On Fri, 2006-01-13 at 22:18 +1100, John Vandenberg wrote: > Hi, > > I have a number of questions and suggestions for 0.44 that I have > throw together; feel free to answer a few at a time and come back to > others.
My $0.02. > > 1. should patches for review contain quilt.changes updates? My > initial impression is that this would be a good idea, as the contents > of this file are part of the published package; but, it would hinder > re-ordering patches. perhaps this can be avoided by adding a script > that generates a quilt.changes update from the patch header, to be > used prior to commit. Nah. Like you said, it's just adds overhead. And the person that submits the patch isn't necessarily the person that commits it. I think the committer should add the text to quilt.changes. > 2. should the patch header be used for the cvs commit message? I would think so. > 3. should I start a Contributions FAQ file? if so, in doc/ or the > base directory? Can you elaborate on what this would be used for? Like a "HOWTO contribute"? Or more of a "these people have contributed"? > 4. I often add features to quilt because I need to use them then and > there, and then months later on forget what additional patches I have > installed on that box. Timestamps and version numbers are useless; > the only way I can deduce what it can do is to know the list of > patches that were applied. does this sound like a reasonable addition > to development versions of quilt? Need more info. If you've used quilt to patch your local copy of quilt, aren't the changes in your patches directory? > 5. I also forget where I was up to with a workpit, and come back to it > wondering whether if I need to do a fork/refresh and diff the two > patches. I would like to add a status.in to tell me what the current > state of play is. My thoughts were that it should display: > > Patch: $(quilt top) > Status: (up-to-date|stale|missing) > > $(quilt header) > > The missing option would appear whenever there are changes against > .pc, but no saved patch. A quilt status command might be cool. I could see myself using it at least. > 6. I would like to add a shell.in, that opens a shell with patchfns > loaded, so I can trial certain operations that doesn't warrant > cracking open the source and adding a feature. However, this command > could produce dire consequences in the wrong hands; is it ok to > install this? should it issue a strong warning to the user? Erm, not sure if this should be in quilt itself. Starting a shell and sourcing patchfns shouldn't be that hard... or maybe I'm missing something? > 7. Is it ok if I start moving the autoconf tests to separate files > into m4, and submit them to the ac-macro archive. > > http://autoconf-archive.cryp.to/ Are you looking to do that and push the resulting stuff back into quilt? If so, I have no idea. But even if you aren't, quilt is licensed under the GPL so I don't see anything stopping you. josh _______________________________________________ Quilt-dev mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/quilt-dev
