Re: [boost] Boost regression notification (2006-09-22 [RC_1_34_0])

From: Lars Gullik Bjønnes (larsbj_at_[hidden])
Date: 2006-09-23 15:57:28

David Abrahams <dave_at_[hidden]> writes:

| larsbj_at_[hidden] (Lars Gullik Bjønnes) writes:
| 
| > Douglas Gregor <dgregor_at_[hidden]> writes:
| >
| > | Boost Regression test failures
| > | Report time: 2006-09-22T12:06:02Z
| > | 
| > | This report lists all regression test failures on release platforms.
| > | 
| > | Detailed report:
| > |   http://engineering.meta-comm.com/boost-regression/CVS-RC_1_34_0/developer/issues.html
| > | 
| > | 293 failures in 14 libraries
| >
| > This number seems to just be going up lately. 
| 
| Really?  I've been knocking it down hard, and making steady progress
| on _my_ problems, anyway.   A reconfiguration of test
| machines at OSL caused the entire Boost.Python library to start
| failing on one of their machines yesterday, but the problem has been
| diagnosed.  It's a momentary glitch, I assure you.

It is not the momentary glitches that worry. Perhaps it is the
"buisiness as usual" on the mailinglist and not the all-out focus on
getting the 1.34 release out that worry me most. Note that this is the
feeling I get when I read the mailing lists quite cursory, so it might
very well not be accurate at all. 

This is a list of
failures in regressions report 2006-08/09:

1.  271 failures in 23 libraries
2.  268 failures in 24 libraries
3.  268 failures in 24 libraries
4.  268 failures in 24 libraries
5.  268 failures in 24 libraries
6.  268 failures in 24 libraries
7.  268 failures in 24 libraries
8.  238 failures in 23 libraries
9.  239 failures in 23 libraries
10. 240 failures in 24 libraries
12. 204 failures in 24 libraries
14. 693 failures in 32 libraries (Ok, glitch. No need to worry.)
15. 692 failures in 32 libraries
16. 692 failures in 32 libraries
18. 692 failures in 32 libraries
19. 175 failures in 24 libraries
20. 171 failures in 21 libraries
31. 232 failures in 22 libraries (another small glitch?)

Regressions reports 2006-09:

1.  158 failures in 19 libraries 
2.  149 failures in 19 libraries 
3.  150 failures in 20 libraries (steady decline, good.)
4.  150 failures in 20 libraries 
5.  149 failures in 20 libraries 
6.  149 failures in 19 libraries 
7.  101 failures in 17 libraries (really getting somehwere now.)
8.  100 failures in 17 libraries
9.  106 failures in 18 libraries (... what? going up again?)
10. 107 failures in 18 libraries
11. 109 failures in 18 libraries
12. 161 failures in 17 libraries (Hmm...)
14. 88 failures in 17 libraries  (Sudden drop? Also a glitch?)
15. 209 failures in 20 libraries (Massive increase?)
16. 278 failures in 24 libraries (...and another...?)
17. 288 failures in 19 libraries
18. 238 failures in 18 libraries
19. 220 failures in 19 libraries
22. 293 failures in 14 libraries (Glitch at OSL)

| > Will a 1.34 release ever happen?
| 
| Yes.  You're welcome to pitch in by clearing inspection reports and/or
| regressions if you want to help move things along.

I inspection reports does not worry me much, and imho they are getting
too much attention (seemingly) over the regression reports.

We have at least upgraded our development software to use 1.34 from
cvs, so that we can discovere problems therein before the actual
release.

And personally I try to have a report on platform(s) that concern me
and that I might be able to help with.

-- 
	Lgb