RE: Emacs freezes every now and then on my windows 7 pc

[Drew Adams]

> > Each week I hope for a new Windows build that works enough 
> > to check the fix for some bug I filed for a previous build.
> > But in some cases I have not been able to check fixes
> > because of new bugs (including crashes).  And the Windows
> > builds are no longer done weekly, it seems, which aggravates
> > this problem.
> 
> Again, this is expected.

No, this degree of volatility is not expected, if expectation is based on past
experience.  This is not near the mean (or median or mode) of the frequency
distribution, in my guesstimation.

New bugs and disruptions are to be expected, of course.  But this amount of
disruption (at the C level in particular, e.g. crashing) is not typical,
fortunately.

As I think you hinted in emacs-devel, this is more like what everyone might have
experienced if you had developed bidi, from the outset, as the main dish,
instead of as a side dish that later, when more or less perfected, was
incorporated into the main dish.

> I would actually advise people who don't build Emacs themselves
> not to use the development snapshots for anything serious.

Oh, no need to worry.  I wouldn't think of using them for anything serious,
believe me.  (Not that I do anything serious with any build or release.)

And the matter has nothing to do with whether someone builds Emacs personally or
uses a binary built by someone else.  Who builds Emacs is irrelevant here.

And the more frequently prebuilt Windows binaries are provided, the easier and
quicker are the reporting and debugging of bugs.  That's an obvious benefit.
Costs include the effort to build and post the binaries, of course.

> > There seems to be a fair amount of volatility in the C code 
> > at this time....  This is, fortunately, not typical of Emacs
> > development; it is presumably temporary.  Let's hope things
> > settle down soon.
> 
> Sorry, but I'm afraid this is here to stay, at least for a while.
> There are a few contributors to Emacs who tend to make disruptive
> commits at least once a week.  I see no signs that they will stop any
> time soon.  The only time these attitudes can be kept in check is
> during a pretest, or a feature freeze.

On n'arrete pas le progres. ;-)