[erlang-questions] Missing compile warning

[Vlad Dumitrescu]

On Mon, Mar 3, 2014 at 5:49 PM, Björn Gustavsson <bjorn@REDACTED> wrote:

> > On Fri, Feb 28, 2014 at 7:54 PM, tom kelly <ttom.kelly@REDACTED> wrote:
> >>
> >> -module(no_warn).
> >> -compile(export_all).
> >>
> >> -record(my_rec,{field1, field2}).
> >>
> >> my_function1(#my_rec{field1 = {tag, _}, field2 = f2}) -> 1;
> >> my_function1(#my_rec{field1 = {tag, 1}, field2 = f2}) -> 2; % line 7
> >> my_function1(_) -> 3.
> >>
> >> my_function2(#my_rec{field1 = {tag, _}}) -> 1;
> >> my_function2(#my_rec{field1 = {tag, 1}}) -> 2; % line11
> >> my_function2(_) -> 3.
>
> It is perhaps a documentation bug.
>
> Some warnings can be trusted 100 per cent.
> For example, if there are no warnigs for unused variables
> there are no unused variables.
>
> The absence of warnings for clauses that
> don't match does *not* mean that that all clauses
> can be matched. It merely means that
> the compiler (during optimisation and
> pattern matching compilation) did not find
> any clauses that obviously would not be
> reached.
>
>
Hi,

I'm not sure if I can agree with this. Non-matchable clauses should in my
opinion be marked as errors, because they expose a logical flaw (usually a
misordering of the clauses), and in this case I would have them be 100%
accurate.

Given clauses without guards like f(A)->ok; f(B)->ok. then I think we
should get a warning/error if B is the same as or more specific than A.
This probably applies when there are guards too. If that's the case, it
shouldn't be too difficult to implement.

What am I missing? Is it difficult just because it is done after some
optimization passes that might mess things up? Or is there some other issue?

best regards,
Vlad
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20140303/f007f333/attachment.htm>