My memory is clearly that B did not use line numbers. However, it
did report line numbers for errors. The 'ed' command had an 'n'
command that would number the lines of the file -- I think this was a
result of an early customer for Unix being the patent department -- at
the time, patents had to have exactly 50 lines on each page with no
blanks and the lines had to be numbered... When writing about a
program, it was handy to include line numbers so the document could
refer more easily to lines in the program, but they weren't fed to the
compiler.
Steve
(Funny how this conversation makes me feel like one of a few surviving
members of a tribe speaking a soon to be dead language...)
----- Original Message -----
From:
"Alec Muffett" <alec.muffett(a)gmail.com>
To:
"Random832" <random832(a)fastmail.com>
Cc:
<tuhs(a)minnie.tuhs.org>
Sent:
Sun, 9 Apr 2017 12:03:13 +0100
Subject:
Re: [TUHS] Non-US Unix Activities
On 9 April 2017 at 07:34, Random832 <random832(a)fastmail.com [1]>
wrote:
I've posted a few images at
https://dropsafe.crypticide.com/article/12714 [2]
Username "iy7" spattered around the text,
was Alan Cox.
I suppose the most notable thing is use of asterisk, where C coders
would
expect backslash? And the lack of types.
Are those BASIC-style line numbers?
I honestly can't remember; I think it was the Aberystwyth sysadmin
(Rob Ash) who write a B prettifier (the cited BTIDY) but I can't
remember whether B as a language used/ignored line numbers, or if this
was just a prettyprint thing.
There are mentions of line numbers
in
https://www.bell-labs.com/usr/dmr/www/btut.pdf [3] so perhaps they
are a compiler convenience for diagnostics?
-a
--
http://dropsafe.crypticide.com/aboutalecm [4]
Links:
------
[1] mailto:random832@fastmail.com
[2]
https://dropsafe.crypticide.com/article/12714
[3]
https://www.bell-labs.com/usr/dmr/www/btut.pdf
[4]
http://dropsafe.crypticide.com/aboutalecm