Hadn't known of git-trac before. Since the tool has some sage settings 
hard-wired into it, perhaps it should be called git-sage instead, to avoid 
clashes?

Recently I've been using git the hard way for pretty much everything except 
pushing, for which I used "./sage -dev push --ticket" in the hope that that 
will take care of setting any fields on the Trac ticket that need setting. 
I had been using the dev scripts more at first, but the mapping between 
tickets and branch names was a real pain, so anything which can simply 
deduce the ticket number from the branch name sounds like a good thing to 
have. Seems I'll have to learn a new way now, but the change seems simple 
enough.

Reading the documentation now for the first time, some things are not 
perfectly clear. For example, when creating a ticket, where does the ticket 
description come from? The printed code snippet doesn't show it, and the 
text doesn't mention it. The text also doesn't say where the content of the 
remote branch comes from. So should I do this step at a point where I 
already have some modifications locally, or can I do this at any time? Does 
it matter what branch I'm on locally? What's the point of creating a remote 
branch if I don't have any code at this point? The note about how I could 
also open a ticket using my web browser is already in the next section, 
which is confusing.

The “Finishing Up” section isn't perfectly clear either. It mentions “the 
trac server”, but that could be the git server associated with trac or the 
trac web interface as opened in a browser. I assume the latter. In which 
case I'd add my name into the “Authors” field accessible under the “Modify 
Ticket” section, just as the ticket status is. The note about “first line” 
is confusing to me. The “Merging” section gives special consideration to 
the “master” branch, but shouldn't the same consideration also apply to the 
“develop” branch? Perhaps that section should also discuss the dependencies 
field for Trac tickets, since when one ticket merges code from another, 
you'd usually want it to depend on that other as well. You might also want 
to offer some guideline about the converse: if some code depends on some 
other code, should you merge branches even though doing so will add code to 
the diff? In #16571 <http://trac.sagemath.org/ticket/16571> I decided not 
to do so, but I had been reading the docs (for the dev scripts) searching 
for some guideline on this.

So if you want to rewrite the documentation some more, you might want to 
address some of these things.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to