26 comments
Comment actions Permalink

Davide Baroncelli wrote:

Am I the only one to find it slower than previously?


It's always been slow and unnecessary on our end.
1 out of 3 times it replaces a couple of classes (which take 5-10
seconds) even when I only change static JavaScript files.

Amnon

0
Comment actions Permalink

Davide Baroncelli wrote:

Am I the only one to find it slower than

previously?

It's always been slow and unnecessary on our end.
1 out of 3 times it replaces a couple of classes
(which take 5-10
seconds) even when I only change static JavaScript
files.


Do you find it sometimes reloads even if you specify never reload?

0
Comment actions Permalink

I've also noticed this.

I've disabled it -- it's faster to restart our server than to wait for the hotswap to finish (30+ seconds for just a few classes).

0
Comment actions Permalink

Glen Stampoultzis wrote:

Do you find it sometimes reloads even if you specify never reload?

Haven't tried to disable it completely. Can you remind me where I can
disable it?

Amnon

0
Comment actions Permalink

Could you please send me a CPU snapshot?

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"

"Anders Wallgren" <anders_wallgren@alum.mit.edu> wrote in message news:13847166.1136514844778.JavaMail.itn@is.intellij.net...

I've also noticed this.

>

I've disabled it -- it's faster to restart our server than to wait for the hotswap to finish (30+ seconds for just a few classes).



0
Comment actions Permalink

I'm not sure what type of CPU snapshot you're refering to. Could you enlighten me, and I'll be happy to forward the info.

FWIW, the CPU seems to be idle during the 90+ seconds it takes to load a single method change.

0
Comment actions Permalink

Hello Anders,

I'm not sure what type of CPU snapshot you're refering to. Could you
enlighten me, and I'll be happy to forward the info.

FWIW, the CPU seems to be idle during the 90+ seconds it takes to load
a single method change.


See the last comment to this KB article: http://support.jetbrains.com/kb/entry.jspa?categoryID=5&externalID=3
.
There are instructions how to provide CPU snapshot.

--
Serge Baranov
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"




0
Comment actions Permalink

OK, I've uploaded a snapshot to ftp://intellij.net/.uploads. It is called 4109_anders_16.01.2006_18.19.04.zip.

I did the following steps:

  • started my program, let it initialize

  • touched a single file

  • started the CPU snapshot

  • rebuilt

  • single class reload took 100 seconds

  • stopped the cpu snapshot.

0
Comment actions Permalink

Hello Anders,
According to the snapshot, compilation took ~ 4 seconds, actual reload took 4 seconds. The other threads listed in the snapshot do
run for 100 seconds but what they do is waiting for another processes or for data from sockets. The conclusion is that the hotswap
itself works relatively fast and the bottlenack is somewhere else.
Do you have any third-party plugins installed? If so, how does IDEA's hotswap behave without them?


--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"

"Anders Wallgren" <anders_wallgren@alum.mit.edu> wrote in message news:27267839.1137464670430.JavaMail.itn@is.intellij.net...

OK, I've uploaded a snapshot to ftp://intellij.net/.uploads. It is called 4109_anders_16.01.2006_18.19.04.zip.

>

I did the following steps:

>

  • started my program, let it initialize

  • touched a single file

  • started the CPU snapshot

  • rebuilt

  • single class reload took 100 seconds

  • stopped the cpu snapshot.



0
Comment actions Permalink

I tried the following steps to make it work:

  • Removed all third-party plugins.

  • Removed all plugins

  • Did a fresh install of IDEA (removed all preferences, all settings, etc.)


The problem remains: reloading classes takes 90+ seconds.

0
Comment actions Permalink

Did you remove the plugins from both "plugins" directories (one under IDEA home and another under "config" directory)?

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"


0
Comment actions Permalink

Also, please don't be offended by my additional questions, I just try to eliminate all possible reasons of slowness that first come
to my mind. Especially when I see from the snapshot that hotswap actually takes much less time (4 sec) instead of 100 sec for which
the UI appears to be frozen.

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"


0
Comment actions Permalink

Davide, could you send me the CPU snapshot like Anders did?

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"


0
Comment actions Permalink

Hello Eugene,
Ant task which performs Hotswap works much faster then IDEA one almost instantly. But it connects, hotswaps and then disconnects from app. May be it related to debugging stuff synchroniation?

Thanks,
Dmitry

0
Comment actions Permalink

Davide, could you send me the CPU snapshot like
Anders did?


Don't worry, in my case either the problem disappeared with an EAP upgrade or it was not related specifically to hotswap (which now performs "normally", which means it's not immediate but lasts an acceptable number of seconds - let's say 3-7). Anyway, as I wrote in another thread:

1) I can't upload snapshots because I have no ftp access at work ( :| )
2) The only performance problems I experience are related to increases in garbage collection rate and durations when IDEA (or some of the plugins?) leaks memory (especially when a Full GC hits a memory area swapped-out by windows). I'd like to be able to directly analyze the memory snapshots by myself: couldn't we have a crippled version of yourkit GUI which is only able to open idea snapshots?

0
Comment actions Permalink

Hello Dmitry,

Ant task which performs Hotswap works much faster then IDEA one almost instantly. But it connects, hotswaps and then disconnects
from app. May be it related to debugging stuff >synchroniation?


This is what I'm trying to find out from snapshots.
If you have similar problem and it is possible for you to provide a snapshot, I'll be glad to examine it.


--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"


0
Comment actions Permalink

No worries -- I think we all understand how difficult it can be to narrow these things down.

I finally uninstalled IDEA, then removed the install directory, the ~/.IntelliJIdea50, all traces of IDEA I could find.

Then I did a fresh install, loaded my project and repeated the how swap measurement, and it was still taking the same time.

I will try a trivial project to determine if it may be related to my project. We're using Spring, Hibernate, etc., so maybe there's something going on there.

0
Comment actions Permalink

In article <25112818.1137777433768.JavaMail.itn@is.intellij.net>,
anders_wallgren@alum.mit.edu says...

No worries -- I think we all understand how difficult it can be to narrow these things down.

I finally uninstalled IDEA, then removed the install directory, the ~/.IntelliJIdea50, all traces of IDEA I could find.

Then I did a fresh install, loaded my project and repeated the how swap measurement, and it was still taking the same time.

I will try a trivial project to determine if it may be related to my project. We're using Spring, Hibernate, etc., so maybe there's something going on there.

For an additional datapoint, I use both Spring and Hibernate and haven't
seen hotswap problems.
--
-


David H. McCoy


-


0
Comment actions Permalink

This is still happening for me in Demetra build 5131.

0
Comment actions Permalink

Hello Anders,
We've just got an interesting feedback (many thanks to Chris Thielen):
http://www.jetbrains.net/jira/browse/IDEA-6540?page=comments#action_45886

You seem to also use linux, please check whether the problem above (ant its solution) is applicable in your environment.

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"

"Anders Wallgren" <anders_wallgren@alum.mit.edu> wrote in message news:26806344.1139358605718.JavaMail.itn@is.intellij.net...

This is still happening for me in Demetra build 5131.



0
Comment actions Permalink

Unfortunately, I'm experiencing this problem on Windows XP, not linux.

0
Comment actions Permalink

You mentioned you debugged a server app. Does it also run on WinXP?

--
Best regards,
Eugene Zhuravlev
Software Developer
JetBrains Inc.
http://www.jetbrains.com
"Develop with pleasure!"

"Anders Wallgren" <anders_wallgren@alum.mit.edu> wrote in message news:6569027.1140202231723.JavaMail.javamailuser@localhost...

Unfortunately, I'm experiencing this problem on Windows XP, not linux.



0
Comment actions Permalink

Yes, IDEA and the server app are running on the same Windows XP machine.

0
Comment actions Permalink

Hello,

Has this issue been resolved? I have the same problem in IDEA 6.0.4 build 6148.

I'm running on Windows XP with JDK 1.6.0.

I took some snapshots as described above but they don't show any abnormal activity.

Regards
Anders

0
Comment actions Permalink

i have this in 6148 too. on xp, reloading takes way too long.

I noticed it only happens when my module is running with 1.6, idea is running with 1.5. I changed my module to test with 1.5 and its back to normal swap speed.

Message was edited by:
Andre Mermegas

0
Comment actions Permalink

It seems they have resolved the issue in Selena and it will be released in 6.0.5

http://www.jetbrains.net/jira/browse/IDEADEV-13821?rc=1

Anders

0

Please sign in to leave a comment.