Re: Unexpected result using file:read("l")

[Lars Müller <appgurulars@...>]

Using a global substitution paired with an anchor that limits it to the end of string feels somewhat odd to me (also shouldn't it be "\r" instead of "\\r" in the pattern?). As an alternative I'd suggest simply using matching:

line = line:match"(.*)\r?"

On 24.03.22 19:55, Oliver Kroth wrote:

> Hi,
>       
>       
>       opening a text file in non-binary mode (no 'b' in mode) on Linux
>       won't help you with a file that was written with CR+LF as line
>       endings.
>       
>       I use to snip off a terminating \r:
>       
>       
>       line = file:read('*l')
>       
>       if not line then break end
>       
>       line = line:gsub( '\\r$', '' )
>       
>       
>       Funny enough, the socket filter '*l' removes all '\r' characters,
>       thus it may indeed may come unexpected that file:read('*l') does
>       not.
>       
>       
>       If one has to read a lot of files line by line, one may consider
>       to define a iterator factory lines that crates iterators that snip
>       off the terminating '\r', like:
>       
>       
>       for line in lines( file ) do
>       
>        ...
>       
>       end
>       
>       
>       
>       Regards,
>       
>       
>       Oliver
>       
>       
>       Am 24.03.22 um 16:02 schrieb Francisco Olarte:
>       
> > Mitchell:
> >         
> >         
> >         On Thu, 24 Mar 2022 at 15:39, Mitchell
> >         <lists@triplequasar.com> wrote:
> >         
> > > I don't have a Windows box to test this
> > >           on, but if I recall correctly, opening the file in binary mode
> > >           on Windows should allow 'l' to also trim '\r'. I could be
> > >           misremembering though.
> > >           
> > >         
> >         It is the other way round. Opening a file in binary gives you
> >         
> >         uninterpreted bytes. The C runtime ( on which lua builds ) is
> >         supposed
> >         
> >         to translate whatever you os uses for end of line into a '\n'.
> >         So a
> >         
> >         file "ab\012bc\015de\015\012" is read the same in every OS in
> >         binary
> >         
> >         mode, but you may have problems (CP/M like CRLF, unixlike LF or
> >         
> >         Macintosh(classic, not osx) CR ).
> >         
> >         In default (text) mode they could be read differently. Also,
> >         normally
> >         
> >         it is not a good idea to use binary mode for text files, or
> >         viceversa.
> >         
> >         
> >         Francisco Olarte.
> >         
> >         
> >         
> >         
> >         
> >         
> >         
> > > Cheers,
> > >           
> > >           Mitchell
> > >           
> > >         
> >       
>