On Sun, 1 Feb 2004, Riccardo wrote:

Hello,

since upgrading to 1.2.10 did not help, I pathed manually every file including gtk.h by doing #undef func_data before.

- I notice I get 'stcasecmp' implicit declaration warning.

Fixed in CVS.

- the vsnprintf was already reported and shouldn't be a matter.

Yes.

at the final linking inside "src" I get errors about duplicate symbols (.strcmp, strcmp, .pov)

??

This time I got the WAIT_ANY error on aix too (I wonder why I didn't get it last time!, maybe because I am using gnu-make ?) and subsituted it with -1.

Probably because you didn't use the CVS code. The CVS code is free of WAIT_ANY. Well, I made a new tarball with it... With regard to glib/gtk 1.2.10, effectively glib has no func_data inside, but gtk+ is full of it! Maybe simply undefining it in dillo (or any source) yields some obscure side-effects. You can re-compile gtk+-1.2.10 using something like this script: <q> #!/bin/bash # # Call with: find -name "*" -exec ./fix {} \; # if test -n "`echo $1 | sed 's/l$//' | grep "\.[chp]$"`"; then echo $1 mv $1 $1.BAK sed 's/\([^_]\)func_data/\1func_par/g' $1.BAK > $1 rm $1.BAK fi </q> It will replace "func_data" with "func_par" all along the code. That way, you'll not need to tweak dillo or any other program with regard to func_data. How? unpack e clean gtk+-1.2.10 cd gtk+-1.2.10 copy "fix" to this directory chmod +x fix find -name "*" -exec ./fix {} \; ./configure make ... Then grab dillo-0.8.0-pre.Feb02.tar.gz from dillo.org/test/ and try it! Cheers Jorge.-