osd@hou2d.UUCP (Orlando Sotomayor-Diaz) (03/10/85)
From: Orlando Sotomayor-Diaz (The Moderator) <cbosgd!std-c>
mod.std.c Digest Sat, 9 Mar 85 Volume 4 : Issue 9
Today's Topics:
\e and \l
more on "long int file pointers considered harmful"
trigraphs
----------------------------------------------------------------------
Date: Fri, 8 Mar 85 20:29:19 PST
From: ihnp4!amdahl!sjl (Steve Langdon)
Subject: \e and \l
To: ihnp4!cbosgd!std-c
I believe that the only strong argument against these constructs is that the
committee should not add new features to the language. If my memory is correct
the escape character is used in ISO 2022 (the major international character set
standard) to shift into alternate sets of characters. The linefeed character
may be used to indicate a new line, but is not required to have this function.
I can see legitimate uses for both of these constructs. The arguments against
them seem to be primarily aimed at possible misuse, or based on an unvoiced
objection to two escapes sequences generating the same character (ie. \n == \l).
When discussing C it appears inappropriate to argue on the basis that a feature
can be misused. The alternatives available in the language as it exists today
are all less portable than these new alternatives. The major question that
remains is whether the added portability is worth enlarging the language.
Steve Langdon ...!{ihnp4,hplabs,sun,nsc}!amdahl!sjl (408)746-6970
------------------------------
Date: Fri, 8 Mar 85 21:20:36 est
From: decvax!minow
Subject: more on "long int file pointers considered harmful"
To: ima!stevel@cbosgd.ATT.UUCP, std-c@cbosgd.ATT.UUCP
In mod.std-c, Steve Ludlum (ima!stevel) questions my comment that
a long can't contain enough information to reposition a file,
wondering if I'm using a "tiny C" with short longs.
The problem isn't that some C's have short longs, but that a long
(of 32 bits) isn't big enough to hold the underlying operating
system's file pointer. Disk files can hold an awful lot of information
and are getting bigger even as we speak. If my math is correct,
a 32-bit long int cannot address a 500,000 block (@512 bytes) file.
Furthermore, it seems to me that hardwiring file pointers to a
particular datatype of a particular size is short-sighted.
Martin Minow
decvax!minow
------------------------------
Date: Sat, 9 Mar 85 09:31:41 pst
From: Eugene D. Brooks III <ucbvax!brooks@lll-crg.ARPA>
Subject: trigraphs
To: cbosgd!std-c@BERKELEY
The use of trigraphs of the form ??char to enter C on terminals that don't
support full ascii seems a bit much. Terminals that don't support full ascii
are rapidly going to the junk heap. We got rid of the ASR33's here at LLNL
a long time ago. The use of ? and the addition of an escape \? to get ?
seems like a portability disaster. The character ? is a common one to use
when a program prompts for information. If this is eventually added to
the ANSI standard it seems that it ought to be added as a compiler OPTION
so that its people with ascii terminals, by far the majority, dont have
to deal with this nonsense.
------------------------------
End of mod.std.c Digest - Sat, 9 Mar 85 19:57:19 EST
******************************
USENET -> posting only through cbosgd!std-c.
ARPA -> ... through cbosgd!std-c@BERKELEY.ARPA (NOT to INFO-C)
In all cases, you may also reply to the author(s) above.