nanogui: Thread: "stub.c" question (nxlib-0.44)


[<<] [<] Page 1 of 1 [>] [>>]
Subject: "stub.c" question (nxlib-0.44)
From: "Paul Bartholomew" ####@####.####
Date: 28 Jun 2005 01:47:44 +0100
Message-Id: <BAY108-F28CD7208FB2B40B46178E2EFE10@phx.gbl>

Hello -

I've got microwindows-0.90/nano-X running under X-Windows.  I'm building 
under Fedora Core 4 (includes gcc 4.0, which has introduced a few problems 
(which I've 'hacked' to resolve)).

The microwindows/nano-X demos seem to run OK.

I'm trying to get fltk-1.1.6 running using nxlib-0.44.  Running their 
"test/hello" or "test/clock" apps (after starting the "nano-X" server), I 
get core dumps.  I've found that "XAllocSizeHints()" and "XAllocWMHints()" 
are returning NULL pointers (which the code in "Fl_x.cxx" doesn't like).

I see that these are in nxlib's "stubs.c" (returning "0" - I assume this 
means that they are 'not implemented'??).

I modified Fl_x.cxx to use "calloc()" instead of "XAllocSize/WMHints()", and 
it seems to work.

I assume I could just modify 'stubs.c' to do the same.

But, is this the 'correct' way to do things in the "X" client/server model?  
On a 'real' X system, would the allocation occur on the server-side 
(meaning: I shouldn't do an allocation on the client-side to workaround the 
problem)?  If that's the case on a 'real' X setup, is that also true for the 
nano-X client/server model?

Or, is it OK to just use 'calloc()' for this type of allocation?

Thanks in advance,

- Paul B.


[<<] [<] Page 1 of 1 [>] [>>]


Powered by ezmlm-browse 0.20.