Skip to Content.
Sympa Menu

sm-discuss - [SM-Discuss] stable-rc prometheus run nearing completion

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: seth AT swoolley.homeip.net
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: [SM-Discuss] stable-rc prometheus run nearing completion
  • Date: Thu, 1 Jun 2006 10:47:49 -0700

stable-rc prometheus run nearing completion... film at eleven.

After ten days of watching compiles, I'm about 1/2 done triaging bugs and
prometheus is nearing the end of its run with 90% of the spells done. That
means it will be done in time for the weekend. I will spend most of Saturday
triaging the rest of the bugs and based off input from Arwed, I will be
making a must fix list and a new (much smaller) test list for those fixes.
Anything you have gotten approval to integrate/want to integrate, it
must be fixed over this weekend and the early part of next week. If the
fixes Arwed and I want in are not in, stable release will be delayed, so if
you are an unlucky soul assigned to fix something, please get on it or I will
be forced to fix it myself and we'll remember not to assign you
responsibilities in the future ;)

For those interested, about 10000 compiles were done since each spell was
attempted to compile at least once (unless its dependencies failed on a
previous spell cast).

For the future, now that we have a pretty complete list of "basic" bugs for
the (future) stable grimoire, I'd like to see us work toward fixing
those failures. When stable-rc is re-opened, we'll be a lot more (read:
very) flexible about letting fixes into stable-rc, but for right now, commit
stuff to test and when stable release happens test automatically becomes
stable-rc.

I'll also be running prometheus on that new stable-rc grimoire right away, so
we should have much better warning of failures than this time. When I
get any idle time, I'm going to add in a build-on-commit system for test
grimoire commits -- just to catch the initial problems and act as a simple
limited regression suite. I am also thinking of ways to parallelize
prometheus. The parallel prioritization will probably be to 1) find
different hardware, 2) work on different sections, 3) work on different
grimoires, 4) work on different prompt answers, but I'm undecided as to how
this will be done currently.

Seth




Archive powered by MHonArc 2.6.24.

Top of Page