On Jun 24, 2007, at 8:37 AM, Ryan Schmidt wrote:
On Jun 23, 2007, at 18:16, Paul Guyot wrote:
On Jun 24, 2007, at 6:12 AM, Juan Manuel Palacios wrote:
About this commit though.... it was discussed on the this list some time ago whether GSoC work should be committed to trunk or to a branch, and the consensus was that, this being SCM and therefore moving back and forth being an easy operation, GSoC students should commit to either based on their own judgement. Seems to me like you clearly don't feel comfortable enough to commit this work to trunk, so why not put it in an appropriate branch? Seems to me like a more standard and expected place to find experimental code like this.
What's the difference with subversion between a branch and a directory such as users/sfiera/registry2.0/?
If it were a complete branch of base, anyone could just check it out, compile it, and play with it. When registry2.0 is just a directory by itself like that, though, it requires more work: at least checking out base, then checking out (or exporting) registry2.0 into the correct place inside it. Maybe more changes; I don't know, since I don't really know how base works. If it were all codified into a branch, I wouldn't need to know. I'd just check it out and use it.
True. Yet, registry2.0 directory is completely inert C code for now, it's just a set of native Tcl commands. Copying this source code to trunk doesn't mean it will be used for any operation. It will not even be compiled with a ./configure && make sequence in base/ directory. Out of what Chris checked in, only tests can be executed. I believe this is why he didn't check in a full copy of base/. The rationale for checking the code in is that (a) he can use our SCM (b) we can see what he's up to. I had a previous version of this code by mail, but I appreciate that he publishes this in his users/ directory, so everyone can see. Paul