On Thu, Sep 04, 2003 at 09:57:58AM +0200, Stefan Gybas wrote:
In order to be LSB 1.3 certified Debian must comply with these specifications (see http://www.opengroup.org/lsb/cert/docs/faq.tpl#general5)
I don't see an rc requirement for LSB certification, the wording of our release goal is "Packages must not conflict with requirements of the LSB, v1.3". The thing you're quoting says: What Specifications does LSB Certification cover? *LSB Certification currently covers the following Specifications:* The Linux Standard Base Specification 1.3 The Linux Standard Base Specification for IA32, IA64, PPC32, PPC64, S390 and S390X The OpenI18N Specification (formally the Li18nux 2000 Globalization Specification Version 1.0 with Amendment 4) Signficiantly, this is from an FAQ for LSB certification, not from the LSB standards document. IOW, OpenI18N is part of LSB certification but seperate from the LSB specification. AJ, you need to clarify what the release goal is. (Currently: "Packages must not conflict with requirements of the LSB, v1.3. (eg, if you provide a library specified in the LSB, you must be compatible with the LSB specification of that library)") That mentions the LSB specification, but not LSB certification.
Then OpenI18n spec only requires multi-byte handling for programs listed in chapter 4 (which includes the shell and other basic utilities).
Which is half-assed. If utf8 support is a goal we need to support it across the board.
I think this is a goal than can be achieved for sarge without delaying the release.
If someone had explained things better six months ago, maybe. Only in this last email has there been enough detail to understand what the issue is. IMO it is way too late in the cycle to add such a massive and invasive patch. The one you provided is larger, by an order of magnitude, then any other patch in the coreutils package. I like to keep my patches small for a reason: I don't want to introduce problems or massive differences from the upstream version; this patch is likely to do both. I'd much rather have that sort of change introduced in the upstream version rather than as a debian-specific change. I talked to coreutils upstream about the multibyte issues about 9 months ago, and the feeling then was that they might be addressed by this summer. I think that the issue on that end is the same as it is on mine--lack of time to properly integrate such a massive chage to the source code for a relatively low priority item. Among other issues, I seriously doubt that this patch covers all multibyte issues in the coreutils package. It might cover all of the issues in the test suite, but that's why you don't write to the test suite. This goes back to my feeling that if we make utf8 support a release requirement we need to support it fully. Mike Stone
