compile error in r16

[sendtopms]

rebar compile ==> examples (compile) ==> locks (compile) src/locks_agent.erl: error in parse transform 'locks_watcher': {{badmatch, {ok, {locks_watcher, [{abstract_code, no_abstract_code}]}}}, [{locks_watcher,get_exprs,2, [{file,"src/locks_watcher.erl"}, {line,111}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,71}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,96}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,100}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,95}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,98}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,98}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,100}]}]} make: *\ [compile] Error 1

[uwiger]

I can't reproduce that. The no_abstract_code error means that it can't find the debug_info. I'm not sure how that can be, since the locks_watcher.erl module has the -compile(debug_info) hard-coded in the source.

[sendtopms]

I am Erlang OTP R16 (R16B03-1), is that an issue?

[uwiger]

I have tried with R16B02 and R16B03. Have you tried running rebar clean compile?

[sendtopms]

I tried but got same error

$ rebar clean compile ==> examples (clean) ==> locks (clean) ==> examples (compile) Compiled src/gdict.erl src/test_cb.erl:39: Warning: behaviour locks_leader undefined Compiled src/test_cb.erl ==> locks (compile) Compiled src/locks_watcher.erl src/locks_leader.erl:129: Warning: opaque type election() is not exported Compiled src/locks_leader.erl Compiled src/locks_app.erl Compiled src/locks.erl Compiled src/locks_cycles.erl Compiled src/locks_sup.erl src/locks_agent.erl: error in parse transform 'locks_watcher': {{badmatch, {ok, {locks_watcher, [{abstract_code, no_abstract_code}]}}}, [{locks_watcher,get_exprs,2, [{file,"src/locks_watcher.erl"}, {line,110}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,70}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,95}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,99}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,94}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,97}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,97}]}, {locks_watcher,transform,1, [{file,"src/locks_watcher.erl"}, {line,99}]}]}

[uwiger]

Is your locks repository current? Can you try with a clean checkout of master?

[sendtopms]

Thank you Ulf, I actually got old version of rebar, I updated it and now it is compiling.

[uwiger]

Glad it was resolved.

[sendtopms]

thanks

[sendtopms]

Ulf, Can you guide me to the simple example of begin_transaction and end_transaction. Do I need to implement lock_leader for all the cases or can I just use begin/end transaction?

[uwiger]

Sure. You don't need to use lock_leader in order to use locking. Lock_leader is implemented on top of the locker.

Eshell V5.10.3  (abort with ^G)
1> application:start(locks).
ok
2> F = fun() ->
           {ok,A} = locks:spawn_agent(),
           try locks:lock(A,[lock],write),
                io:fwrite("~p: ~p has the lock!~n", [time(),self()]),
                timer:sleep(15000)
           after
              locks:end_transaction(A)
           end, receive stop -> ok end
        end.          
#Fun<erl_eval.20.80484245>
3> {P1 = spawn(F), P2 = spawn(F)}.
{<0.41.0>,<0.42.0>}
{10,26,34}: <0.41.0> has the lock!
{10,26,49}: <0.42.0> has the lock!
4> erlang:is_process_alive(P1).
true

In the above case, we have two processes trying for the same lock. P1 gets it first, sits on it for 15 seconds, then ends the transaction, at which time P2 gets the lock.

If the process that started a transaction dies, the lock is automatically released.

5> f(P1),f(P2), {P1 = spawn(F), P2 = spawn(F)}.
{10,31,40}: <0.47.0> has the lock!
{<0.47.0>,<0.48.0>}
6> exit(P1,kill).
{10,31,48}: <0.48.0> has the lock!
true

[sendtopms]

Thank you Ulf, I could not locate locks:spawn_agent() instead can I use begin transaction locks:begin_transaction

[uwiger]

Ouch, sorry. That is a function I have not yet pushed (note to self: push those changes soon)

locks:begin_transaction() does [roughly] the same thing, but is synchronous, and has a different return value.

[sendtopms]

Ohh.... when can I expect new APIs? For meantime should i use {ok, A} = locks_agent:start_link(). ?

[uwiger]

Yes, that's the one I used to use. Then I introduced spawn_agent() for performance reasons.

Most of the upcoming changes are performance-related, specifically aimed at scalability.

[sendtopms]

Great expecting to use it in my project.

[uwiger]

I pushed my local changes to the branch uw-optimizations and created a pull request, #4 .

Feedback is welcome. I would like to do some more testing work before merging it.

[sendtopms]

Thanks for the new code. I will try to test it later today.