| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
| |
system. It no longer grows with time that I can tell, but I am sure there is some additional cleanup and optimizations which can still be done later.
|
| |
|
|
| |
that any class type need this defined.
|
| |
|
|
| |
Going to check if this fixes the memory leak issue on some systems (have one I can test on now).
|
| |
|
|
| |
currently-registered default terminal is rather than just xterm (although xterm is still the default/fallback).
|
| |
|
|
| |
installed/removed during a session.
|
| |
|
|
| |
itself can be accessed as needed.
|
| |
|
|
|
|
| |
1) Add in all the bits to make it auto-update as needed (optional init argument)
2) Have it also keep track of what previous files were *removed* when it updates.
3) Tweak the update routine to try and make it even faster.
|
| |
|
|
|
|
|
|
|
|
| |
This now uses a new data class which provides for delta updates to the application list, allowing for later checks/updates to be performed a *lot* faster, in addition to providing information about which *new* application entries were discovered during the check in addition to the raw app list.
As an added bonus, the new class-based nature of the routine will make it possible to embed the detection/update routines directly into the class rather than requiring some external class to monitor the system and request updates.
Example Times:
First run: same as old routine.
Second run with one new app entry: new routine is ~1/4 of the time of the old routine (20ms vs 78ms on my test box)
|
|
|
Now the utilites are arranged by category (core, core-utils, desktop-utils), so all the -utils may be excluded by a package system (or turned into separate packages) as needed.
|
