Welcome, Guest. Please login or register.

Author Topic: user-startup -- Really useful?  (Read 9515 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline Karlos

  • Sockologist
  • Global Moderator
  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 16879
  • Country: gb
  • Thanked: 5 times
    • Show all replies
Re: user-startup -- Really useful?
« on: October 24, 2003, 01:26:08 PM »
I have to admit, I find this discussion somewhat pointless. The user-startup file is completely optional and always has been.

If you don't think it's a good idea, don't use one, if you do, then do. To my mind, the existing concept of an 'if present then execute' script is fine and simply reflects the modular design of the OS as a whole.

Personally I haven't found any speed penalties as a result of executing the script compared to 'inlining' its contents into the startup-sequence. If you want to speed up assigns and what not, use a specific late-assignment tool or MCP.
int p; // A
 

Offline Karlos

  • Sockologist
  • Global Moderator
  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 16879
  • Country: gb
  • Thanked: 5 times
    • Show all replies
Re: user-startup -- Really useful?
« Reply #1 on: October 24, 2003, 01:36:34 PM »
It does indeed :-D
int p; // A
 

Offline Karlos

  • Sockologist
  • Global Moderator
  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 16879
  • Country: gb
  • Thanked: 5 times
    • Show all replies
Re: user-startup -- Really useful?
« Reply #2 on: October 25, 2003, 03:44:35 PM »
@Scholle

If your user-startup is being called by 'run' there is something badly wrong.

Quote

The speed aspect is obvious: When you have lots of programs in WBStartup AND a big user-startup, there's a good chance that they are executed in parallel which means lots of hard drive access back and forth


user-startup and wbstartup programs in parallel? Not on any system I have used.

A normal boot should be as follows

Commands in s:Startup-sequence are called sequentially (some are run, usually hacks/patches). Drivers are initialised, preferences are loaded etc.

If there is a user-startup, it is executed, again sequentially, causing the normal startup-sequence script to be suspended until it completes. It is not executed concurrently.

Workbench is then loaded, after any user-startup sequence has been processed.

Programs in WBStartup are launched in order of their start priority. Most use 'donotwait' which instructs workbench not to wait for the program to finish before continuing.

This is usually the busiest time, programs are being launched and workbench itself is being set up, images loaded etc. Your user-startup script should have long finished before this happens.

If you feel that you are getting too much strenuous activity at this point, add the 'delay' argument to loadwb, which causes workbench to wait for existing shell stuff to complete.
int p; // A
 

Offline Karlos

  • Sockologist
  • Global Moderator
  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 16879
  • Country: gb
  • Thanked: 5 times
    • Show all replies
Re: user-startup -- Really useful?
« Reply #3 on: October 25, 2003, 04:43:44 PM »
Man, that is really strange. I can see the thrashing that that would cause :-o

Hopefully removing the run >nil: part should help a great deal...
int p; // A