Welcome! Log In Create A New Profile

Advanced

Next beta release?

Posted by Anonymous User 
Anonymous User
Next beta release?
August 02, 2010 12:59PM
Hi everyone. I apologize if this is an overasked question, (I did my homework, by the way), but I'm curious when the next beta release is. With the June 9th release of r404, you mentioned in the release topic that the next release could be expected in a week or so. Well, here we are at the beginning of August. Any accurate release dates?
Re: Next beta release?
August 02, 2010 02:08PM
Heya tgp1994.

Right now, the developers are attempting to fix a critical issue before issuing a new version. I'm sorry that this has caused a lot of delay on launching new PeerBlock release.
Anonymous User
Re: Next beta release?
August 02, 2010 02:21PM
Tippy Wrote:
-------------------------------------------------------
> Heya tgp1994.
>
> Right now, the developers are attempting to fix a
> critical issue before issuing a new version. I'm
> sorry that this has caused a lot of delay on
> launching new PeerBlock release.


Could you perhaps shed some light on this issue?
Re: Next beta release?
August 02, 2010 02:53PM
The issue is Issue #331 (link: [code.google.com] ), the memory leak. This issue would be the most evident when under a heavy traffic use in long time period, I believe.
Anonymous User
Re: Next beta release?
August 02, 2010 03:04PM
Hmm. Seems to me that PB is keeping the log in the memory rather than writing it to the disk in realtime. Just my thoughts though.

Best of luck to the dev team, and thank you for that link.
Re: Next beta release?
August 03, 2010 09:37PM
avatar
Quote

Hmm. Seems to me that PB is keeping the log in the memory rather than writing it to the disk in realtime. Just my thoughts though.
Nope, though I wish it was something that easy...  sad smiley 

It appears as though we're leaking a small amount of memory when logging to history.db.  Haven't been able to trace down the exact source of the leak yet, though it does not appear to show itself if you disable history logging.  It does appear to be a "day one bug", meaning it's one that was also present in the original PG2 source code . . . as well as PeerBlock 1.0.  Since it's not a regression (it wasn't added during the current 1.1 development cycle) we're considering releasing it even with this bug still in there - it shouldn't be any worse in this regard than what's out currently, and a whole slew of other things we've fixed / changed since 1.0 was released make it a whole heck of a lot better overall.

Assuming we do end up deciding to release 1.1 with the memory leak unfixed (in which case we'll have it high on our list of things to fix for the first 1.1.x "bugfix release"), we should only have one more major issue which we need to take care of . . . which is the fact that if we have a truly retardedly-large history.db file we should just delete it while we're updating people's settings to prevent this from occurring again.  Currently we just slowly start archiving it away, which can cause PeerBlock to go unresponsive for awhile as it whittles away at the file.  This is only a one-time issue, but especially given that we have over a million downloads of our Stable Release this means a lot of people will be experiencing problems if we don't fix this sooner rather than later.

Sooooo, all this means that it should be Real Soon Now that we should be releasing a new Beta, shortly followed by a new Stable Release.  And after that, we'll get the 1.1+ branch ready for an even newer Beta Release.  We've been working on 1.1+ while we've been stabilizing 1.1, which means that we should have a new Beta ready pretty quickly. 

The primary slowdown from my standpoint (as Lead Dev) is that I just started a new job, and it's tough finding time to devote to PeerBlock right now - this should get better over the course of the next few weeks though, as I settle into a new routine. 

Hope that helps shed some more light as to what's going on in this space,

        ---  Mark  ---




Lead developer of PeerBlock
Anonymous User
Re: Next beta release?
August 03, 2010 10:23PM
Hmm... is it perhaps whatever API is used to store in the *.db format?

I'm fine with the time being Real Soon Now, all I just wanted to know was the holdup. Knowing that makes me feel better smiling smiley (One of the leading causes of accidents on a highway, I might add)

I've seen a lot of programs die because the developer gets a job/enters college. Don't let PG get in the way though; put your life first. smiling smiley
Re: Next beta release?
August 03, 2010 11:56PM
avatar
That's what happened to PG....lead developer opened a new business and it consumed all of his time. But you are right tgp1994 "put your life first." So if we have to wait for a new beta while your schedule gets organized then we will wait. We'll all back you Mark. That's my 2 cents worth.



Edited 1 time(s). Last edit at 08/03/2010 11:59PM by RobrPatty.
Anonymous User
Re: Next beta release?
August 04, 2010 12:05AM
What was I thinking... I meant to say PB, rather than PG, but I think you understood my meaning none the less smiling smiley
fxm
Re: Next beta release?
August 04, 2010 04:20AM
MarkSide Wrote:
-------------------------------------------------------

> It appears as though we're leaking a small amount
> of memory when logging to history.db.  Haven't
> been able to trace down the exact source of the
> leak yet, though it does not appear to show itself
> if you disable history logging. 

FWIW 'memory leak' would fit my symptoms exactly - except that I have logging to file disabled.
Could it be that the leak isn't in the actual logging but somewhere in the preliminaries?
Re: Next beta release?
August 06, 2010 10:39PM
avatar
Quote

FWIW 'memory leak' would fit my symptoms exactly - except that I have logging to file disabled. Could it be that the leak isn't in the actual logging but somewhere in the preliminaries?

Hmm, I don't know...  I've tried all the tricks I generally use to chase down memory leaks (UMDH, DebugDiag, etc.), but while I could get various stack traces that appeared to implicate our database-writing routines . . . they were somewhat suspect.  And the amounts of leak they were detecting weren't corresponding with the performance counters' idea of how much memory PeerBlock was allocating. 

We still haven't gotten PeerBlock running under a Debug build in a reasonably-performant manner yet, either, which is a PITA.  This means that I can't use any of the off-the-shelf memory-leak-detection libs I would otherwise have used.  sad smiley  Might be worth trying to look deeper into why the program crashes out when STL debugging is disabled, and/or creating our own custom allocator to help track allocs/frees...

Thanks for the words of support though, guys, I do appreciate it.  And don't worry, I have no intentions of abandoning the project like how the original PG2 author(s?) did back in the day.  I just ordered a new laptop in fact, one which will be easier to take on the train with me for the >1 hour commute I have each way, each day . . . which should let me get some "work" done on PeerBlock in the meantime, instead of just zoning out listening to music. 

Hopefully I'll be able to settle into a new routine soon now...  In the meantime, night_stalker_z's been working on stuff too, though mostly in the 1.1-followon release "trunk" branch...  As always, if any other devs out there are interested in joining us and helping out, we'd be more than happy to hear from you!

        ---  Mark  ---




Lead developer of PeerBlock
Anonymous User
Re: Next beta release?
August 06, 2010 11:10PM
Oh, so the memory leak is appearing in debug and release builds? I recall compilers inserting quite a bit more code into the program during debugging. Just out of curiosity, what compiler do you happen to be using?
Re: Next beta release?
August 06, 2010 11:16PM
tgp1994 [www.peerblock.com]

That's what MarkSide uses.
Anonymous User
Re: Next beta release?
August 07, 2010 04:50AM
What?
fxm
Re: Next beta release?
August 07, 2010 06:53AM
MarkSide Wrote:
-------------------------------------------------------

> I've tried all the tricks I
> generally use to chase down memory leaks (UMDH,
> DebugDiag, etc.), but while I could get various
> stack traces that appeared to implicate our
> database-writing routines . . . they were somewhat
> suspect.  And the amounts of leak they were
> detecting weren't corresponding with the
> performance counters' idea of how much memory
> PeerBlock was allocating. 

I have a related suspicion, but it could be a red herring.

Could a major leak require two triggers? Or could it involve something other than the page file? I ask because my [xp sp3] mysterious crashes seem to occur only after PB has been running for long enough that an automatic list-update might have been due.

When PB disappears, Security Essentials (MS realtime AV) also turns up dead [reporting "service stopped" and/or "out of memory"]. There have been other reports here of PB failures citing lack of resources [which I took to mean 'memory' as opposed to GDI heap].
Sorry, only registered users may post in this forum.

Click here to login