solaris status

[Phil Edwards <pedwards at disaster dot jaj dot com>]

For the record, I've been hacking at building under Solaris.  It's not
pretty, and I can't even post a useful patch yet.  The build is still
failing.  (It just gets farther before dying, which is still hopeful.)

All of the problems so far have derived from wide character stuff.  Chunks of
the libio have conditional declarations like:

    #if defined _LIBC || defined _GLIBCPP_USE_WCHAR_T
     correct wchar stuff
    #endif

or

    #if defined _LIBC || defined _GLIBCPP_USE_WCHAR_T
     correct wchar stuff
    #else
     not-so-correct workarounds
    #endif

If _GLIBCPP_USE_WCHAR_T is undefined, then the missing/incorrect declarations
cause problems later on (e.g., __c_wfile_type, __conv_type).  I've been
kludging workarounds just by defining empty structs and a lot of null
pointers, but there's a couple points I haven't been able to work out yet.

I'm thinking two things:  first, more of the wide-character-using stuff
should be conditionally compiled, based on _GLIBCPP_USE_WCHAR_T.  Second, to
aid in porting, might we add --disable-wchar to configure?  That would
disable _GLIBCPP_USE_WCHAR_T even if the necessary support functions were
present.

But maybe that isn't the best way to go?


Phil


-- 
pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.