fix: use opj_uint_ceildiv instead of opj_int_ceildiv when necessary

[mayeut]

There are a bunch of loc where we can see a usage of opj_int_ceildiv:

(OPJ_UINT32)opj_int_ceildiv((OPJ_INT32)a, (OPJ_INT32)b);

where a & b are OPJ_UINT32.

This can lead to overflow/underflow for some a/b combinations.
Replace those calls by opj_uint_ceildiv instead to always get a correct result.

This also allows some valid single tile images with huge tile size to be decoded properly.
Fix #1438

Requires data from uclouvain/openjpeg-data#24

[rouault]

can you git commit --amend & git push -f to force CI to restart now that uclouvain/openjpeg-data#24 has been merged ?

[mayeut]

I'm seeing the same tests failures for "regular" on master in my fork: https://github.com/mayeut/openjpeg/actions/runs/6278478237/job/17052321449

Not sure how to go forward.

[rouault]

Not sure how to go forward.

maybe try renaming the existing tools/travis-ci/knownfailures-Ubuntu22.04-gcc11.3.0-x86_64-Release-3rdP.txt file as tools/travis-ci/knownfailures-Ubuntu22.04-gcc11.4.0-x86_64-Release-3rdP.txt since there's appear to have been an upgrade from gcc 11.3 to 11.4

[Jamaika1]

Why do I need to use OPJ_HAVE_INTTYPES_H for GCC Windows?

j2k.c: In function 'opj_j2k_dump_MH_info':
j2k.c:11341:35: error: expected ')' before 'PRIu32'
11341 |     fprintf(out_stream, "\t tx0=%" PRIu32 ", ty0=%" PRIu32 "\n", p_j2k->m_cp.tx0,
      |            ~                      ^~~~~~~
      |                                   )
j2k.c:45:1: note: 'PRIu32' is defined in header '<inttypes.h>'; did you forget to '#include <inttypes.h>'?
   44 | #include "opj_includes.h"
  +++ |+#include <inttypes.h>
   45 |
j2k.c:11343:35: error: expected ')' before 'PRIu32'
11343 |     fprintf(out_stream, "\t tdx=%" PRIu32 ", tdy=%" PRIu32 "\n", p_j2k->m_cp.tdx,
      |            ~                      ^~~~~~~
      |                                   )
j2k.c:11343:36: note: 'PRIu32' is defined in header '<inttypes.h>'; did you forget to '#include <inttypes.h>'?
11343 |     fprintf(out_stream, "\t tdx=%" PRIu32 ", tdy=%" PRIu32 "\n", p_j2k->m_cp.tdx,
      |                                    ^~~~~~
j2k.c:11345:34: error: expected ')' before 'PRIu32'
11345 |     fprintf(out_stream, "\t tw=%" PRIu32 ", th=%" PRIu32 "\n", p_j2k->m_cp.tw,
      |            ~                     ^~~~~~~
      |                                  )
j2k.c:11345:35: note: 'PRIu32' is defined in header '<inttypes.h>'; did you forget to '#include <inttypes.h>'?
11345 |     fprintf(out_stream, "\t tw=%" PRIu32 ", th=%" PRIu32 "\n", p_j2k->m_cp.tw,
      |                                   ^~~~~~

[rouault]

@Jamaika1 does #1483 do the job ?

[mayeut]

@rouault,

with #1450 (which goes with "Remove support for non-C99 compilers (like VS2010) that don't support snprintf()"), support for MSVC versions prior to vs2015 is dropped (https://stackoverflow.com/questions/2915672/snprintf-and-visual-studio-2010).

This means that all supported MSVC versions do have stdint.h & inttypes.h now.
For non windows platforms, those headers were mandatory.

I can either add a definition for PRIu32 in the current opj_inttypes.h or just make those headers mandatory altogether.

I'd rather go with the second option.

Any thoughts ?

[rouault]

I'd rather go with the second option.

ah yes, but @Jamaika1 how come inttypes.h isn't detected on your gcc windows ? Are you using an antiquated version ?

[mayeut]

@Jamaika1, were you defining -DOPJ_HAVE_INTTYPES_H=0 in your cmake configure step rather than letting cmake find/define those ?

[Jamaika1]

I'm amateur who decided to manually assemble ffmpeg in gcc. There are lot of additions and bugs.
https://github.com/Jamaika1/plugins_ffmpeg_vvc_evc_htj2k
I use c/c++11.
I have this record:
As I understand it, in mingw64 (no POSIX) I should not use inttypes.h, sys/types.h, sys/types.h, sys/stat.h

#if (defined(__STDC__) && __STDC__ && __STDC_VERSION__ >= 199901L) || (defined(__GNUC__) && (defined(_STDINT_H) || defined(_STDINT_H_)) || defined (HAVE_STDINT_H))
#include <stdint.h>

   typedef int16_t celt_int16;
   typedef uint16_t celt_uint16;
   typedef int32_t celt_int32;
   typedef uint32_t celt_uint32;
#elif defined(_WIN32) 

#  if defined(__CYGWIN__)
#    include <_G_config.h>
     typedef _G_int32_t celt_int32;
     typedef _G_uint32_t celt_uint32;
     typedef _G_int16 celt_int16;
     typedef _G_uint16 celt_uint16;
#  elif defined(__MINGW32__)
     typedef short celt_int16;
     typedef unsigned short celt_uint16;
     typedef int celt_int32;
     typedef unsigned int celt_uint32;
#  elif defined(__MWERKS__)
     typedef int celt_int32;
     typedef unsigned int celt_uint32;
     typedef short celt_int16;
     typedef unsigned short celt_uint16;
#  else
     /* MSVC/Borland */
     typedef __int32 celt_int32;
     typedef unsigned __int32 celt_uint32;
     typedef __int16 celt_int16;
     typedef unsigned __int16 celt_uint16;
#  endif

#elif defined(__MACOS__)

#  include <sys/types.h>
   typedef SInt16 celt_int16;
   typedef UInt16 celt_uint16;
   typedef SInt32 celt_int32;
   typedef UInt32 celt_uint32;

#elif (defined(__APPLE__) && defined(__MACH__)) /* MacOS X Framework build */

#  include <sys/types.h>
   typedef int16_t celt_int16;
   typedef u_int16_t celt_uint16;
   typedef int32_t celt_int32;
   typedef u_int32_t celt_uint32;

#elif defined(__BEOS__)

   /* Be */
#  include <inttypes.h>
   typedef int16 celt_int16;
   typedef u_int16 celt_uint16;
   typedef int32_t celt_int32;
   typedef u_int32_t celt_uint32;

#else 

[rouault]

in mingw64 (no POSIX) I should not use inttypes.h

inttypes.h is a standard headers that all C99 and C++11 compatible compilers (including mingw64) should provide