From: Max L. <max...@ve...> - 2006-07-29 20:21:56
|
I managed to compile the SVN version of the code today for the first time, and found that PTMender doesn't work, and all of the tests fail. All attempts to run PTMender result in a dialog box saying "Could not open _PTStitcher_tmp_0000xx for writing" and then the program exits. I've spent a few hours on it this afternoon and think I've tracked down the problem to the new metadata shuffling code. Specifically, the problem is that in "panoTiffSetImageProperties" (line 1019 in tiff.c), the calls that set the samplesPerPixel and bitsPerSample are trying to use a value of zero, which is what the metadata struct contains. Here's what I see in GDB at this point: Breakpoint 4, panoTiffSetImageProperties (file=0x3f7c90) at tiff.c:1019 1019 returnValue = 1: returnValue = 1 (gdb) p *metadata $26 = {imageWidth = 1168, imageHeight = 411, isCropped = 1, xPixelsPerResolution = 150, yPixelsPerResolution = 150, resolutionUnits = 2, samplesPerPixel = 0, bitsPerSample = 0, bytesPerLine = 0, rowsPerStrip = 1, compression = {type = 32773, predictor = 0}, iccProfile = {size = 0, data = 0x0}, cropInfo = {fullWidth = 1190, fullHeight = 418, croppedWidth = 1168, croppedHeight = 411, xOffset = 0, yOffset = 0}, copyright = 0x0, datetime = 0x0, imageDescription = 0x0, artist = 0x0, bytesPerPixel = 0, bitsPerPixel = 0} (gdb) n 1040 if (returnValue && metadata->bitsPerSample == 32) 1: returnValue = 0 It looks to me like these fields probably ought to have been populated at this point in the code, but they are still zero. This causes PTMender to give up with the error message. Because I'm not familiar with the thinking behind how all of this metadata code should work, I thought I'd post here and see if someone has more familiarity with this and knows how to fix the problem without creating more problems! I've tried this with both JPEG and TIFF files as input, because I thought that the "source" metadata might be different depending on the input image format. The error appears with both image types. Max > > I started renaming the library to pano13. It is amazing in how many > places it is hardcoded :) At least in the Unixes it should be building > pano13 and installing under pano13/ > > In the process I also updated the version to 2.8.5pre1. I am going to > start using -preN suffixes to make sure that we don't have binaries > with the same library number, but different functionality. For > instance, the idea is that we release binaries, and in the next > functional change we update the version to the next number. > > I have tested the library under OS X and Linux and it seems to be > working (in this delta I also killed one bug I recently introduced). > > > > > > -- > Daniel M. German "I'm crazy, but I'm not stupid" > Jackie Chan > http://turingmachine.org/ > http://silvernegative.com/ > dmg (at) uvic (dot) ca > replace (at) with @ and (dot) with . > > > > ------------------------------------------------------------------------- > Take Surveys. Earn Cash. Influence the Future of IT > Join SourceForge.net's Techsay panel and you'll get the chance to share your > opinions on IT & business topics through brief surveys -- and earn cash > http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV > _______________________________________________ > PanoTools-devel mailing list > Pan...@li... > https://lists.sourceforge.net/lists/listinfo/panotools-devel |