I would say this enhancement is welcome.
My question is if we should commit the partial patches or to keep them
pending until the whole solution is available.
regards,
bogdan
Ovidiu Sas wrote:
Hi all,
As Juha mentioned it into the lcr doc (see TODO), the second query
should be rewritten in C. This is my intention and I would like to do
it in three steps:
step 1: remove the first raw query (done).
step 2: load the lcr table into the memory.
step 3. replace the secondary query with C code.
I prefer to do it in three steps because because it slips easier in my
schedule and also it's much more easy to review small pieces of code
with well define functionality.
-ovi
On Tue, 6 Dec 2005, Juha Heinanen wrote:
Bogdan-Andrei Iancu writes:
> "SELECT %.*s.%.*s, %.*s.%.*s, %.*s.%.*s, %.*s.%.*s, %.*s.%.*s,
%.*s.%.*s
> FROM %.*s, %.*s WHERE '%.*s' LIKE %.*s.%.*s AND '%.*s' LIKE
> CONCAT(%.*s.%.*s, '%%') AND %.*s.%.*s = %.*s.%.*s ORDER BY
> CHAR_LENGTH(%.*s.%.*s), %.*s.%.*s DESC, RAND()"
>
> I would say the patch makes sense only if it solves completely the
> problem....
yes, and the real solution is to avoid the db query altogether and write
the code in c. this can be done by first extending reload_gws() to load
also the routes to memory and then by rewriting load_gws() so that it
uses in-memory information only.
-- juha
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
