial heads up notice was
(a) Not very detailed
(b) Delivered a few hours before the commits happened, over a
weekend.
That is not enough time to allow for a response, particularly on a
global project like FreeBSD for changes affecting the entirety of the
repository. While the change itself might be technically quite simple,
the *potential* for breakage is huge, and we should be considering the
warnings we put out in e-mail based on this potential for damage, not
how technically simple the change is.
5. The changes had clearly not been tested, as shown by the various
problems that Bruce and others are seeing.
I would have liked to have seen two things happen:
1. When the discussion started on -i18n, a HEADS UP should have been
posted to other lists involved (-doc and -current probably), explaining
what the topic was, and inviting interested parties to subscribe to
-18n for the duration.
2. A detailed HEADS UP, with at least 72 hours notice, detailing the
changes that were going to be made, and pointers to the discussion
on -18n on the mailing list archives.
This should also have included a pointer to a patch that implemented
the change. so that interested third parties could try it out.
Yes, it's effort to put that together. That's one of the crosses
to be borne as a FreeBSD committer.
Going forward, what do we do?
1. As far as I can see, we don't need to back out the change in doc/. I
haven't yet done a full CVSup and build to test it (cable modem is
down, I'm back to dialup, and it won't be fixed until tomorrow).
I expect there will be changes to be made to the www/ tree to
compensate for this, including changing a few URLs. That's a hurdle
I think we can jump with no problems (anyone with differing opinions
should chime in now). As far as I'm aware, because the www/ build
doesn't remove any old files, all the links on the website still work
(correct me if I'm wrong).
2. I don't know about the release-notes section of the tree. As far as
I'm concerned, Bruce is the final authority on that chunk -- if the
change is causing him problems that will take time to resolve then we
should back out the change there, live with the fact that it's
inconsistent, and then plan how to remove the inconsistency without
introducing world breakage.
3. Unfreeze the doc/ tree.
N
- --
FreeBSD: The Power to Serve http://www.freebsd.org/
FreeBSD Documentation Project http://www.freebsd.org/docproj/
--- 15B8 3FFC DDB4 34B0 AA5F 94B7 93A8 0764 2C37 E375 ---
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (FreeBSD)
Comment: For info see http://www.gnupg.org
iEYEARECAAYFAjsmVXcACgkQk6gHZCw343UgmgCeIns7X55b8Y4EEuvZG0LVCnt6
ljcAn3dnqnXBGGkQCmewXXO8e/OPmsrE
=sf4o
-----END PGP SIGNATURE-----
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-doc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?01061218463104.37769>