[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: binmode

From: Akim Demaille
Subject: Re: binmode
Date: 19 Sep 2001 16:57:42 +0200
User-agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.4 (Artificial Intelligence)

>>>>> "Tim" == Tim Van Holder <address@hidden> writes:

Tim> binmode $fh->SUPER;
>> It might be right, I just don't understand it!  Why not binmode
>> $fh?

Tim> I don't know - I was making it up as I went along.  It's quite
Tim> possible binmode $fh would be enough; it's just that OO Perl is
Tim> somewhat outside my current Perl knowledge.  

Actually, just forget about OO here, it's merely a set of values ($<
@, %, etc.) bound to that ``typeglob''.  binmode looks at its file
handle part, there is nothing related to inheritance etc.

Hence, IMHO, SUPER is totally superfluous: there is no difference.

Tim> Of course, this second draft also dropped too much; it would
Tim> cause _every_ file to be in binary mode, and we don't want that
Tim> for input files (besides the CR issus, putting stdin in binmode
Tim> locks up the console under DJGPP).  

Maybe you want to check whether open is asked to open an output or
input file.

Tim> Also, autom4te's main work is done through a call to system()
Tim> which uses regular redirection, so that output file (which seems
Tim> to be the main output file) will have DOSish line endings under
Tim> DOS (which makes the resulting configure scripts unrunnable under
Tim> most Unices).  I don't see a good way out of that (unless the
Tim> system() is changed to an XFile pipe, but I'm not sure whether
Tim> that's worth it).

There are certainly means to know a posteriori whether a file is used
as output.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]