#5419 closed defect (invalid)
activation/deactivation stalls
| Reported by: | joe@… | Owned by: | macports-tickets@… |
|---|---|---|---|
| Priority: | Normal | Milestone: | |
| Component: | base | Version: | 1.0 |
| Keywords: | Cc: | markd@… | |
| Port: |
Description
Long stalls right after activation/deactivation of ports. Sometimes the stalls are so long that canceling and retrying is necessary.
Will test this further to find a way to reliably reproduce this problem.
Change History (5)
comment:1 Changed 19 years ago by joe@…
comment:3 Changed 18 years ago by danielluke (Daniel J. Luke)
I think it's INVALID, there's a lock that gets taken so two instances of 'port' don't clobber something or other by writing at the same time and the 'stall' is one of the port processes wating for the lock.
comment:4 Changed 18 years ago by markd@…
| Resolution: | → invalid |
|---|---|
| Status: | new → closed |
(In reply to comment #3)
I think it's INVALID, there's a lock that gets taken so two instances of 'port' don't clobber something or other by writing at the same time and the 'stall' is one of the port processes wating for the lock.
Closing per Daniel's comments.
comment:5 Changed 18 years ago by joe@…
I don't think this is still a problem, but I'll alert people if this becomes a problem again.
I was trying to deactivate libiconv while at the same time working on another compile, and ran into this problem. It almost seems like this problem occurs more frequently when DarwinPorts attempts to multitask, or is perhaps under a heavy load.
Could there be a problem with two processes simultaneously trying to claim a lock on a file?