[Debian-med-packaging] Bug#788559: [Help] Cmake detection problem (Was: Bug#788559: vxl: FTBFS: fatal error: libc.h: No such file or directory)

[Andreas Cadhalpun]

Control: tag -1 patch

Eureka!

On 25.06.2015 11:27, Andreas Tille wrote:
> I've checked this issue and i boils down to the fact that the header
> file in question is included only conditional.  Unfortunately this
> condition is not properly set.  Cmake is supposed to detect an existing
> libc.h header file in dcmtk:
> 
> $ grep -R HAVE_LIBC_H | grep CHECK_INCLUDE_FILE
> v3p/dcmtk/CMakeLists.txt:CHECK_INCLUDE_FILE( libc.h        DCMTK_HAVE_LIBC_H )
> 
> But this does not seem to work properly.  Any idea what might went
> wrong here?

The problem is not limited to DCMTK_HAVE_LIBC_H:
  # The configuration has changed, so reset the stored variables to
  # force them to be recomputed
  #
  FOREACH( var
        DCMTK_DCM_DICT_DEFAULT_PATH
[...]
        DCMTK_HAVE_LIBC_H
[...]
        HAVE_DCMTK_SIZEOF_LONG )  # Created indirectly by CHECK_TYPE_SIZE
    SET( ${var} ${var} )  #  <-- This is the problem!
  ENDFOREACH( var )

It tries to reset the variables, but it sets their value to their name.
Apparently this worked (accidentally?) with previous cmake versions, but
doesn't work anymore.
Instead I think it should just unset the variables.
A patch doing that (and thus fixing the FTBFS) is attached.

Best regards,
Andreas

-------------- next part --------------
A non-text attachment was scrubbed...
Name: vxl_unset-cmake-variables.patch
Type: text/x-diff
Size: 977 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/debian-med-packaging/attachments/20150715/ed051889/attachment.patch>