On Thu, Sep 04, 2003 at 12:21:39PM +0200, Stefan Gybas wrote:
More and more parts of OpenI18n are being merged into the LSB specification, so I think the goal is to merge both documents.
Great. I'm not going to write to a goal any more than I'm going to write to a test suite.
However, I must admit that -- like you -- I could not find such a requirement in the LSB 1.3 specification.
At least we agree on that. :)
This surely is the long-term goal. But we need to get there step by step and IMHO should start with the basic utilities, like OpenI18n does.
Yes, which means a *comprehensive* examination of the basic utilities. The test suite isn't that. It also needs to be done on a package-by-package basis. (I'm not impressed with the idea of a coreutils package with a list of utf8-enabled features and a list of features where utf8 is broken--which is, I think, where the patch in 186140 will leave things.)
The patch must have been integrated by Red Hat and SuSE, otherwise their latest distributions wouldn't be LSB 1.3 certified. So it has at least been tested by others - like the patches to diffutils.
Then why haven't they been merged upstream if they're so well tested correct? Other vendors have shoved half-assed patches into their packages in the past, so I don't consider that to be sufficient criteria.
You're not the only one. Could you talk again to the upstream authors?
I do fairly regularly. :)
You are probably correct here, but, again, this is a start.
A start is something you do at the beginning of the release cycle, not the end. I've spent a lot of time trying to make sure that coreutils is compliant with POSIX/SUS/LSB-spec (as has upstream) but this push to have the package be multibyte compliant also is just coming out of nowhere. (The bug was in for quite a while, at a low priority, without any particular comment that it was an important issue.) We're already past the Sep 1 date for major changes to fundamental packages. AJ--do you want to push that date back a couple of months? If the answer is yes then someone can have a reasonable chance at looking over the whole coreutils package for multibyte issues. (Basically, anything that uses a char must be looked at, right?) Even then I can't commit myself to a deadline because I have no way to test multibyte characters, so someone else would have to pick up the ball for this. Mike Stone
