Re: [Gnu-arch-users] New aba command: revert, name clash with tlash

[David Allouche]

On Fri, Mar 26, 2004 at 12:54:42PM -0500, Aaron Bentley wrote:
> David Allouche wrote:
> >
> >"revert" is already the name of a tlash alias for "tla replay --reverse"
> >which is IMHO very handy and natural.
> 
> Hmm...  extract?  reject?

It would make sense to rename the log-preserving undo "reject".

The scenario would be:

  1. star-merge with contributor's branch
  2. reject unwanted patches

"Revert", the verb for "reverse", seems better suited to "replay
--reverse".

> >Eventually aba subcommands will be usable from tlash in the same way as
> >tla subommands (without the wrapper name), maybe you could change the name
> >of your command, or propose a better alias for "tla replay --reverse"?
> 
> Oh, are we out-of-sync?  I thought they worked already.  James has 
> recently merged my latest tlash-compatibility hacks.

I am not yet up to date with that.

If that works okay, I will make some tweaks to fix the existing name
conflicts and add a tlash.d file to add the aba commands directory to
the PATH.

At this point, the integration will be completed, thus the "eventually".

> >BTW, an outstanding issue with tla is the ability ignore the patchlog
> >part of (non-pathological) changesets, or apply only the patchlog part.
> >That would effect several commands, at least "changeset",
> >"apply-changeset", "replay", "changes", and maybe also "undo", "redo",
> >"apply-delta", etc.
> 
> We can factore out the metadata-restoring stuff, if useful.  Both 
> patch-logs and tree-version may be altered by an undo.

The ability to use and view only the patchlog or non-patchlog part of
changesets is a generally useful feature which I believe is okay with
Tom to put in tla.

Pointing at the forest behind the tree at hand, do not look at the
finger.

-- 
                                                            -- ddaa