You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "William A. Rowe, Jr." <wr...@rowe-clan.net> on 2007/12/13 18:52:38 UTC
Re: Subclipse 1.2.4, APR Lib Crash
Jack Hardcastle wrote:
> Subclipse crashes Eclipse whenever I attempt any network-based action.
> Update, commit, retrieve repo listing, etc.
>
> My C is a bit rusty, but to my untrained eyes this appears to be an APR
> memory/socket glitch (?) . I'm e-mailing to both addresses per
> instructions at:
>
> http://subversion.tigris.org/bugs.html
tigris should actually vet their reports first before escallating to
APR, but essentially, you are right, there is a memory glitch, but
this is undoubtedly attributed to a mistake in the pool or allocation
scope in subclipse's source code, that socket was likely created in
the wrong pool which was since destroyed.
There is one edge case in older apr libraries when a certain threshold
is hit, did this happen to occur only after extensive processing?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Subclipse 1.2.4, APR Lib Crash
Posted by Mark Phippard <ma...@gmail.com>.
On Dec 13, 2007 1:52 PM, William A. Rowe, Jr. <wr...@rowe-clan.net> wrote:
> Jack Hardcastle wrote:
> > Subclipse crashes Eclipse whenever I attempt any network-based action.
> > Update, commit, retrieve repo listing, etc.
> >
> > My C is a bit rusty, but to my untrained eyes this appears to be an APR
> > memory/socket glitch (?) . I'm e-mailing to both addresses per
> > instructions at:
> >
> > http://subversion.tigris.org/bugs.html
>
> tigris should actually vet their reports first before escallating to
> APR, but essentially, you are right, there is a memory glitch, but
> this is undoubtedly attributed to a mistake in the pool or allocation
> scope in subclipse's source code, that socket was likely created in
> the wrong pool which was since destroyed.
>
> There is one edge case in older apr libraries when a certain threshold
> is hit, did this happen to occur only after extensive processing?
I am not aware of anyone from Subversion or Subclipse that escalated
this. This is a known issue with a simple workaround:
http://subclipse.tigris.org/faq.html#win32-crash
Subversion 1.5 no longer uses APR-ICONV so this problem will go away with 1.5.
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
Re: Subclipse 1.2.4, APR Lib Crash
Posted by Mark Phippard <ma...@gmail.com>.
On Dec 13, 2007 1:52 PM, William A. Rowe, Jr. <wr...@rowe-clan.net> wrote:
> Jack Hardcastle wrote:
> > Subclipse crashes Eclipse whenever I attempt any network-based action.
> > Update, commit, retrieve repo listing, etc.
> >
> > My C is a bit rusty, but to my untrained eyes this appears to be an APR
> > memory/socket glitch (?) . I'm e-mailing to both addresses per
> > instructions at:
> >
> > http://subversion.tigris.org/bugs.html
>
> tigris should actually vet their reports first before escallating to
> APR, but essentially, you are right, there is a memory glitch, but
> this is undoubtedly attributed to a mistake in the pool or allocation
> scope in subclipse's source code, that socket was likely created in
> the wrong pool which was since destroyed.
>
> There is one edge case in older apr libraries when a certain threshold
> is hit, did this happen to occur only after extensive processing?
I am not aware of anyone from Subversion or Subclipse that escalated
this. This is a known issue with a simple workaround:
http://subclipse.tigris.org/faq.html#win32-crash
Subversion 1.5 no longer uses APR-ICONV so this problem will go away with 1.5.
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org