This is the mail archive of the cygwin 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]
Other format: [Raw text]

Re: cygheap base mismatch detected


On Thu, Feb 16, 2006 at 01:17:07AM -0600, Dill, Jens (END-CHI) wrote:
>Christopher Faylor wrote:
>> On Wed, Feb 15, 2006 at 10:22:30PM +0000, Jens Dill wrote:
>> >What you write makes it appear that CygWin simply will not support
>> >large executables that reference the CygWin DLL and are also launched
>> >from a CygWin shell.  I can't believe that nobody has found a better
>> >workaround than:
>> >
>> >  (a) not using the CygWin DLL (not an option for us)
>> >  (b) not launching from a CygWin shell (not an option for us)
>> >  (c) reducing the memory space required by the program (again, not
>possible)
>> >  (d) tinkering randomly with the stack size
>> >
>> >Does anyone else know of another way around this problem?
>> 
>> How about:
>> 
>>    (e) Looking into the source code and coming up with a fix rather than
>>        a workaround
>> 
>> ?
>
>Unfortunately for me, (e) is impractical.  It's not clear whether it is
>my source code or CygWin's that I need to fix, but if it were mine, the
>application I am porting to CygWin is really large, and I don't have
>the mandate or time to re-architect it as a bunch of small pieces.  If
>I could do that, I'd rewrite the app for native Windows.  If it's the
>CygWin source, however interested I may be in fixing it, I have no
>mandate from my management to do so (and am not likely to get one).
>Furthermore, the problem appears to be something not directly written
>into the source but rather a property of the CygWin-Windows
>interaction, at a level that I am not likely to understand without
>suddenly learning a lot more Windows internals than I ever thought I
>would need to (Bet you didn't think there were disadvantages to being a
>Unix Guru, did you?).

I think you're missing a point here.

Your knowledge or management limitations will not have any effect whatsoever
on anyone who works on Cygwin.

I'm telling you that your best bet is to fix your problem yourself
because it is unlikely that anyone else is going to do it for you.

I'm not trying to be mean.  I'm just telling you what your options are.

>My project time frame doesn't allow for that.   If I read Dave Korn's
>posting correctly (along with the others who talk about adjusting the
>sizes of Fortran arrays to fix the problem), it becomes pretty clear to
>me that the problem is likely to be one of somehow scaling CygWin to 
>fit with large executables.  Surely someone has run into this before,
>and solved it by some other means than either disabling Cygwin or 
>dismembering their executable.

You can use the word "surely" a few more times but it isn't likely that
you are going to find a magic bullet for your problem.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


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