gnu-arch-users
[Top][All Lists]
Advanced

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

[Gnu-arch-users] Re: implicit discussion


From: Miles Bader
Subject: [Gnu-arch-users] Re: implicit discussion
Date: 16 Apr 2004 12:13:38 +0900

Aaron Bentley <address@hidden> writes:
> > changing to a representation that used a single file per directory
> > to hold explicit tags would improve things quite a bit.  However
> > there doesn't appear to be any rush to change the implementation.
> 
> Archives contain the paths of those .id files.  We can't change the 
> representation of explicit without
> a) breaking archive compatibility or
> b) layering an alternate namespace on top of the filesystem.

It doesn't have to break the old reprensetation -- just add an alternate
representation for the same tag namespace, but don't get rid of the code
implement the old representation.

Of course because archives containing a new representation won't be
usable by old versions of tla, a new archive-version would have to be
used.

So when tla++ with the new super-explicit tags gets released, nothing
will change.  When the maintainer of some tree feels it's safe he'd
create a new archive with the new archive version (tagged to the old
archive), and commit a changeset that removes all the .arch-ids/*.id
files and adds the new per-directory id files.  The changeset wouldn't
contain any changes to normal files, just to meta-info files.

[I guess you could also make some sort of compatibility hack that uses
the new representation in project-trees, but then converts it to the old
many-files representation in changesets -- but that seems much too
complicated...]

-Miles
-- 
Is it true that nothing can be known?  If so how do we know this?  -Woody Allen




reply via email to

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