gl_MODULE_INDICATOR

[Eric Blake]

Right now, we have:

# gl_MODULE_INDICATOR([modulename])
# defines a C macro indicating the presence of the given module
# in a location where it can be used.
#                                             |  Value  |   Value   |
#                                             | in lib/ | in tests/ |
# --------------------------------------------+---------+-----------+
# Module present among main modules:          |    1    |     1     |
# --------------------------------------------+---------+-----------+
# Module present among tests-related modules: |    0    |     1     |
# --------------------------------------------+---------+-----------+
# Module not present at all:                  |    0    |     0     |
# --------------------------------------------+---------+-----------+

But I have a condition (namely, test-fclose), where I only want to test
a particular behavior of fclose if I know that fflush was replaced at
the same level as fclose.  That is, if fclose is in lib, but fflush is
only in tests, then the test is likely to fail and so I want to compile
it out.  So here's what I'm thinking:

# gl_MODULE_INDICATOR([modulename])
# defines a C macro indicating the presence of the given module
# in a location where it can be used.
#                                             |  Value  |   Value   |
#                                             | in lib/ | in tests/ |
# --------------------------------------------+---------+-----------+
# Module present among main modules:          |    1    |     2     |
# --------------------------------------------+---------+-----------+
# Module present among tests-related modules: |    0    |     1     |
# --------------------------------------------+---------+-----------+
# Module not present at all:                  |    0    |     0     |
# --------------------------------------------+---------+-----------+

then using:

#if GNULIB_FFLUSH == GNULIB_FCLOSE
  test behavior of fclose on seekable input stream
#endif

which works if both modules are in lib, or both modules are in tests,
but properly skips in libvirt's case where the use of 'getpass' and
'close' pulls fclose but not fflush into lib.

Most users just check for non-zero value, and will not be impacted; but
with the extra information of lib vs. tests, we can do nicer conditional
compilation.  Sound reasonable?  I'll be posting a full RFC patch
shortly, but won't push it until I have feedback.

-- 
Eric Blake   address@hidden    +1-801-349-2682
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature