Le 17/10/2010 16:07, Florent Becker a écrit : > > New submission from Florent Becker <florent.bec...@ens-lyon.org>: > > This is the beginning of a support for "no working repositories". > > That is, repositories without a working directory, not non-working > repositories. > Except of course, that they are currently not tested or complete in anyway, so > it's non-working after all. Please use, test and report any issues while i add > tests and support for other commands. > Here come the grumpy old bits:
> What problem does the proposed feature solve? In repositories where no editing takes place, having a working dir is a waste of space. Not having one also enforces the invariant that these repositories don't have unrecorded changes. > What are the user stories? Jane has limited storage space on her http server, and each of her darcs repositories there store the working files twice. Jane is not happy. Jane made a post-hook to sweep the working-dir clean after each push/pull/apply. And a pre-hook to restore it before. Jane feels dirty. Does her trick even work? Florent wants to implement push to a dumb server over scp. This is not going to be pleasant if repositories on dumb servers have changes in working directories. Florent is not subtle. Florent says: "let there not be a working directory at all". > Does this change any pre-existing workflows? Does this introduce any > incompatibilities? The change ratifies the current practices that there are two kinds of repositories, public repositories which get pushed to/pulled from, and private repositories which get recorded into. A working-dir-less repo is guaranteed to be public(-izable). It creates the incompatibility that some repo are not editable, but this seems desirable. > How do we preserve the conceptual integrity of Darcs? Is the UI for > this feature really Darcs-ish? / > What are the possible unintended interactions with other pre-existing > features? ?? > What are the alternative approaches to solving the same problem? Why > do we prefer this one? We once had(?) pristine-less repos. Now that things are hashed, it is no longer a possibility. > Who are the stakeholders? Who is going to benefit/be affected by this > feature: end-users, repo farms like patch-tag, darcs developers? All of them, but especially repo farms. Florent _______________________________________________ darcs-users mailing list darcs-users@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-users