On 29.01.2012 18:25, Garret Wilson wrote:
> The story ends this way: I've spent days writing migration code which
> traverses the gigabytes of my various repositories and converts legacy
> property names (the ones I've been using for years over DAV+SVN, not
> knowing they were invalid in JavaHL) to "valid" Subversion property
> names (according to the source code; there exist no public
Sorry you wasted your time ... it all looks like a simple
search-and-replace in a repository dump file, not something you'd have
to write a lot of code for.
> Creating this throwaway migration code was many times over more work
> than it would have been simply to loosen slightly the restrictions in
> the Subversion client source code to be more liberal with property names.
Are you sure about that? Have you considered how quoting the names over
HTTP/DAV/XML would work?
I can't help wondering why you didn't ask about valid property names
/before/ you created a bunch of invalid ones. Sounds like you made one
too many assumption.
> On the other hand, writing this migration code was many times over
> /less/ work than all the time and effort I would have spent trying to
> get permission (or even interest) in modifying the Subversion source
> code, or even to publish a specification.
> I did get a bit of constructive discussion to at least understand the
> issue, and I appreciate the "+1" I received for the concept of
> creating a specification and making property name requirements
> consistent across clients. But I have been given no authority to
> create any specification or modify any code,
What kind of authority do you think you need? Perhaps from your manager?
Because you certainly need none from this project if you want to submit
patches, be they documentation or code.
Received on 2012-01-29 19:56:35 CET