[erlang-patches] Correct ls/1 in c.erl

Bengt Kleberg bengt.kleberg@REDACTED
Wed May 22 12:26:55 CEST 2013


Thank you. That was better than I had hoped for.


bengt

On Wed, 2013-05-22 at 12:22 +0200, Björn Gustavsson wrote:
> We found the idea of improving c:ls/1 worthwhile.
> 
> 
> Therefore, we will not simply reject the patch.
> 
> 
> Instead we will use the implementation in the following
> commit:
> 
> 
> 
> https://github.com/bjorng/otp/commit/67d83e15fdee068217df7f14a32fb901d7d4d91b
> 
> 
> 
> in which I have credited Bengt for the idea and test case.
> 
> 
> We will test it a few days in our daily builds before merging
> it to maint.
> 
> 
> /Bjorn
> 
> 
> 
> 
> On Mon, May 20, 2013 at 4:40 PM, Bengt Kleberg
> <bengt.kleberg@REDACTED> wrote:
>         I do not know why file:list_dir/1 has failed with {error,
>         enoent}.
>         Or to be precise, I do not know what failure with {error,
>         enoent} means
>         for the different cases I get from directories, files, links,
>         special
>         files, none-existing files and the combination of read/no-read
>         access
>         for these.
>         
>         
>         
>         bengt
>         
>         On Mon, 2013-05-20 at 16:36 +0200, Björn Gustavsson wrote:
>         > I'll try to clarify my question:
>         >
>         >
>         >
>         > Why test for existence at all if
>         > file:list_dir/1 has already told you that the
>         > file exists?
>         >
>         >
>         > On Mon, May 20, 2013 at 4:22 PM, Bengt Kleberg
>         > <bengt.kleberg@REDACTED> wrote:
>         >
>         >
>         >         The filelib:is_dir/is_file/is_regular functions
>         confuse me. I
>         >         am unsure
>         >         what they will return for directories, files, links,
>         special
>         >         files,
>         >         none-existing files and the combination of
>         read/no-read access
>         >         for
>         >         these.
>         >
>         >         file:read_link_info/1 is more predictable.
>         >
>         >
>         >         bengt
>         >
>         >         On Mon, 2013-05-20 at 15:59 +0200, Björn Gustavsson
>         wrote:
>         >         > On Mon, May 20, 2013 at 1:45 PM, Bengt Kleberg
>         >         > <bengt.kleberg@REDACTED> wrote:
>         >         >         6> file:list_dir("nosuchfile").
>         >         >         {error,enotdir}
>         >         >
>         >         >         I think to use file:read_link_info/1 to
>         establish
>         >         the
>         >         >         existence of the
>         >         >         file/directory.
>         >         >
>         >         > Why?
>         >         >
>         >         >
>         >         > If file:list_dir(Name) returns {error,enotdir}, it
>         means
>         >         that
>         >         > there exists a file named Name (but it is not a
>         directory).
>         >         >
>         >         >
>         >         > If no such file exists, the return value will be
>         >         {error,enoent}.
>         >         >
>         >         >
>         >         > In what situation would it be useful to call
>         >         > file:read_link_info/1?
>         >         >
>         >         >
>         >         >
>         >         > --
>         >         > Björn Gustavsson, Erlang/OTP, Ericsson AB
>         >
>         >
>         >
>         >
>         >
>         >
>         > --
>         > Björn Gustavsson, Erlang/OTP, Ericsson AB
>         
>         
> 
> 
> 
> 
> -- 
> Björn Gustavsson, Erlang/OTP, Ericsson AB




More information about the erlang-patches mailing list