20882 - autoconf test for D_INO in pathsub.c

Status: RESOLVED FIXED

(SeaMonkey :: Build Config, defect, P3)

Description

[Bruce Mitchener]

in pathsub.c, there is an #ifdef for QNX to use a different D_INO value.  But
that is then special cased for Neutrino to not use that.  Sounds like a good
candidate for an autoconf test.

Comment 1

[Jon Granrose]

reassigning to briano as I believe he has a QNX system to test this on.

Comment 2

[Brian Ostrom]

Good idea.  I need to do the same kind of thing for va_list (array or not),
so I'll just deal with both at the same time.

Comment 3

[Brian Ostrom]

From what I've been able to determine, only QNX uses dirent.d_stat.st_ino
instead of the normal dirent.d_ino, so I'm wondering if a test is really
necessary.  Instead, I could simply define D_INO="d_stat.st_ino" for QNX4
and D_INO="d_ino" for everything else and add it to DEFINES in configure.
What do you think?  If an autoconf test _is_ appropriate, any suggestions as
to how it should work?  Everything I've thought of so far has required two
new macro definitions: HAVE_D_INO and D_INO='whatever'.  Thoughts?

Comment 4

[Jerry.Kirk]

Well if only QNX4 is the only oddball... then I think that just doing it
in configure without the test is good enough.

Comment 5

[Brian Ostrom]

I checked in a change to configure.in to set D_INO appropriately, and to
check if va_list is an array.  The code changes relating to this will go
in next week or after M12, whichever comes first.

Comment 6

[Brian Ostrom]

The complete fix is now checked in (both the D_INO and va_list) fixes.
Enjoy.