virtan / eep Goto Github PK

In the "Also" section you describe the function for filtering calls to a specific module:

Collect specific module calls only
1> eep:start_file_tracing("file_name", [], my_module).

It looks like it should actually be to a list of modules:

Collect specific module calls only
1> eep:start_file_tracing("file_name", [], [my_module_1, my_module_2]).

Is it normal that if I have 100 gen_servers running, the callgrind output (default, no grouping) contains a separate gen_server:loop/6 line for each of them? (location is something like "<6868.244.12>: gen_server" with varying pids).

This problem doesn't seem to show up on the screenshots attached to the eep doc.

Is it a kcachegrind issue?

λ uname -a
Linux 3.13.0-85-generic #129-Ubuntu SMP x86_64 x86_64 x86_64 GNU/Linux
λ kcachegrind -v
Qt: 4.8.6
KDE Development Platform: 4.13.3
KCachegrind: 0.7.4kde

The application crashes due to wrong return.

(eep@127.0.0.1)1> application:start(eep).

=INFO REPORT==== 19-May-2016::11:43:47 ===
    application: eep
    exited: {bad_return,{{eep_app,start,[normal,[]]},ok}}
    type: temporary
{error,{bad_return,{{eep_app,start,[normal,[]]},ok}}}

Inclusion of application into release with rebar3 terminates the node

(typhoon@127.0.0.1)1> {"Kernel pid terminated",application_controller,"{application_start_failure,eep,{bad_return,{{eep_app,start,[normal,[]]},ok}}}"}
Kernel pid terminated (application_controller) ({application_start_failure,eep,{bad_return,{{eep_app,start,[normal,[]]},ok}}})
heart: Thu May 19 11:40:00 2016: Erlang is crashing .. (waiting for crash dump file)
heart: Thu May 19 11:40:00 2016: Would reboot. Terminating.

First, i want trace only one module, eep:start_file_tracing("file_name", [], [b2bAtcfLib]). Then i run eep:stop_tracing(), eep:convert_tracing("file_name").. but it reports an error:

`=ERROR REPORT==== 30-Mar-2015::07:15:29 ===
Error in process <0.6572.6> on node 'ppb1_bs6@esekilvxen245' with exit value: {badarg,[{ets,lookup,[3333325396,<0.2044.0>],[]},{eep,callgrind_convertor,2,[{file,"src/eep.erl"},{line,167}]},{dbg,tc_loop,3,[{file,"dbg.erl"},{line,1241}]}]}

`

Hello,

I'm trying to use eep with Erlang/Elixir release which is runned in docker container. I've attached to remote console and executed:

iex(node@ip)1> :eep.start_file_tracing('myapp.trace')
{:ok, [{:matched, :"node@ip", 1843}]}

But after couple of seconds console is closed with error:

iex(node@ip)2> *** ERROR: Shell process terminated! (^G to start new job) ***
command terminated with exit code 137

Any ideas what I'm doing wrong? Thank you.

Hello,

eep:convert_tracing("file_name"). is failed

=ERROR REPORT==== 26-Nov-2013::10:32:13 ===
Error in process <0.52.0> on node '[email protected]' with exit value: {function_clause,[{eep,'-pid_item_to_bytes/1-fun-0-',[undefined,{1,1},[]],[{file,"src/eep.erl"},{line,275}]},{orddict,fold,3,[{file,"orddict.erl"},{line,194}]},{eep,pid_item_to_bytes,1,[{file,"src/eep.erl"},{line,275}]},{eep,convertor_unwind...

As I trace id down line 275 from eep.erl is root cause. The SubCalls dict contains undefined {CM, CF, CA}.

The project name conflicts with the Erlang Enhancement Proposals. Maybe a rename should be considered?

Hello,
I get the error message below when I'll try to convert file to cachegrind format.
OS: Windows 7, Erlang R16B03

=ERROR REPORT==== 3-Aug-2014::13:04:33 ===
Error in process <0.35.0> with exit value: {{badmatch,{error,eacces}},[{eep,save_kcachegrind_format,1,[{file,"src/eep.erl"},{line,298}]}]}

** exception error: no match of right hand side value {error,eacces}
in function eep:save_kcachegrind_format/1 (src/eep.erl, line 298)

`Error in process <7827.24524.146> on node '...' with exit value:
{problem,[{eep,save_kcachegrind_format,1,[{file,"eep.erl"},{line,315}]}]}

** exception exit: problem
in function eep:save_kcachegrind_format/1 (eep.erl, line 315)`

Can I profile a release with some modules compiled with hipe?

This looks like an amazing project, that will make profiling a lot easier than using built-in tools like e.g. eprof that only show text-based output.

Would it be possible to host this on Hex.pm? I run projects that are managed by Hex; it would make installing and updating a lot easier.

Thank you!