On Thu, 2009-12-24 at 11:04 -0500, Julian Foad wrote:
> I think this feature is exactly the sort of thing that would make sense
> as a plug-in to Subversion, and not as part of the core product
It's possible that the entire keyword system would make more sense as a
plug-in to some version control system which had been designed with
plugins in mind. I don't think a single keyword extension to
Subversion, which requires fairly deep working copy support, is good
fodder for a plugin system which might hypothetically be added onto
Subversion at such a late date.
> And then a minute after
> you have updated, the number is already out of date, so what is the
> point of this information? Or rather, why is this information related to
> files in your working copy? It isn't, it's related to the repository.
Although there are implementation issues to worry about, I don't think
there are any difficult specification issues. The possibility of mixed
revisions casts a few questions on what the keyword value should be, but
those questions are easily settled. If you run "svn info" on a file,
you will see the current rev (the rev of the working copy in which the
file resides) and the last-changed rev; the new keyword's value should
reflect the former.
Although people may ask for this feature in terms of the "global
repository revision number," I don't think anyone actually wants a file
to magically display the current repository revision in real time. They
just want to know what revision of a repository is reflected by a
working copy, without any scripting machinery (either because their
build system makes such scripting hard, or because their content is not
software and doesn't have a build system).
It's fine to say "this is too hard to implement in our architecture,"
and in fact we've been saying that for a long time. But it's silly to
say "we don't know what this would do or why anyone would want it" when
it's pretty clear what it would do and why people want it.
Also, see:
http://subversion.tigris.org/issues/show_bug.cgi?id=1975
Received on 2009-12-24 18:27:33 CET