Hi > > > In the long run, I think the best solution is to move those > > > time-consuming tasks out of UI-thread. What do you guys think?
I agree with Cedric for not only bcml task but also build failure data collections moving to another thread. I guess that reason why you feel so slow is coming from build failure colleciton too. I am not sure Cedric is the first persion, but we have used the BCML metric colleciton for more that one and half years but no one complained this before (or they knew the problem but they thought it's trivial :-)). I guess there is another reason (I suspect build faiure calcuation (this would be heavy calculation)) too or something else. Of course, we can BCML task disable as a default, and it would be OK too. Let me try to see what's going on more and I appreciate if you guys monitor the events as well as sensor itself. > I believe I haven't seen the crashes since that time, but I'll watch > this week to be sure. Since my Hackystat plugin is older, I'd suspect > Jupiter more, but it is equally likely to be something else...we'll > see. Is this Jupiter plug-in or Jupiter sensor plug-in? it is reported that there was a case that Eclipse seems hang up 10 seconds on Eclipse startup after Jupiter was installed. I guess it would be the server ping problem to check the new version of the plugn-in (the thread is waiting for the ping reponse so that it seems hanging up for a while, I definately need to move the update checking funciton to another thread so that users would be happy). Cheers, Takuya On Mon, 11 Oct 2004 10:23:54 -1000 Tim Shadel <[EMAIL PROTECTED]> wrote: > I actually disabled both my Hackystat and Jupiter plugins last > Wednesday. I was experiencing 2-3 crashes of Eclipse per day, and the > .metadata/.log file showed no activity near those times. I didn't > check the .hackystat/logs, so the connection is still merely > hypothetical. > > Hackystat plugin: 6.0.707.3x > Jupiter plugin: 2.1.927.3x > Hackystat server: 6.0.707 > > I believe I haven't seen the crashes since that time, but I'll watch > this week to be sure. Since my Hackystat plugin is older, I'd suspect > Jupiter more, but it is equally likely to be something else...we'll > see. > > On Sun, 10 Oct 2004 07:17:49 -1000, Philip Johnson <[EMAIL PROTECTED]> wrote: > > I agree and I've just posted this issue to Jira for Takuya to investigate: > > > > <http://hackydev.ics.hawaii.edu:8080/browse/HACK-67> > > > > I've optimistically included this as something for Takuya to fix for the > > 6.2 stable release. Takuya, let me know if this is not feasible. > > > > Cheers, > > Philip > > > > --On Sunday, October 10, 2004 1:53 AM -1000 "(Cedric) Qin ZHANG" > > > > > > <[EMAIL PROTECTED]> wrote: > > > > > Hi, > > > > > > My eclipse (both on my notebook and workstation) stalls periodically. > > > Every time the UI hangs, there is lots of disk activity. My guess is that > > > eclipse sensor is busy computing BCML file metrics. > > > > > > Takuya, is there a way that I can suppress file metrics collection in > > > eclipse sensor? > > > > > > In the long run, I think the best solution is to move those > > > time-consuming tasks out of UI-thread. What do you guys think? > > > > > > Cheers, > > > > > > Cedric > > > > > > > ================================ Takuya Yamashita E-mail: [EMAIL PROTECTED] ================================
