Debugging unbearably slow in specific situation?

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Debugging unbearably slow in specific situation?

Stephan Herrmann-2
Hi,

Let me check if this is worth a bugzilla: Is anybody else
seeing the following?

In a debug session
(1) debugger had previously stop at various breakpoints
(2) then control flow enters JrtFileSystem.walkModuleImage()

at this point it's practically impossible to debug beyond this
method invocation, because stepping takes "forever" (deep inside
internals of file walking).
Sometimes I can do individual step overs, but I am unable to
reach the bottom of walkModuleImage().

I've seen this and & off for quite a while, always at this same
code location.

Right now, pressing suspend opens the call stack, but keeps
saying "Stepping" and step buttons remain disabled. Only
"Resume" and "Terminate" are enabled.

I see no sign of any locks being involved. All other threads
are basically idle. I already said "Skip All Breakpoints".
After "Resume" for all threads and letting it sit a minute,
I *may* be lucky to be able to continue stepping, but after a
few F7's it may get stuck again.

What's funny: the same debugging works fine, if no breakpoint was
hit before walkModuleImage().

I'd file a bug against JDT/Debug, but frankly I have no clue
if this is fixable in JDT, or a problem of the JVM, or just
a conspiracy against me, that happens only on my machine ;p

I recall that some years ago I had a similar problem, when
the program I was debugging launched Ant in-process. I think
back then it got stuck during XML parsing. Not sure if there's
any relation.

anyone?
Stephan
_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
Reply | Threaded
Open this post in threaded view
|

Re: Debugging unbearably slow in specific situation?

Duft Markus
Hey,

I have encountered similar issues in various areas in our code. I too have no clue whether it is a JVM or JDT issue... Unfortunatly it nearly always happens near closed source code, and I never managed to really reproduce anything reportable - thus I didn't until now :) This happens for us ever since 4.2.x or even earlier...

Just wanted to say you're not (completely) alone :D

Cheers,
Markus

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of Stephan Herrmann
Sent: Tuesday, January 16, 2018 12:57 PM
To: [hidden email]
Subject: [jdt-dev] Debugging unbearably slow in specific situation?

Hi,

Let me check if this is worth a bugzilla: Is anybody else
seeing the following?

In a debug session
(1) debugger had previously stop at various breakpoints
(2) then control flow enters JrtFileSystem.walkModuleImage()

at this point it's practically impossible to debug beyond this
method invocation, because stepping takes "forever" (deep inside
internals of file walking).
Sometimes I can do individual step overs, but I am unable to
reach the bottom of walkModuleImage().

I've seen this and & off for quite a while, always at this same
code location.

Right now, pressing suspend opens the call stack, but keeps
saying "Stepping" and step buttons remain disabled. Only
"Resume" and "Terminate" are enabled.

I see no sign of any locks being involved. All other threads
are basically idle. I already said "Skip All Breakpoints".
After "Resume" for all threads and letting it sit a minute,
I *may* be lucky to be able to continue stepping, but after a
few F7's it may get stuck again.

What's funny: the same debugging works fine, if no breakpoint was
hit before walkModuleImage().

I'd file a bug against JDT/Debug, but frankly I have no clue
if this is fixable in JDT, or a problem of the JVM, or just
a conspiracy against me, that happens only on my machine ;p

I recall that some years ago I had a similar problem, when
the program I was debugging launched Ant in-process. I think
back then it got stuck during XML parsing. Not sure if there's
any relation.

anyone?
Stephan
_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev

SSI Schäfer IT Solutions GmbH | Friesachstrasse 15 | 8114 Friesach | Austria
Registered Office: Friesach | Commercial Register: 49324 K | VAT no. ATU28654300
Commercial Court: Landesgericht für Zivilrechtssachen Graz
_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
Reply | Threaded
Open this post in threaded view
|

Re: Debugging unbearably slow in specific situation?

Christian Pontesegger
In reply to this post by Stephan Herrmann-2
Same here. I too have nothing reproducible to report a bug. I often
experience this when stepping over a method invokation and then wonder
why that method takes ages to return while doing single steps through
that same method would execute much faster.

I did not see this behavior before Oxygen.

Christian
_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev
Reply | Threaded
Open this post in threaded view
|

Re: Debugging unbearably slow in specific situation?

Daniel Megert
Is only debugging very slow, or is the whole IDE hanging?

A bigger new feature is the advanced source lookup. You can try whether disabling this makes the problem go away (Java > Debug > Advanced source lookup).

Dani



From:        Christian Pontesegger <[hidden email]>
To:        [hidden email]
Date:        30.01.2018 08:38
Subject:        Re: [jdt-dev] Debugging unbearably slow in specific situation?
Sent by:        [hidden email]




Same here. I too have nothing reproducible to report a bug. I often
experience this when stepping over a method invokation and then wonder
why that method takes ages to return while doing single steps through
that same method would execute much faster.

I did not see this behavior before Oxygen.

Christian
_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://urldefense.proofpoint.com/v2/url?u=https-3A__dev.eclipse.org_mailman_listinfo_jdt-2Ddev&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=1UITCR5rxUZHSFczvfaNFK4ymEbEiccRX7VKchpqz0Y&m=rWYEa09yIaEXMUsxFzTVuEcYl1OwNstKl0HkjHKHmTs&s=y7ZWYsm6cbBvCaEOPukPbuvzXPz5Jx885Jmg_fKTVG0&e=





_______________________________________________
jdt-dev mailing list
[hidden email]
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jdt-dev