[gdal-dev] assert failure (tif_open)

[Nik Sands]

Thanks again for your help Frank.

As usual, you were spot on and adding the following line to .../port/cpl_config.h.in before building fixed the problem:

#define HAVE_LONG_LONG 1

I've now added this to my own build notes.

Since the problem didn't occur when using the beta versions of Apple's Xcode 5 and only occurred with the final version of Xcode 5, I'm assuming that Apple changed their developer tools/environment with their production release which coincided with their announcement of 64-bit iPhones (iPhone 5S) in such a way that it affected what 'configure' wrote to cpl_config.h.

Cheers,
Nik.


On 27/09/2013, at 9:54 AM, Frank Warmerdam <warmerdam at pobox.com> wrote:

> Nik,
> 
> I assume you are using the "built in" libtiff in GDAL.  In that case the file gdal/frmts/gtiff/libtiff/tif_config.h will include cpl_port.h and includes this line:
> 
> #define TIFF_UINT64_T GUIntBig
> 
> I assume TIFF_UINT64 ultimately is used to define uint64.
> 
> So I think you need to go through gdal/port/cpl_port.h and gdal/port/cpl_config.h to try and figure out how GUIntBig is getting defined.  In cpl_port.h it helps to search for the block of code titled "64bit support".  I would guess you want to use "unsigned long long" in which case you would need to have HAVE_LONG_LONG defined in cpl_config.h.
> 
> I'm not aware of any recent changes in this logic so I'm not sure why it used to work and doesn't now.  
> 
> Best regards,
> Frank
> 
> 
> 
> 
> 
> 
> On Thu, Sep 26, 2013 at 4:33 PM, Nik Sands <nixanz at nixanz.com> wrote:
> Hi list members,
> 
> I've been using GDAL 1.10 as a statically linked library in an iOS app for some time.  I recently recompiled GDAL again from the same local set of source files, using my notes on configuration options that I used last time, and ever since I've been having a problem every time the app attempts to open an image file.
> 
> The app fails an assert() in the included tiff library (in tif_open.c), throwing the error:
> 
>         Assertion failed: (sizeof(uint64)==8), function TIFFClientOpen, file tif_open.c, line 99.
> 
> The line in question is the last line of the code snippet from tif_open.c below:
> 
>         /* The following are configuration checks. They should be redundant, but should not
>          * compile to any actual code in an optimised release build anyway. If any of them
>          * fail, (makefile-based or other) configuration is not correct */
>         assert(sizeof(uint8)==1);
>         assert(sizeof(int8)==1);
>         assert(sizeof(uint16)==2);
>         assert(sizeof(int16)==2);
>         assert(sizeof(uint32)==4);
>         assert(sizeof(int32)==4);
>         assert(sizeof(uint64)==8);  //  <- SIBABRT HERE
> 
> The comments indicate that there may be a problem with the "configure" options I used when building GDAL, however I'm using exactly the same options (copy and pasted) as I used last time when it worked OK.
> 
> So this leads me to think that perhaps something has changed in my environment, and the only thing I can think of is that I may be using a different version of GCC (I'm using the 'gcc' included with Apple's 'Xcode' IDE, and since the last compile of GDAL, I've upgraded from a beta build of Xcode to the production build of Xcode 5).
> 
> I've tried it with GDAL 1.10.1 as well and get exactly the same problem.
> 
> I'm stumped as to how to proceed with resolving this issue.  Can anybody help me to overcome this?
> 
> Note that the problem occurs when running on the real iOS device (armv7s) and on the iOS simulator (i386).
> 
> Cheers,
> Nik.
> 
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev
> 
> 
> 
> -- 
> ---------------------------------------+--------------------------------------
> I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
> light and sound - activate the windows | http://pobox.com/~warmerdam
> and watch the world go round - Rush    | Geospatial Software Developer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20130927/842f1e8b/attachment-0001.html>