[m-rev.] Small bug in io.file.m

[Volker Wysk]

Am Mittwoch, dem 15.03.2023 um 14:06 +1100 schrieb Julien Fischer:
> Hi Volker,

Hi!

> 
> On Tue, 14 Mar 2023, Volker Wysk wrote:
> 
> > In io.file.m, line 839:
> > 
> >    is_error(Error, "can't find file type: ", MaybeIOError, !IO),
> > 
> > This doesn't include the file name in the IOError, like it's common.
> 
> In the current io module and its submodules, the generation of I/O error
> messages is entirely up to the I/O facilities of the target platform. In
> the above case, FileName is passed to file_type_2/6, which is the
> predicate that creates the value Error; some target will mention the
> file name (e.g. in the C# and Java grades), some don't seem to (e.g.
> Linux / glibc).

When doing a "ln -s aaa aaa" and apply file_type to it, I get:

can't find file type: Too many levels of symbolic links

When creating a file "aaa" with insufficient permission to read it, and try
to read it, I get:

aaa: Permission denied

In the first case, the file name isn't included, but in the second one, it
is. My target is Linux with an asm_fast grade.

I think, the file name is needed as part of an io.error message. It should
always be included.


Cheers,
Volker
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20230315/fc6c4e96/attachment.sig>