AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2008.07.28 08:49 "[Tiff] CVS access", by Mateusz Łoskot
2008.07.28 13:01 "[Fwd: Re: [Tiff] CVS access]", by Edward Lam
2008.07.28 14:02 "Re: [Tiff] CVS access", by Mateusz Loskot
2008.07.28 23:17 "Re: [Tiff] CVS access", by Ryan Schmidt
2008.07.28 23:57 "Re: [Tiff] CVS access", by Frank Warmerdam
2008.07.29 16:37 "Re: [Tiff] CVS access", by Gary McGath
2008.07.29 17:43 "Re: [Tiff] CVS access", by Graeme Gill
2008.07.29 20:17 "Re: [Tiff] CVS access", by Gene Amtower
2008.07.29 21:01 "Re: [Tiff] CVS access", by Bob Friesenhahn
2008.07.29 22:04 "Re: [Tiff] CVS access", by Gene Amtower
2008.07.30 09:32 "Re: [Tiff] CVS access", by Andrew Brooks
2008.08.11 20:55 "Re: [Tiff] windows 64 bit build", by Bob Friesenhahn
2008.08.11 19:35 "Re: [Tiff] windows 64 bit build", by Mikhail Kruk
2008.08.11 17:57 "[Tiff] windows 64 bit build", by Mikhail Kruk
2008.08.11 19:29 "Re: [Tiff] windows 64 bit build", by Bob Friesenhahn
2008.08.12 00:41 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.11 20:03 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.11 20:51 "Re: [Tiff] windows 64 bit build", by Bob Friesenhahn
2008.08.12 00:36 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.12 02:44 "Re: [Tiff] windows 64 bit build", by Bob Friesenhahn
2008.08.12 03:53 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.12 04:04 "Re: [Tiff] windows 64 bit build", by Mikhail Kruk
2008.08.12 12:54 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.12 04:47 "Re: [Tiff] windows 64 bit build", by Bob Friesenhahn
2008.08.12 13:04 "Re: [Tiff] windows 64 bit build", by Edward Lam
2008.08.13 04:23 "[Tiff] tif_win32.c patch proposal (was: windows 64 bit build)", by Edward Lam
2008.08.13 05:32 "[Tiff] Re: tif_win32.c patch proposal (was: windows 64 bit build)", by Bob Friesenhahn
2008.09.04 14:12 "[Tiff] Re: tif_win32.c patch proposal", by Edward Lam

2008.08.13 05:32 "[Tiff] Re: tif_win32.c patch proposal (was: windows 64 bit build)", by Bob Friesenhahn

If you can submit a tested patch, I will be happy to commit it.

Ok, here's a patch attached for discussion. The approach I previously mentioned ran into a slight snag. Once you've created an fd for a HANDLE, the C runtime library (CRT) takes ownership of it. Thus, when you call _close(fd), it closes the associated HANDLE as well. Plus, in _tiffCloseProc(), there was no way (that I found) to query the CRT to return the associated fd given the HANDLE. So in the end, I did something more drastic, albeit cleaner.

It seems that this approach is drastic enough that we should wait for some feedback from potentially impacted users before moving ahead with it.

Being more from the Unix mind-set than the Windows mind-set, I like your approach better than the approach which is currently used.

Bob
======================================
Bob Friesenhahn
bfriesen@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer, http://www.GraphicsMagick.org/