[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: preparing for tc2

Quoting Joey Hess (joeyh@debian.org):
> Several fixes have gone into testing, but I still have these issues listed as
> important and unresolved in testing:
>  - some broken m68k images (fix: mke2fs fix needed; UNKNOWN)

Yep ? The IRC meeting lead to nothing on that topic.

>  - broken sparc64 module loading (fix: busybox-cvs 20040415-3)
>    The buybox-cvs fix cannot go in until we've dealt with netcfg. Either we
>    update netcfg, or we need a newer busybox-cvs that turns udhcp back
>    on to avoid breaking the netcfg in testing. I plan to do the latter
>    by a NMU soon, if nothing else is done.

We had a few people, namely Giuseppe Sacco and Martin Krafft, who
volunteered for netcfg testing..... maybe they should coordinate with
Joshua and Thomas, now.

>  - mips Installs on r4k-ip22 needs 36 mb ram (fix: unknown -- new
>    glibc?)
>    If it's the new glibc, then just rebuilding the images using the rc1
>    branch should fix these problems. Can someone check this?

Thiemo Seufer mentioned during the meeting he is OK to check this
thing, but he wasn't sure about having enough time

It seems that a quite large agreement is achieved about the new
languagechooser/countrychooser scheme. having it in tc2 would seem to
be a very important improvement, and is also needed for BiDi support
in 2nd stage.

This is not strictly release critical and I still remember we both
agreed to postpone these after rc1 (so, after tc2) at the end of

However, lot of time passed since then....and I tested this stuff
nearly daily, always changing languages here and there (I even have
now a framework for testing netboot images and possibly find nasty
crash bugs like the one you found for Arabic).

So, I think I need to convince you that pushing
languagechooser/countrychooser/lowmem (the 3 are needed in the same
time) to testing will be an enhancement and NOT introducing new bugs
like it may have been in the past.

Oh, and while I'm at it, choose-mirror, if pushed to testing needs
either a solid review of the code in sort-countries....or commenting
this script temporarily if unsure (thus needing a new upload). BTW,
this script can be "set -e".

tc2 needs of course a small string freeze. I would say that 3-4 days
may be enough as long as a one-week delay for warning translators
before the freeze is allowed. A string freeze over a week-end is
better (say, from a Thursday 18:00 up to a Sunday 18:00 UTC)

Reply to: