On Fri, Feb 10, 2012 at 00:51, kcrisman <kcris...@gmail.com> wrote: > I think it's great that we've made the updates to the developer guide > that explain queues, and I have been using them a lot more for the > last year or so. But remember, a lot of the infrastructure was put in > place in order to aid new developers who known *nothing* about > branches, git, revision control, anything. > > For this sort of newcomer, having all tools available as "close to > home" as possible - like "sage -clone" - lowers the learning curve > enough that they might want to learn enough to be more efficient. If > you want to make more than one patch, or apply a patch and then write > a reviewer patch, going back to an "untouched" Sage is hard without > using queues or cloning.
I am not really convinced that "Sage branches" are any more (or less) intuitive or "close to home" than patch queues. Sure, copying directories seems easy enough, but in the end you still need to produce a patch somehow. But in any case I would really like to get rid of "Sage branches" from the Sage infrastructure itself. Of course you would be entirely free to make multiple Sage library directories in $SAGE_ROOT/devel and point the symlink $SAGE_ROOT/devel/sage at them, if you so desired, just as Jason is doing with his multiple sagenb directories. The documentation could even walk the user through doing this. But having Sage *expect* this to be the situation is very annoying, IMO. I should ideally be able to keep the Sage library in $HOME/src or wherever, call the directory whatever I want, and point the $SAGE_ROOT/devel/sage symlink at it (just like is perfectly possible with sagenb). However, this currently *breaks* Sage. Sage will not start unless the actual directory, not a symlink, is situated in $SAGE_ROOT/devel/foo where foo is some string starting with "sage-". Would you have any objection to getting rid of the actual commands `sage -clone`, `sage -b <foo>` (just plain `sage -b` would stay of course), and friends, so that this behavior can be removed? Again, you could easily do this stuff manually: $ cd $SAGE_ROOT/devel/ $ mv sage vanilla $ hg clone vanilla sandbox # or just `cp -r vanilla sandbox` $ ln -s sandbox sage I don't think there's that much difference between the above and the below, to a newbie: $ sage -clone sandbox $ sage -b sandbox -Keshav ---- Join us in #sagemath on irc.freenode.net ! -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
