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: WriteFile() whacks st_atime patch


----- Original Message -----
From: "Christopher Faylor" <cgf@redhat.com>
I actually proposed the link option since it fixes the problem at the
> source.  If you link mutt with setftime.o you will not be constantly
> saying "Didn't you read the FAQ, you idiot?  You have to set
> CYGWIN=ftime!" mutt will just work.
>
> The issue of mutt not working correctly has been a hot mailing list
> topic.  However, I don't think that the issue of atime not being set
the
> same way UNIX sets it has really been of generic enough interest to
> warrant YA CYGWIN variable.

I concur, but with slightly different reasoning.

The overhead of the fix will seriously impact anything trying performing
lots of writes, whether or not it cares about atime. I.e. a newsspool or
squid. Installing mutt shouldn't be a cause to cripple all the other
software on the system.

So _any_ solution needs to be specific, or low overhead. Mutt and
related packages can have makefile changes propogated upstream, or at a
minimum they will be in the source package here, and anyone porting new
software across that interacts with the files mutt cares about *will
have trouble anyway*. ((binmode|automode|textmode).o | setmode()) at a
minimum.

Rob


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