Matthieu Moy <[EMAIL PROTECTED]> writes: > Stephen Leake <[EMAIL PROTECTED]> writes: > >> If we use letters for now, and gradually convert code to use >> dvc-diff-ewoc->status, then at some point we can switch to using symbols. >> >> On the other hand, maybe it's not all that much work to just go >> straight to symbols. I guess I'll look at that some more. > > I'd say: don't discard the idea of symbol too fast. It would take some > time, but save some time in the future ... > >> Is the list of symbols I suggested a good start? > > It seems to be, yes. There's also another thing to take care of, which > is the type of file (file, directory, symlink). That should be a > different field in the ewoc (it's currently also a letter IIRC).
Ok. If I'm adding fields, I'd want to use a defstruct. >> If the user edits a file, and then renames it, and then runs >> dvc-status, won't it show up as 'missing old-name' and 'unknown new-name'? > > Depending on whether the user ran "XXX mv old-name new-name" to > actually rename the file. Right. I (and members of my team at work) can't be bothered to remember to do that; I want to edit files "just like I used to without this fancy CM system", and expect DVC to remind me of what needs to be done. But clearly DVC needs to handle all such cases. > But even if we use a singleton 99% of the time, we should not restrict > the model to singletons. Using a list is probably not more > difficult. Right. This is getting into a lot of patching. I suspect it would be better if I had a public database (if that's the right term for bzr) somewhere, that you could pull from. I have a website, but I'd have to pay more to set up a bzr server there. I can do that, but is there another option? -- -- Stephe _______________________________________________ Dvc-dev mailing list [email protected] https://mail.gna.org/listinfo/dvc-dev
