Bug: bfconvert fails to write valid TIFF header for some planes

[rleigh]

See: QA 9412

Run:

% bfconvert -channel 0 in.czi o_%z.tif


% ls -ltr o_*
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_0.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_1.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_2.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_3.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_4.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_5.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_6.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_7.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_8.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_9.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_10.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_11.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_12.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_13.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_14.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_15.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_16.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_17.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_18.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_19.tif
-rw-rw-r-- 1 rleigh rleigh 427258 Aug 19 10:43 o_20.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_21.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_22.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_23.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_24.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_25.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_26.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_27.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_28.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_29.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_30.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_31.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_32.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_33.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_34.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_35.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_36.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_37.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_38.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_39.tif
-rw-rw-r-- 1 rleigh rleigh 427250 Aug 19 10:43 o_40.tif

% ls -1tr o_* | xargs file
o_0.tif:  TIFF image data, little-endian
o_1.tif:  TIFF image data, little-endian
o_2.tif:  TIFF image data, little-endian
o_3.tif:  TIFF image data, little-endian
o_4.tif:  TIFF image data, little-endian
o_5.tif:  TIFF image data, little-endian
o_6.tif:  TIFF image data, little-endian
o_7.tif:  TIFF image data, little-endian
o_8.tif:  TIFF image data, little-endian
o_9.tif:  TIFF image data, little-endian
o_10.tif: TIFF image data, little-endian
o_11.tif: TIFF image data, little-endian
o_12.tif: TIFF image data, little-endian
o_13.tif: TIFF image data, little-endian
o_14.tif: TIFF image data, little-endian
o_15.tif: TIFF image data, little-endian
o_16.tif: TIFF image data, little-endian
o_17.tif: TIFF image data, little-endian
o_18.tif: TIFF image data, little-endian
o_19.tif: TIFF image data, little-endian
o_20.tif: TIFF image data, little-endian
o_21.tif: data
o_22.tif: data
o_23.tif: data
o_24.tif: data
o_25.tif: data
o_26.tif: data
o_27.tif: data
o_28.tif: data
o_29.tif: data
o_30.tif: data
o_31.tif: data
o_32.tif: data
o_33.tif: data
o_34.tif: data
o_35.tif: data
o_36.tif: data
o_37.tif: data
o_38.tif: data
o_39.tif: data
o_40.tif: data

Looking at the data in the above files, the last 20 are missing the initial 8 byte TIFF header; all bytes after that look OK. I'm unsure what's making bfconvert miss the header just for these planes, but that looks like the root cause of the problem here.

[mlinkert]

NB: the correct QA is 9452, not 9412.

I could duplicate this on develop alone, but rebuilding with ​https://github.com/openmicroscopy/bioformats/pull/1555 merged in solved the problem.