On 02/12/2010 01:25 PM, Andrey Asadchev wrote:
> couple questions:
> 1) I am trying to index /usr/include at once using semanticdb.sh. Once
> the database is complete, how can I disable parsing include files from
> that directory while still retaining cached tags?
Semantic keeps a size/date reference in it's databases, and will reparse
if the cache is out of date. It shouldn't reparse them over and over,
but it will reference the databases to build up indexed tables for fast
symbol lookup later.
If it is doing full reparses, then some debugging is needed to figure
> 2) right now it seems semantic DB with global back end expects to find
> GTAGS in every directory. If the directory and all its children are
> indexed in one file, is it possible to instruct
> semantic to use that tags for all subdirectories?
For a particular buffer, use semanticdb-adebug-project-database-list to
see if GNU Global is enabled there. It runs a "global" query to
determine the root, and thus share one GTAGS table for a diretory tree.
M-x cedet-gnu-global-show-root RET
to see what it thinks.
> couple bugs?
> 3) I am encountering some problems where semantic finds function like
> macro. Especially things go bad in boost/mpl which uses macros
> heavily. Does current semantic able to handle such macros?
The CVS version of CEDET has a new set of macro handles for #if
statements that has not been heavily tested yet. If you are using pre6,
the macro handlers have been improved a lot since that release.
> 4) in some g++ installations limits.h must include compiler-proper
> limits.h from /usr/lib/... using include_next directive. One of the
> offending macros is LLONG_MAX and its derivatives.
You can use semantic-c-describe-environment to see how it interpreted
this macro. If you are using the CVS version of CEDET and this problem
persists, you could email the smallest C++ file that causes the issue