The throttle idea is to control how much time is spent looking for a
symbol, thus the 'throttle' term. It was derived by getting the code
to maximum search paramters, then cutting sections out a piece at a
time and putting it behind the throttle so the speed could be
improved. As it is generating discussion on parts of the search, it
is already fullfilling a useful purpose.
The 'unloaded' feature does seem like it needs another version to
load cache files. I've added notes to the code to add the option.
>>> Bruce Stephens <bruce+cedet@...> seems to think that:
>If I understand things correctly (and I almost certainly don't),
>semanticdb-find-default-throttle is mixing two quite different kinds
> `file' - The file the search is being performed from.
> This option is here for completeness only, and
> is assumed to always be on.
> `local' - Tables from the same local directory are included.
> This includes files directly referenced by a file name
> which might be in a different directory.
> `project' - Tables from the same local project are included
> If `project' is specified, then `local' is assumed.
> `unloaded' - If a table is not in memory, load it. If it is not cached
> on disk either, get the source, parse it, and create
> the table.
> `system' - Tables from system databases. These are specifically
> tables from system header files, or language equivalent.
> `recursive' - For include based searches, includes tables referenced
> by included files.
> `omniscience' - Included system databases which are omniscience, or
> somehow know everything. Omniscience databases are found
> in `semanticdb-project-system-databases'.
> The Emacs Lisp system DB is an omniscience database.
>file, local, project, system, and recursive seem to be about what
>tables will be searched for a tag, but doesn't say that they'll be
>loaded (so if they're somehow already loaded then they'll be searched,
>but otherwise they won't). (By "searched", I really mean this
>tooltips thing that shows function prototypes when you leave point
>alone for long enough; I'm sure there are other kinds of search, but I
>can never seem to remember how to get to them.)
>And unloaded and omniscience indicate how tables might be loaded, if
>they haven't already been. (Maybe not omniscience, but unloaded is
>Presuming that's right, why isn't unloaded a separate flag? And why
>isn't it two separate flags (one for automatically loading cached
>tables, and one for parsing if necessary)?
>SF email is sponsored by - The IT Product Guide
>Read honest & candid reviews on hundreds of IT Products from real users.
>Discover which products truly live up to the hype. Start reading now.
>cedet-semantic mailing list
Eric Ludlam: zappo@..., eric@...
Home: http://www.ludlam.net Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net GNU: http://www.gnu.org