Stefan Sperling wrote:
Julian Foad wrote:
> Shelving: [...] uses a crude, complete "svn checkout" to create a shelf
How much work do you think is required to fix that?
Was there a plan already or is design work required?
It needs design and quite some work, so off the radar until someone
invests further in it (whether financially or as a volunteer).
Sounds like we should reinstate shelving-v1 on trunk and put the
shelving-v2 [edit: it's v3] code on a branch, for now?
Maybe yes.
> svn x-wc-copy-mods:
Great. Could this become an optional mode for 'svn copy'?
No, it's unlike "copy" and more like "diff-and-patch", but with an API
that supports all committable changes, not only those that the "diff"
and "patch" serialization format currently supports.
> svn info --x-viewspec=
>
> This is not ready to be promoted to non-experimental. This command itself
> is not too bad; what is missing is a proper efficient implementation of
> "apply a viewspec". I consider that forces us to keep this "report the
> current viewspec" command as "experimental".
In that case I would prefer to move it to a branch instead of having
it on trunk.
In terms of whether it is visible by default to an users, I agree. In
terms of where the source code lives, I don't mind -- I would be equally
happy if it were on trunk and disabled by a compile time or run time
feature flag.
In particular, I think it is important to keep the visible behaviour
"stable" which implies I think we should stop exposing these
experimental features by default; and at the same time we don't want to
make anything more difficult to develop than necessary, so maybe we
should not be too eager to push things onto branches.
- Julian
