Re: [Tora-develop] Alpha release - status update

[<iv...@cv...>]

Quoting Petr Vanek <pe...@ya...>:
>
> my observations:
>     - tora takes quite lot of CPU when is the window visible: about ~8%
> on my homeoffice desktop (old 2 cores), 2% on my office desktop (8
> cores). The CPU usage is reduced when is the window hidden behind
> another app window.
8% is really too much. Each Editor's syntax analyzer has own background thread
which periodically scans the whole source. Regardless if it was  
changed on not.
This also need some tweaking.
Which optimization level are you using? This should not be so slow.
Does it take CPU when some other tool (other than SQL Editor) has focus?

>
>     - that cool code folding is sometimes quite strange. Example:
>
>         [+] /*
>              foo
>              bar
>
>         [+] */

This is natural qscintilla behaviour :) At this how QsciLexerSQL works  
at this moment.
I agree this need some tweaking. Or maybe it should be "fixed" in  
qscintilla upstream.

>
>     - sql editor parser(?) requires empty line to find statement correctly:
>
> select sysdate from dual;
> select sysdate from dual;
>
>     cursor anywhere in these 2 lines:
> Thu Oct 31 12:14:27 2013
> trotl_stat.cpp:234    void trotl::SqlStatement::execute_describe()
> ORA-00911: invalid character
>
> error at SQL position 24 [1.25]:
> select sysdate from dual;
>                         ^
>

"This is not a bug but the feature" :) This is implemented in  
toSyntaxAnalyzerNL class.
When you click in bottom right corner onto "QtSql" you can switch from  
"generic Lexer/Parser/Analyzer"
onto Oracle specific one.
The Oracle one is much smarter, but beware - it's color scheme is horrible.
So far I did not align color schemas and they can not be configured yet.

Ivan



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.