[Top][All Lists]

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

Re: [Gnu-arch-users] Inconsistency of Added-files and Removed-files

From: Mikhael Goikhman
Subject: Re: [Gnu-arch-users] Inconsistency of Added-files and Removed-files
Date: Mon, 16 Aug 2004 22:34:23 +0000
User-agent: Mutt/

On 16 Aug 2004 11:54:44 -0700, Tom Lord wrote:
>     > From: David Allouche <address@hidden>
>     > Having been recently converted to test-driven programming, I'm writing
>     > tons of test case for pyarch and uncovering some glitches in tla.
>     > I have just found that when explicit id-tags are used, the patchlog
>     > header "New-files" does not mention explicit id-tags, but the
>     > "Removed-files" headers does mention them.

Both New-files and Removed-files headers obviously include .arch-ids/
files. The problem is only with import that indeed should be fixed.

Here is, for example, a page generated from a non-import log:


>     > I have not checked the way Moved-files (or whatever the name of the
>     > header summarizing renames) behaves with respect to explicit id-tags
>     > since it is not (yet) supported by PyArch and my current priority is
>     > upgrading the _existing_ functionality to support file name escaping.

Renamed-files include .arch-ids/ files as expected, see the same page.

>     > I _can_ write a workaround for the inconsistence if the correct
>     > behaviour would be not to show explicit id-tags ever in patchlogs, so I
>     > would like to know what _is_ the correct behaviour.
> These headers are not heavily used.  Their primary use is, for human
> consumption (e.g., grep for a filename).
> Therefore, I think, the explicit id files can be safely omitted from
> the headers.

Please don't reduce the log information, fix the import log instead.

Moreover, I would like to request several more log headers, like the ones
to report permission or symlink changes, and the one to report changeset
type (normal, tag, import, base).

I don't see a good reason to make additional tla call (much more expensive
than log) just to get an accurate changelog summary. Is not this the whole
patchlog purpose?


reply via email to

[Prev in Thread] Current Thread [Next in Thread]