Has anyone successfully built and used DNG4PS-2 on Linux. I have it built and it seems to convert to uncompressed DNG without an issue but Lightzone, rawstudio, and UFRaw all crash when I try to open the resulting file. DNG4PS-2 just hangs when I attempt to convert to a compressed DNG.
Hello, I have built the latest beta
http://dng4ps2.homeftp.org/beta/dng4ps2-0.2.2beta-25-sources.zip and found the same crashing problem with compressed DNG's. I emailed the author with details about the crash.
And yes, haven't find any program that could open those files. Imagemagick (convert crw_1907.DNG crw_1907.tiff) tells me:
convert: ShadowScale: Rational with zero denominator (num = 1). `crw_1907.DNG'.
The windows binary run under WINE, however, works perfectly, as verified by others too. The resulting DNG's open fine. dcraw -i -v says:
Filename: crw_1907.DNG
Timestamp: Sun Jun 15 09:58:12 2008
Camera: Canon DIGITAL IXUS 860 IS DNG
ISO speed: 80
Shutter: 1/8.0 sec
Aperture: f/2.8
Focal Length: 4 mm
Secondary pixels: no
Embedded ICC profile: no
Decodable with dcraw: yes
Thumb size: 256 x 192
Full size: 3336 x 2480
Image size: 3292 x 2466
Output size: 2466 x 3292
Raw colors: 3
Filter pattern: GBRGGBRGGBRGGBRG
Daylight multipliers: 1.610551 0.986053 1.876544
Camera multipliers: 1.610551 0.986053 1.876544 0.000000
When I do the same for the native linux conversion result, it says:
Filename: crw_1907.DNG
Timestamp: Sun Jun 15 09:58:12 2008
Camera: Canon DIGITAL IXUS 860 IS DNG
ISO speed: 80
Shutter: 0.0 sec
Aperture: f/inf
Focal Length: -2147483648 mm
Secondary pixels: no
Embedded ICC profile: no
Decodable with dcraw: yes
Thumb size: 256 x 192
Full size: 3336 x 2480
Image size: 3336 x 2
Output size: 0 x 3336
Raw colors: 3
Filter pattern: GBRGGBRGGBRGGBRG
Daylight multipliers: nan nan nan
so there is something wrong in the header I quess.
By examining the files with hexdump and diff, it seems that the files are only slightly different. The most differences are just right after the <?xpacket end="w"?> tag. And to me it seems that the native linux version just has the same information, but in slightly different position and with some extra bytes. For example, the wine version has hex sequence: 15 3c 00 00 10 27 00 00 74 ea ff ff 10 27, and linux version has: 15 3c 00 00 00 00 00 00 10 27 00 00 00 00 00 00 74 ea ff ff ff ff ff ff 10 27.
Could it be something to do with 32 bit and 64bit machines, since the windows binary is 32 bit, and my native linux version is built for 64 bit? That could explain why some of the byte sequences are longer.
I'll try to build a 32 bit version and see how it performs.
