2011.04.12 18:01 "Re: [Tiff] Bug 1941, Was: Libtiff v4.0.0beta7 released", by Edward Lam

On 4/12/2011 1:05 PM, Lee Howard wrote:

1. http://bugzilla.maptools.org/show_bug.cgi?id=1941

In reviewing that bug report there is nothing concluded for committal. If an agreeable patch appears there it would be committed to CVS.

It's agreeable by me. :) Some of the objections were/are:

1. It changes API semantics for tif_win32.c. I think this fear is unfounded given than we've defaulted to tif_unix.c for at least years now (CVS change was made in Mar 2005). Changing the default to tif_win32.c without the patch is more dangerous to me now because a lot of client apps have been building with tif_unix.c all along.
2. It was feared that older build systems like Visual C++ 6.0 might not work. I noted that this issue should not be a problem in theory because of the MSDN documentation. This point is probably now moot anyhow because people have enough trouble even finding MSVC 2005 these days.
3. In terms of source code compatibility, I offered an approach that should be sufficient to alert library users when compiling libtiff 4 by issuing a #error when USE_WIN32_FILEIO is defined. This is already a lot better than the change to TIFFFindFieldInfo() which silently does nothing.

Best Regards,

-Edward