< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index  

Re: [reSIProcate] Fwd: [reSIProcate-commit] resiprocate 7077 nash:inside resip/stack/Transaction::process method contains complex


Personally I was trying for the minimal changes, but it's simply inevitable as "if I fixed the memory management here, it still could go wrong elsewhere".
 
Boost SharedPtr usage has been around for some times now, anyway after trying to follow the logic inside TransactionState::process, I have to say it's simply too easy to get lost of message ptr's ownership, I would say if there were comments aside telling the ownership has got taken would be very helpful. And I thought it might be OK to check in as 1.10 just released.
 
I would say impact 'might' be performance and existing codes need to be changed to tell "safer memory management", I would propose a code review on commit 7077 revision, to identify any suspicious codes.
 
 Nash
 
On 4/17/07, Scott Godin <slgodin@xxxxxxxxxxxx> wrote:

I agree as well.

 

Scott

 

From: resiprocate-devel-bounces@xxxxxxxxxxxxxxxxxxxx [mailto:resiprocate-devel-bounces@xxxxxxxxxxxxxxxxxxxx] On Behalf Of Jason Fischl
Sent: Monday, April 16, 2007 5:22 PM
To: Byron Campen
Cc: resiprocate-devel
Subject: Re: [reSIProcate] Fwd: [reSIProcate-commit] resiprocate 7077 nash:inside resip/stack/Transaction::process method contains complex

 

I'm also not very keen on this change without doing lots of thinking first. Can I suggest that we back this out for now.

Thanks
Jason

On 4/16/07, Byron Campen <bcampen@xxxxxxxxxxxx> wrote:

          Yikes! Can we put this in a branch somewhere so we have time to digest it? I get the distinct impression that this large of a change is going to cause a lot of headaches for a lot of people. Mind you, I can see why this sort of change might be desirable, but this seems a tad aggressive. I would also like to have a chance to look into the performance impact that this work has. I don't want to get lynched when it comes time to push 1.2 out...

 

Best regards,

Byron Campen

 

Begin forwarded message:



Date: April 16, 2007 5:07:24 AM CDT

Subject: [reSIProcate-commit] resiprocate 7077 nash: inside resip/stack/Transaction::process method contains complex

 

Project

resiprocate

New Revision

7077

Committer

nash (Nash Tsai)

Date

2007-04-16 05:07:24 -0500 (Mon, 16 Apr 2007)

Log

 inside resip/stack/Transaction::process method contains complex 

 logics and message ptr can be easily leaked if it's not maintained by
 original coder, therefore I changed to SharedPtr usage and made a bunch
 of code changed to make it compile.
 
 Also another place worth mention as TimeLimitFifo::add method could

 result memory leak usage, and it's fixed on the way for above changes.
 
 The tests under rutil/stack/dum has passed as well.
 
 resip/stack/Security add to allow disable server authentication check

 
 
 
 

Modified:

_______________________________________________

resiprocate-commit mailing list

 


_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel

 


_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel