Okay, a couple weeks back, right after we released TG 2.2.0, we
discovered a few things about the sourceforge system that makes it no
longer desirable for using as our tracker. We had been thinking of
using Github back when we switched last year, but (at the time) we
chose SF.
Now, we're going to make the switch to Github. I'm in the process of
writing a migration tool that will copy all the pieces over for us.
One question, though, has arisen, and this change is both significant
and inconsequential. All of our previous systems have used one issue
tracker for the entire project. Github uses one tracker per
repository, and we have three repositories (core, docs, and
tutorials). This means that I have to do something different for
copying ticket information, and I want to ask for feedback before I do
it. Here's the options:
1. Copy every ticket into every repository's issue tracker. Good:
Ticket numbers are preserved. Bad: Tickets are fully duplicated, which
could make it difficult for people to track the tickets they genuinely
care about. Also means we have to close a bunch of them
inappropriately (a ticket applies only to docs, so it has be closed on
core and on tutorials).
2. Copy only relevant tickets, but use placeholders for non-relevant
tickets. Good: Ticket numbers are preserved. Bad: Lots of tickets in
each repository that will read "Placeholder. Please ignore."
3. Copy only relevant tickets. Good: Each repository's issues are kept
to being just the relevant items. Bad: Ticket numbers are not
preserved.
I'm leaning towards number three, personally. However, I know that
there are some people who will prefer (or even demand) one of the
others, so as to preserve ticket numbering.
What do you think? Which one would you prefer?
--
Michael J. Pedersen
My Online Resume:
http://www.icelus.org/ -- Google+
http://plus.ly/pedersen
Google Talk:
m.ped...@icelus.org -- Twitter: pedersentg