summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Add missing specMatthew Sackman2011-08-171-0/+1
|
* In Erlang 'Ye Olde Versiony', you can't just call 'error'Matthew Sackman2011-08-171-2/+2
|
* More testing of which_childrenSimon MacMullen2011-08-171-5/+18
|
* {error, already_present}Simon MacMullen2011-08-172-4/+7
|
* Get algorithmic complexity of function right.Matthew Sackman2011-08-171-6/+10
| | | | | | | | Yes, it's 4 more lines, but: > L = [{X,X} || X <- lists:seq(1,1000000)]. > timer:tc(erlang,apply,[fun () -> mirrored_supervisor:merge_proplists(L, L) end, []]). {190570,[...]} whereas with previous implementation, after eight minutes, this still hadn't completed.
* OopsSimon MacMullen2011-08-151-1/+1
|
* Aggregate which_children, implement count_children.Simon MacMullen2011-08-152-2/+25
|
* Merge heads.Simon MacMullen2011-08-152-7/+28
|\
| * Make delete / restart / terminate work groupwide.Simon MacMullen2011-08-152-7/+28
| |
* | Simplification of comment; improve consistency of function arg ordering; ↵Matthew Sackman2011-08-151-13/+12
|/ | | | make mnesia table an ordered_set and always put the Group first - this will improve performance of delete_all and is generally nicer.
* That one isn't needed either.Simon MacMullen2011-08-111-2/+1
|
* Use {Id, Group} as key in Mnesia, and remove update since, err, write will ↵Simon MacMullen2011-08-112-31/+28
| | | | overwrite anyway...
* Merge headsSimon MacMullen2011-08-101-3/+10
|\
| * Fix stop raciness.Simon MacMullen2011-08-101-3/+10
| |
* | CosmeticMatthew Sackman2011-08-101-1/+1
| |
* | CosmeticMatthew Sackman2011-08-101-3/+1
|/
* Should explain this...Simon MacMullen2011-08-091-1/+3
|
* Flatten.Simon MacMullen2011-08-091-10/+4
|
* Slight test for ignore too.Simon MacMullen2011-08-091-0/+10
|
* Blow up if we're asked for simple_one_for_one, more tests, try to handle ↵Simon MacMullen2011-08-092-18/+56
| | | | Mod:init/1 returning ignore.
* Matthias thinks this is obvious :)Simon MacMullen2011-08-091-4/+0
|
* Clear mnesia table up if we're the first mirror to start, since otherwise we ↵Simon MacMullen2011-08-081-15/+32
| | | | can leak when all mirrors go down, mnesia does not and children were added dynamically.
* More doc.Simon MacMullen2011-08-081-0/+3
|
* Don't delete from mnesia if child is running.Simon MacMullen2011-08-082-2/+6
|
* Exit differently.Simon MacMullen2011-08-081-1/+1
|
* Explain cardinality relationship between modules and process groups.Simon MacMullen2011-08-081-0/+4
|
* The most important commit so far.Simon MacMullen2011-08-082-2/+2
|
* Dialyser points out I was returning the wrong thing here. Oops.Simon MacMullen2011-07-212-5/+5
|
* Bad OTP! No quotes!Simon MacMullen2011-07-211-1/+1
|
* Remove demonitor_all_peers since we always stop after invoking it.Simon MacMullen2011-07-211-24/+10
|
* Simplify group shutdown by repeated casting.Simon MacMullen2011-07-181-13/+11
|
* It might not be 'killed'. And we should return an error here.Simon MacMullen2011-07-181-3/+3
|
* I spent a long time trying to get the teardown of the complete supervision ↵Simon MacMullen2011-07-181-7/+18
| | | | hierarchy associated with each test to be completely synchronous and failing. It's much easier to create a new group for each test.
* already_startedSimon MacMullen2011-07-182-23/+24
|
* Handle being torn down while in the transaction.Simon MacMullen2011-07-181-2/+4
|
* Not even sure what that was there for. Oops.Simon MacMullen2011-07-181-2/+1
|
* Remove a sleep() by waiting for the supervisor and any children we care ↵Simon MacMullen2011-07-181-9/+21
| | | | about to go away.
* Catch more exit reasons (with a fork of rabbit_misc:with_exit_handler/2).Simon MacMullen2011-07-181-5/+16
|
* If start_child is called twice in a row we used to hang. Um, that's not ↵Simon MacMullen2011-07-182-7/+16
| | | | good. Fix.
* Defer all of the startup work to the second stage startup, so that we can't ↵Simon MacMullen2011-07-181-12/+15
| | | | get a DOWN message while we're still half started and then fall over.
* This was copied from R14B03 supervisor.erl, but there it was exported. I ↵Simon MacMullen2011-07-151-3/+0
| | | | don't think we want to export any types from here, people can get them from supervisor.
* More descriptive names, don't clash with erlang:monitor/2.Simon MacMullen2011-07-151-11/+12
|
* Merging default to bug24230Matthew Sackman2011-07-1422-723/+360
|\
| * fix bug introduced by bug 24216 (tx untangling)Matthias Radestock2011-07-141-1/+1
| | | | | | | | forgot to propagate an API change
| * cosmeticMatthias Radestock2011-07-141-1/+0
| |
| * Merge bug24216 into defaultRob Harrop2011-07-083-39/+44
| |\
| | * cosmeticMatthias Radestock2011-07-081-3/+2
| | |
| | * merge bug24244 into defaultMatthias Radestock2011-07-071-21/+25
| | |\
| | | * inliningMatthias Radestock2011-07-071-6/+6
| | | |
| | | * One last tweak.Simon MacMullen2011-07-061-3/+2
| | | |
View Lorry Controller Status