Re: [Gnu-arch-users] Re: tla import versus tla commit

[Thomas Lord]

    > From: Stefan Monnier <address@hidden>

    > > For a single user, with an almost entirely "local" environment,
    > > implementing undo (even of commits) is quite practical, although a tad
    > > tedious.   Arch needs to keep a more detailed track of the revision
    > > libraries it knows about, the mirrors, the timing of mirror updates,
    > > and the location of project trees.   Matthew Dempsky has recently
    > > started working on this problem, although his time is very limited (he
    > > is a student) and so no particular rate of progress can be promised.

    > Sounds like a lot of work, with a lot of room for "errors" (i.e. things we
    > can only assume to hold even though they may not always do so).
    > And for very little benefit (in my case at least) since most/all my 
archives
    > are pretty quickly disseminated to revlibs and stuff on various machines.

    > I think we'll be better off following an approach similar to NNTP (where 
the
    > problem is also that once data is sent, you can't in general trace back 
all
    > the copies): create a new kind of changeset which modifies a previous one.

Ick.  Sorting out such a collection of changesets by hand seems
horridly tedious and, operating globally (rather than in the
special-case local form I advocate for above) in the arch namespace
but extended with some kind of undo semantics sounds hopelessly
intractible (what is to become of mirrors, for example?  When can they
safely be used in your Brave New World other than when an
authoritative master server can be contacted?).


    > This way we can change the past (e.g. change the log message of an old
    > revision), and the info will be disseminated using the same mechanisms as
    > every other change: no need to keep track of mirrors, caches, revlibs, ...

Build higher-level data structures atop the arch namespace if you want
to do that.

-t