view release on metacpan or  search on metacpan

src/Source/LibTIFF4/ChangeLog  view on Meta::CPAN

	We also modify logic so that in update mode the directory is not
	marked dirty on read, but only when something is changed.  This
	means we need to keep track of updates to the stripmap stuff in
	TIFFAppendToStrip().

2007-12-10  Frank Warmerdam  <warmerdam@pobox.com>

	* tif_jpeg.c: Improve ability to switch between encoding and decoding
	in the jpeg code (gdal bug #2033).

2007-11-23  Frank Warmerdam  <warmerdam@pobox.com>

	* tif_dir.c, tif_dirread.c, tif_dirwrite.c, tif_read.c, tif_write.c,
	tiffiop.h: Added TIFF_BUF4WRITE flag to indicate if contents of the
	rawcp/rawcc buffer are for writing and thus may require flushing.
	Necessary to distinguish whether they need to be written to disk when
	in mixed read/write mode and doing a mixture of writing followed by
	reading.  http://trac.osgeo.org/gdal/ticket/1758

2007-11-23  Andrey Kiselev  <dron@ak4719.spb.edu>

	* configure.com, libtiff/tif_vms.c: Better OpenVMS support. Patches
	from Alexey Chupahin.

2007-11-02  Frank Warmerdam  <warmerdam@pobox.com>

	* tif_write.c: Rip out the fancy logic in TIFFAppendToStrip() for
	establishing if an existing tile can be rewritten to the same location
	by comparing the current size to all the other blocks in the same
	directory.  This is dangerous in many situations and can easily
	corrupt a file.  (observed in esoteric GDAL situation that's hard to
	document).  This change involves leaving the stripbytecount[] values
	unaltered till TIFFAppendToStrip().  Now we only write a block back
	to the same location it used to be at if the new data is the same
	size or smaller - otherwise we move it to the end of file.

	* tif_dirwrite.c: Try to avoid writing out a full readbuffer of tile
	data when writing the directory just because we have BEENWRITING at
	some point in the past.  This was causing odd junk to be written out
	in a tile of data when a single tile had an interleaving of reading
	and writing with reading last.  (highlighted by gdal
	autotest/gcore/tif_write.py test 7.

	* tif_predict.c: use working buffer in PredictorEncodeTile to avoid
	modifying callers buffer.
	http://trac.osgeo.org/gdal/ticket/1965

	* tif_predict.c/h: more fixes related to last item, keeping a
	distinct pfunc for encode and decode cases as these were getting
	mixed up sometimes.
	http://trac.osgeo.org/gdal/ticket/1948

2007-11-01  Frank Warmerdam  <warmerdam@pobox.com>

	* tif_predict.c/h, tif_lzw.c, tif_zip.c: Improvements so that
	predictor based encoding and decoding works in read-write update
	mode properly.
	http://trac.osgeo.org/gdal/ticket/1948

2007-10-24  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_dirread.c: Fixed problem with bogus file triggering
	assert(td->td_planarconfig == PLANARCONFIG_CONTIG) in
	ChopUpSingleUncompressedStrip

2007-10-22  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_jpeg.c: Resolved buffer incrementation bug that lead to faulty images
	at best, access violation at worst, when subsampled JPEG compressed imagery
	is decoded without the JPEG_COLORMODE feature

2007-10-11  Frank Warmerdam  <warmerdam@pobox.com>

	* html/index.html: Update "people responsible" section.

2007-10-05  Frank Warmerdam  <warmerdam@pobox.com>

	* tools/tiff2pdf.c: Fix problem with alpha setting in some cases
	as reported on the mailing list.

2007-10-01  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* changed some more incorrect %lud printf flags to %lu

2007-09-29  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_dirread.c: Strip chopping interfered badly with uncompressed
	subsampled images because it tried to divide subsampled rowblocks,
	leading to all sorts of errors throughout the library for these
	images. Fixed by making strip chopping divide in row counts that
	are a multiple of vertical subsampling value.

2007-09-28  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_dirread.c: Logical cast working around compiler warning

	* tif_read.c: Correction of some error flags and parameter lists

2007-09-27  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_dirread.c: Made calculation of td_maxsamplevalue more robust
	when dealing with large bitspersample values, shutting up purification
	tools that warn about truncation, though it remains incorrect and
	indicates a conceptual problem there.

	* tif_open.c: Moved early exit in case of 'h' flag (to disable reading
	of first IFD) to proper place because it badly interfered with memory
	mapping, resulting in mapping flag even with dummy mapping functions
	that returned 0 whilst at the same time the mapping tif_size wasn't
	set, thus resulting in continuous incorrect beyond-eof errors.

2007-09-24  Joris Van Damme  <joris.at.lebbeke@skynet.be>

	* tif_dirinfo.c: Fixed (MSVC) compiler reports about
	inconsistent use of const in tiffFields and exifFields definition

2007-09-20  Frank Warmerdam  <warmerdam@pobox.com>

	* tif_dirwrite.c: Always write tile/strip offsets and sizes
	using LONG8 type when output format is BigTIFF.  The