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

Re: please consider debconf 1.4.30.13 (t-p-u)



On Thu, Apr 07, 2005 at 04:56:21PM -0400, Joey Hess wrote:
> Some overly strict checks in debconf-set-selections make it very hard to
> extend the preseed file format in backwards compatible ways. This could
> lead to problems post-sarge, since etch d-i will likely be used to
> install sarge systems, and the preseed files are used for both d-i and
> second stage preseeding.

> For example, if a new debconf question type is added post-sarge, and is
> used by d-i, it would be problimatic to preseed the answers to such
> questions, since the preseeding file would also be parsed by sarge's
> debconf, which would choke on the unknown question type.

> Another example is some per-question preseeding w/o setting the seen
> flag that I am working on for post sarge. Sarge's current debconf would
> choke on a preseed file that used this enhancement.

> I think it's best to anticipate these problems and try to solve them by
> making sarge's debconf-set-selections less strict in its checks so the
> preseed file format can be extended later. I came up with a simple patch
> to do this, and have uploaded it in debconf 1.4.30.13 to t-p-u.

> I've attached a diff of all changes since debconf 1.4.30.12. 1.4.30.12
> is itself not yet accepted into testing, although it only involves
> adding some translated strings. If you really don't want to accept the
> changes in 1.4.30.12, I can back those out and make a new upload with a
> much smaller diff from 1.4.30.11.

Approved.

Cheers,
-- 
Steve Langasek
postmodern programmer

Attachment: signature.asc
Description: Digital signature


Reply to: