This is the mail archive of the cygwin-developers@cygwin.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: "make -j2" hang is not fixed in current repository


To clarify my last email:

- What I tried to convey is that recent events have made it less than
  likely that I will be fixing major bugs that require immense amounts of
  time to track down.

- I also expressed dismay that the bug-report/bug-fix ratio is not
  very close to 1 on the cygwin-developers mailing list.

- I provided a pointer to my original message so that you could refresh
  yourself with whatever vague explanation I gave about how I thought
  I'd originally fixed the problem.

- I vented on how much I hate getting stack traces in email.  I've
  just told Corinna something similar in personal email, in fact.

- And, finally, I was stated the obvious.  It has been more than a
  month since I suggested that this was fixed.  I'm not completely fresh
  on what I did so I can't offer the insight now that I could have
  potentially offered then.

Those were my five points.

You *are* *obviously* trying to fix bugs.  I am sure that I am sending
you mixed messages when I fix the bugs that you are working on.  I try
to explain what I've done when I do fix the bugs so that it is obvious
why I did what I did.  However, I am sure that it must be frustrating,
regardless.

(I actually went through a similar scenario with Kazuhiro.  I felt bad
because he often sent in fixes which I thought were not quite right.  I
ended up redoing things more to "my liking'.

Strangely, I haven't had to do that much lately.  I think that is a
sign that this crude feedback relationship worked.)

Your extremely detailed bug reports are appreciated.  This last one,
however, had more raw information than I am prepared to digest.  It was
200+ lines, the majority of which was stack trace.

You state that you are not qualified to fix this problem but I have to
differ with you.  You are obviously a *very* sharp person.  I really
don't believe that the cygwin source code is beyond you.  Your recent
analyses of problems have been impressive.

However, you may not have time to fully grok and debug this particular
problem and, as I have said, that is understandable.

The bottom line is that maybe this is just a bug that is not going to
get fixed.  I've noted this before.  It's a shame because I thought I
actually had finally squashed the problem after your last bug report but
I guess that I was overly optimistic.

Although, I'm now on iteration 53 of a full -j2 build of the winsup
directory without any pauses.  That particular test case couldn't get
much beyond 17 on my (733MHZ SMP W2K) system before.  So, I've at
least ameliorated *a* problem.  Apparently it wasn't your problem.

cgf


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