Ienup Sung wrote: > Roland Mainz wrote at 07/17/06 21:08: > > Ienup Sung wrote: > > > >>Building ON doesn't require any locales except the C/POSIX locale > >>which is the system default locale embedded in and provided by > >>the libc. > >> > >>Since people can choose not to install locales including the Japanese > >>locales you mentioned or delete locales after installation, build machines > >>may or may not have the locales. > > > > > > That's very unfortunate since this eliminates almost all "simple" ways > > to catch breakage of the multibyte locales during build time of > > OS/Net... ;-(
BTW: I commited http://polaris.blastwave.org/changeset/342 to the ksh93-integration tree to catch missing locales. > At this point, ON gate has no hard dependency to localized components > to make it self-contained in my understanding. And the current > model of integration and I18N framework naturally allow and draw people to > do the G11N work including testing at the G11N workspaces/gates. (I'm not > saying one way is better than another but just and merely stating > the current situation.) > > I though think you might have some interesting suggestion and idea that > many people would like to hear about. (Possibly related to your idea, > one time a few people also consider about having a test-purpose made-up > locales but that idea didn't really gain any traction due to maintenance > cost and such.) The idea we had (and implemented) is to run the test suite provided by AST/ksh93 as part of the default OS/Net build AND run it in different locales to catch locale-specific bugs (which already unearthed a couple of problems during development of the prototype, that's why not all locales on my wishlist are enabled yet). Effectively this handles a couple of situations/problems, including: - Someone breaks libc/libnsl/libsocket/etc. and ksh93 goes mad elsewhere (and we get the blame). - Bite the matching engineer back immediately instead of doing a full iteration of "hit bug; file bug report ; assign bug report to engineer ; search for the bug ; fix bug ; commit fix back" when he/she breaks something - Catch errors/mistakes during code updates, e.g. when I update the ksh93 sources and forgot one piece up update. - Catch errors in other locales (with the main focus on multibyte locales). The tests are run once per locale - which currently include "C" and "en_US.UTF-8" - but I'd like to include at least "hi_IN.UTF-8", "ja_JP.UTF-8", "ja_JP.PCK", "ko_KR.EUC", "zh_TW.BIG5" and "zh_CN.GB18030" in this list (once the |libast::strtod| issue (see http://mail.opensolaris.org/pipermail/ksh93-integration-discuss/2006-July/000539.html) has been fixed) to cover a large range of languages + encodings (note that some applications (not ksh93 which only uses the widechar API) have hardcoded paths for UTF-8 locales so we have to include tests for multiple sets of encodings per language in some cases ("ja_JP.UTF-8" vs. "ja_JP.PCK" is a good example)). ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) [EMAIL PROTECTED] \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;) _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
