I have a pretty reliable permanent hang after updating to r9062 this evening. The hang isn't immediate, but happens pretty reliably after editing tracked files for 10 minutes or so.
Inspecting a sample of the process suggests it is scm:background_status that is hanging. My repo is git. I tried disabling the SCM bundle, but that doesn't appear to have helped.
The sample, if it's helpful: http://pastie.org/3411939
-MinRK
On 19/02/2012, at 11.42, MinRK wrote:
I have a pretty reliable permanent hang after updating to r9062 this evening. The hang isn't immediate, but happens pretty reliably after editing tracked files for 10 minutes or so.
Can’t think of why nor something I have experienced.
If others see similar, let me know.
Inspecting a sample of the process suggests it is scm:background_status that is hanging. My repo is git. I tried disabling the SCM bundle, but that doesn't appear to have helped.
The sample, if it's helpful: http://pastie.org/3411939
I need the ‘Binary Images’ part (at the bottom) to be able to translate the addresses to source/line numbers.
The background_status though is only a two line function, so not sure how much it’ll help.
On 19/02/2012, at 11.42, MinRK wrote:
I have a pretty reliable permanent hang after updating to r9062 this evening. The hang isn't immediate, but happens pretty reliably after editing tracked files for 10 minutes or so.
Btw: To “downgrade” one can switch back to Normal Releases and hold down shift when clicking Check Now.
I have a pretty reliable permanent hang after updating to r9062 this evening. The hang isn't immediate, but happens pretty reliably after editing tracked files for 10 minutes or so.
I'm not sure it is the same issue, but TM2 r9062 just hanged on a Cmd+V paste command: C code source, 1 line (~20 chars) pasted.
Cheers, Manu
On 19/02/2012, at 18.33, Emmanuel Blot wrote:
I'm not sure it is the same issue […]
If only you had run ‘sample TextMate’… ;)
Anyway, someone had same issue on IRC and I got a sample with the missing load addresses and reworked the code involved. This has been pushed as a test build, so those affected please test r9063 (by holding option when doing Check Now) and let me know if the problem persists (I didn’t actually find an issue, but I avoided using std::tr1::enable_shared_from_this which was indirectly involved in this).
On Sun, Feb 19, 2012 at 12:54 PM, Allan Odgaard mailinglist@textmate.org wrote:
If only you had run ‘sample TextMate’… ;)
I ran it, but I forgot to dump it :-( Anyway, here is a new sample trace as TM2 r9062 hung on performing the update towards r9063 ;-)
Thanks, Manu
On 19/02/2012, at 19.40, Emmanuel Blot wrote:
[…] here is a new sample trace as TM2 r9062 hung on performing the update towards r9063 ;-)
Looks like it could be same issue.
I pushed r9064 where I removed the code that used the mutex (indirect form a shared_ptr) which seemed to play an active role in the deadlock.
On Sun, Feb 19, 2012 at 10:26, Allan Odgaard mailinglist@textmate.orgwrote:
On 19/02/2012, at 19.40, Emmanuel Blot wrote:
[…] here is a new sample trace as TM2 r9062 hung on performing the update towards r9063 ;-)
Looks like it could be same issue.
I pushed r9064 where I removed the code that used the mutex (indirect form a shared_ptr) which seemed to play an active role in the deadlock.
Working with 9064 all day today, and not a hang yet so it seems to be fixed. Thanks!
I'll make sure to include the full sample next time.
-MinRK
textmate mailing list textmate@lists.macromates.com http://lists.macromates.com/listinfo/textmate
-
Allan Odgaard -
Emmanuel Blot -
MinRK