Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#1032 Performance optimization on long lines

Won't_Implement
closed
Neil Hodgson
None
1
2013-11-08
2013-11-07
Sworddragon
No

On SciTE 3.3.5 if a very long line (for example 1 million characters or more) is in the visible area the performance will decrease heavily. Any input actions will be delayed for ~1 second.

But as soon as this line gets out of the visible area all is fine again. So it seems SciTE is making some internal optimizations to process only the visible lines. I'm wondering if this could also be done in a horizontal way: All content of a line which is beyond the horizontal scrollbar does get this sort of optimization too.

Discussion

  • Neil Hodgson
    Neil Hodgson
    2013-11-07

    • status: open --> closed
    • assigned_to: Neil Hodgson
    • Group: Completed --> Won't_Implement
    • Priority: 5 --> 1
     
  • Neil Hodgson
    Neil Hodgson
    2013-11-07

    A faster Scintilla would be an improvement so there could be an issue written up for each Scintilla feature saying it should be faster. Without a detailed explanation on how to improve performance, these reports would just be non-actionable noise.

    Closing as will never be worked on in this form.

     
  • Sworddragon
    Sworddragon
    2013-11-08

    A faster Scintilla would be an improvement so there could be an issue written up for each Scintilla feature saying it should be faster.

    Maybe, but have you ever tried to reproduce this? It is not a micro-optimization as it could be done for another features.

    Without a detailed explanation on how to improve performance, these reports would just be non-actionable noise.

    The description is already in the startpost. But here is another one:

    On my system with a resolution of 1680x1050 and a fixed font SciTE has place for 230 characters (rows) per line and 62 lines until the scrollbars appears. Now SciTE processes only the visible lines but all (in)visible rows. The request is just to do not process the invisible rows anymore (as it is already implemented for lines so maybe it is not an extensive task).

     
    Last edit: Sworddragon 2013-11-08