[erlang-bugs] Regression on match in 17.0-rc1

[Anthony Ramine]

The problem comes from cerl_clauses. The more I look at it, the more I think this module shouldn’t be used like this.

-- 
Anthony Ramine

Le 3 févr. 2014 à 20:19, Anthony Ramine <n.oxyde@REDACTED> a écrit :

> Smallest Core sample:
> 
> module 'test' ['main'/1]
>    attributes []
> 'main'/1 =
>    %% Line 6
>    fun (_cor0) ->
>        case call 'io':'format' ([102|[110]]) of
>            <Y = X> when 'true' ->
>                Y
>    end
> end
> 
> Running this through sys_core_fold triggers the bug.
> 
> -- 
> Anthony Ramine
> 
> Le 3 févr. 2014 à 19:54, Anthony Ramine <n.oxyde@REDACTED> a écrit :
> 
>> Bug is in sys_core_fold. Culprit is probably e12b7d5331c58b41db06cadfa4af75b78b62a2b1.
>> 
>> -- 
>> Anthony Ramine
>> 
>> Le 3 févr. 2014 à 19:34, José Valim <jose.valim@REDACTED> a écrit :
>> 
>>> There is a regression in 17.0-rc1. The following code will execute io:format/2 twice:
>>> 
>>> -module(test).
>>> -export([test/0]).
>>> 
>>> test() ->
>>> X = Y = io:format(<<"foo~n">>),
>>> X,
>>> Y.
>>> 
>>> I ran it with:
>>> 
>>> $ erlc test.erl
>>> $ erl -s test test
>>> 
>>> Above is the minimum test case extracted from a more complex expression where I matched against a record:
>>> 
>>> Value = #elixir_env{} = expr()
>>> 
>>> Thank you,
>>> 
>>> José Valim
>>> www.plataformatec.com.br
>>> Skype: jv.ptec
>>> Founder and Lead Developer
>>> _______________________________________________
>>> erlang-bugs mailing list
>>> erlang-bugs@REDACTED
>>> http://erlang.org/mailman/listinfo/erlang-bugs
>> 
>