gproc process dies after multiple subscribes

[aberman]

1. Open an erlang shell
2. Type: application:start(gproc).
3. Type : gproc_ps:subscribe(l, test).
4. Repeat step 3

Result: gproc crashes and you see an error:

\ exception error: bad argument in function gproc:reg/1 called as gproc:reg({p,l,{gproc_ps_event,test}})

The supervisor does start the process again, and if you type the subscribe again, you'll get 'true' returned.

Expected result: A tuple like {error, already_subscribed}, should be returned and the gproc process should not die.

[uwiger]

Well, actually, it's the shell process that dies. The gproc processes remain unaffected, since the subscribe/2 function only registers a property (a process-local operation). The gproc API rather consistently enforces that trying to register the same entity twice raises an exception. However, this should of course be documented in gproc_ps:subscribe/2.

The following shell interaction illustrates that it's the shell process that dies.

$ erl
Erlang R15B (erts-5.9) [source] [64-bit] [smp:4:4] [async-threads:0] [hipe] [kernel-poll:false]

Eshell V5.9  (abort with ^G)
1> application:start(gproc).
ok
2> regs().
...
gproc                 <0.37.0>     gproc:init/1                        30    0
gproc_bcast           <0.39.0>     gproc_bcast:init/1                  26    0
gproc_monitor         <0.38.0>     gproc_monitor:init/1                28    0
gproc_sup             <0.36.0>     supervisor:gproc_sup/1             961    0
...
3> self().
<0.31.0>
4> gproc_ps:subscribe(l,test).
true
5> gproc_ps:subscribe(l,test).
** exception error: bad argument
     in function  gproc:reg/1
        called as gproc:reg({p,l,{gproc_ps_event,test}})
6> self().
<0.44.0>
7> regs().
...
gproc                 <0.37.0>     gproc:init/1                        55    0
gproc_bcast           <0.39.0>     gproc_bcast:init/1                  26    0
gproc_monitor         <0.38.0>     gproc_monitor:init/1                28    0
gproc_sup             <0.36.0>     supervisor:gproc_sup/1             961    0
...

[aberman]

Ah, ok, in that case I guess you can close this issue then unless you want it to remind you to document the exception. Thanks @uwiger !

[uwiger]

Docs updated.