Ruby/Tk 100% CPU problem maybe fixed

ruby-talk
1

Hi,

I renewed Tk.mainloop. It may have no 100% CPU problem.
Please get it from CVS and check whether it has proper
timer values or not.

You may be able to replace Ruby/Tk on Ruby1.6.8 by it.

If you want, please try at your own risk.

···

–
Hidetoshi Nagai (nagai@ai.kyutech.ac.jp)

2

Hidetoshi NAGAI wrote:

Hi,

I renewed Tk.mainloop. It may have no 100% CPU problem.
Please get it from CVS and check whether it has proper
timer values or not.

You may be able to replace Ruby/Tk on Ruby1.6.8 by it.

If you want, please try at your own risk.

under win32 there is also a thread with
priority “highest” - really ugly,…

daniel

3

Thank you for your report.

···

From: daniel offstuff@aon.at
Subject: Re: Ruby/Tk 100% CPU problem maybe fixed
Date: Tue, 10 Jun 2003 06:21:45 +0900
Message-ID: 3ee4f4c4$0$22264$91cee783@newsreader02.highway.telekom.at

under win32 there is also a thread with
priority “highest” - really ugly,…

win32… I don’t have a test environment for win32-ruby.
Could you tell me details of your ruby environment?
If your ruby is running on only one thread (main thread)
and the thread is Tk.mainloop, the cost of such process is
almost equal to the cost of Tcl/Tk. (I think so.)
Is Tcl/Tk running under lower priority?
Even if you change parameters on tcltklib.c (line 72–77;
especially DEFAULT_NO_EVENT_WATI on line 74 and
NO_THREAD_INTERRUPT_TIME on line 77), do you get the same result?

                              Hidetoshi Nagai (nagai@ai.kyutech.ac.jp)