This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Undefined reference to '_ctype_'?


On Sat, Feb 26, 2000 at 04:06:30AM -0500, Glenn Spell wrote:
>On 25 Feb 2000 around  2:51PM (-0800) Kendall Bennett wrote:
>>Or is there a stable, later snapshot that I can install and 
>>use effectively
>
>In my opinion, the most stable and the best-of-the-best all-time
>Cygwin package was the 19991204 development snapshot. I'd like
>to see that included with the new Net Release.
>
>bash-2.03$ uname -a
>CYGWIN_95-4.0 GS.FAY.NC.US   22.0(0.16/3/2) 1999-12-5 00:16:58 i586 unknown
>
>Unfortunately, it's not presently available.
>
>Someone (Earnie ?) mentioned that we might want keep that one like
>DJ did with the 19990115. Someone else stated that yes we might want
>to do that but not yet since keeping it would detract from checking
>out the later packages (apparently because 19991204 was so highly
>regarded and because development was changing focus from fixing bugs
>to something else) and concluded with the observation that besides,
>we still had plenty of time to make the decision about keeping it.

We are not going to have a >3 month old snapshot in the net release.

Your assessment of what happened in the mid-December time frame is
inaccurate.  What happened in mid-December is that I came up with a new
method for delivering signals in cygwin that has the potential of making
cygwin faster and more robust -- especially under Windows 95/98.

I spent considerable time retooling the signal code to use this new
method.  For the most part, it has been a success.  There are bugs to be
worked out, as I knew there would be.  Although it seemed to me that
most of the problems had been worked out by January.

I wanted people to try the new snapshots to work the bugs out.  I had
not interest in keeping a "stable" snapshot around because then people
would 1) not use the succeeding snapshots and 2) start reporting
problems in the "stable snapshot" that were fixed in modern versions.

Lately, the code has been through a second wave of modifications and is
relatively unstable as a result.

This week, I have spent considerable time trying to track down the
various "bash is taking all of my CPU" problems despite the almost total
lack of debugging information offered by any of the people reporting the
problem.

>Actually, what might be useful is a branch in development. One
>branch of "stable" and another branch of "cutting edge". The
>stable branch could be based on 19991204.

Sorry.  I have no interest in maintaining two separate branches.

>Perhaps Redhat would consider an increase in resource allocation
>for the Cygwin Project.

That's an interesting observation.  Why would Red Hat (two words) be
interested in doing this?  What would the financial motivation be?

And, what problem would you be attempting to solve here?  Are you
assuming that with more Cygwin engineers on staff we'll have more people
to answer the 525th repetition of the "What is this _ctype_ thing?" It's
extremely unlikely that, if Red Hat did hire more staff they would be
dedicated to answering questions in the mailing list.

It's also not likely that Red Hat would hire someone to do Cygwin net
releases.  Again, there is no financial justification to do this.

For now, you're stuck with the volunteer efforts of myself, DJ, Corinna,
Mumit, Sergey, Ron Parker, and a few others.  I don't expect that that
will change any time soon, although, FWIW, you'll soon be seeing a
"@cygnus.com" appended to Corinna's email address.

cgf

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]