Ok. I found the bug. The following patch it. It is already in the git
repository (debug-tree branch).
best
max
--- a/src/src/System/Language/cpp_language.cpp
+++ b/src/src/System/Language/cpp_language.cpp
@@ -464,6 +464,8 @@ static bool in_preprocessing (string s, tree t) {
string s2= s;
while (line_number(t2) > 0) {
t2= line_inc(t2,-1);
+ // line_inc return tree(ERROR) upon error.
+ if (!is_atomic(t2)) return false;
s2= t2->label;
if (!end_preprocessing(s2)) return false;
if (begin_preprocessing(s2)) return true;
On Nov 25, 2011, at 3:34 PM, Gubinelli Massimiliano wrote:
> Hi Joris,
> in https://gitorious.org/~mgubi/texmacs/new-scheme i added a branch
> debug-trees which implements the continuation patch and some patch for the
> t->obs bug and also some debugging to find further bugs. In this branch
> texmacs fails on the check_atomic function (which I added for debugging)
> during the typesetting of the full manual in cpp_language::get_color, but I
> do not understand why since the code is inlined and it is not easy to
> precisely point out the code line which is wrong.
>
> best
> max
>
> ps: I shift the discussion in texmacs-dev so maybe somebody would like to
> give a look...
_______________________________________________
Texmacs-dev mailing list
Texmacs-dev@gnu.org
https://lists.gnu.org/mailman/listinfo/texmacs-dev
