- From: Terje Bless <link@pobox.com>
- Date: Sat, 11 Sep 2004 14:38:08 +0200
- To: QA Dev <public-qa-dev@w3.org>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Bjoern Hoehrmann <derhoermi@gmx.net> wrote:
>Changing `check` to use SGML::Parser::OpenSP will obsolete some code
>which depends on `onsgmls` for which we do not have replacements yet,
>these are in particular the
>
>* outline * sp * noatt * esis * errors
>
>configuration options. I would suggest to drop "esis" without any
>replacement, re-implement "outline", "sp", and "noatt" as Perl SAX
>handlers and re-implement "errors" in terms of escapeHTML() and
>Data::Dumper::Dumper() on the list of "problems" the various problem
>reporters report. Any better ideas?
As a rough outline it sounds good. I may be more ambivalent about the
specifics. :-)
The current «esis» option (which dumps out the raw ESIS output from `onsgmls`)
is purely a debugging option — so nuking it is a valid option — but a very
usefull one at times. I think it could be usefully reimplemented similarly to
your proposed «errors» implementation, except working on a complete event tree
from the parsing. IOW, it would complement (or be completely obsoleted by) the
new «sp» implementation.
>If we cannot find people willing to implement replacements, I suggest we
>drop their functionality aswell, less code, less testing, ... :-)
I will not accept sacrificing core user-visible features on the altar of m12n!
If we don't have implementations of these features on the m12n core, we do not
release the m12n-ized version.
This doesn't, of course, prevent m12n progressing; only switching the v.w.o
deployed service to adopting it.
- --
"Hath no man's dagger here a point for me?" - Leonato, Governor of Messina.
See Project Gutenberg <URL:http://promo.net/pg/> for more.
-----BEGIN PGP SIGNATURE-----
Version: PGP SDK 3.0.3
iQA/AwUBQULxrqPyPrIkdfXsEQIeVwCeM4HH5XesCrRXtP8BuLTSkxcTQgQAoNR0
43O3qsqDkt77EGCrBRea3LRH
=rApF
-----END PGP SIGNATURE-----
Received on Saturday, 11 September 2004 12:38:13 UTC