From ianh Mon Nov 03 16:34:48 1980
>From 7619715 Fri Oct 31 22:08:08 1980 forwarded by ianh
ianh,
	pxp don't work???!!!! e.g.
    53                 |    repeat
    55              143.---|    endparaph := eof;
    56                     |    if not eof then 
    57                   89.---|    if nextchar in [' ', '	', '('] then 
    58                       86.---|    case nextchar of
    60                                      ' ', '	':
    60                            3.---|        readnextchar;
    61                                      '(':
    61                            2.---|        if input^ = '*' then 
    63                              112.---|        skipcomment
    63                         -110.---|        else 
    63                             -110.---|        endparaph := true
    63                             |    end
    63                    3.---|    else 
    64                        3.---|    endparaph := true
    64                     |until endparaph
	N.B. negative indices!!!


From ianh Mon Nov 10 17:01:23 1980
>From 7619715 Thu Nov 06 15:36:15 1980 forwarded by ianh
Ian H.,
	I sent a copy of the following to /usr/pascal but response was "can't
send...saved in /user1/6.643/6.643/7619715/dead.letter"?? so I am sending a copy 
to you:
	Re negative indices generated by pxp: the program was a lexical
analyser, namely "llexan.p", which takes data from a file containing a
plus program (the file may be standard input or a second argument) -
the one used was "fr". Both "llexan.p" and "fr" are in my directory(7619715
- class 6.643(Ken Robinson is lecturer)) and are now readable.
	"llexan.p" has a number of #include files - these are all readable also.	The commands used were pix -z %<fr, and pxp -z % with no intermediate
modifications.
					Greg Gamble



From root Sat Jun  6 14:33:00 1981
>From daemon Tue Jun  2 22:28:15 1981 forwarded by ianh
Subject: net mail from ianh at elecvax

>From daemon Tue Jun  2 22:29:19 1981 netmail from elecvax
Subject: net mail from root at basservax

>From root Tue Jun  2 22:22:32 1981 netmail from basservax
To: ianh
Subject: New pc, pi and px (from 4.1bsd)


We have obtained a prerelease of 4.1bsd and it contains
a new version of the above programs. Also (shock horror!)
a version of px written in C! Are you at all interested
in this. I noticed that some of your fixes to px and pi
had been applied to the new release, but I am no longer
up to date with your latest changes (fixes). Could it be
possible (if you are interested in examining the latest
stuff) for you to test it for your changes and others
and let me know what you think. It is presently stored
in sccs format in %s/cmd/pascal (4bsd), and I haven't yet got
the 4.1bsd from tape. You can find normal versions of
pc, pi and px in %s/cmd/p[cix].
				Thanks
				  Chris




From root Sat Jun  6 14:33:11 1981
>From daemon Tue Jun  2 22:28:27 1981 forwarded by ianh
Subject: net mail from ianh at elecvax

>From daemon Tue Jun  2 22:29:30 1981 netmail from elecvax
Subject: net mail from root at basservax

>From root Tue Jun  2 22:26:36 1981 netmail from basservax
To: ianh
Subject: previous mail corrections

The source for pc is in %s/cmd/pc0.
To get at it say

	newgrp pascal
	cd %s/cmd/xxx	where xxx is in pascal, pc0, pi or px.

		Chris




From richardg Sat Jun 20 14:44:43 1981
Subject: -s flag on pi

Pi with -s should only make the first 8 chars of identifiers significant,
as this is all that is guarranteed. I had a working pi (-s) program that
failed on pc because of this.
	Thanks Ian
	Richard

From kevin Thu Jul 30 18:55:25 1981
Subject: pascal

Not even God knows this time.....
		K

From root Thu Sep 10 14:38:02 1981
>From munro Wed Sep  9 16:41:49 1981 forwarded by ianh
Subject: px - use of loop control variable after loop.


	You will remmember the change you made some time ago,
whereby variables are initialised to nasty values, thus causing
people to find their mistakes sooner.

	I have just been tutoring a student who had been useing
a loop control variable after the loop, while it was still
officially undefined, in the belief that it would retain the
value it had before the loop. ( Other myths exist. )
My suggestion is that after the loop the control variable be
assigned a nasty value, so as to catch this kind of mistake
earlier.

						Munro.


