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

Re: Character styles

	
>   You will actually see the raw codes if you use Release 6 without the
>   compatibility system (not recommended), or if you examine the files
>   using some other operating system. [...]

>   If you actually have a case in which a perfectly-OK file is written from
>   7.0, and then it is read back into 7.0, and the codes are visible, then
>   there is a bug, and it should be reported.  It doesn't sound like
>   anything I've ever heard of.

There are some  ways to  actually get  confronted with  the raw codes
when running 6.x with 6-7 compatibility loaded. And you may will
even hear it! Simply try to hardcopy a file written in 7.0 from a
6.1 & 6-7 machine with :Hardcopy File from the command processor.
You will here the notification delivery ring the bell; An simple
example; that's the file's contents

D,#TD1PsT[Begin using 006 escapes](0 0 (NIL 0) (NIL NIL NIL) "CPTFONT")(1 0 (NIL 0) (NIL :BOLD-EXTENDED NIL) "CPTFONTB")(2 0 (NIL 0) (NIL :ITALIC NIL) "CPTFONTI")0;;; -*- Mode: Lisp; Package: User; Base: 10.; Fonts: CPTFONT, CPTFONTB, CPTFONTI; -*

-
;2;;========================================================================

0(DEFUN 1FOO0 ()
  (PRINT 'BAR)
  'BAZ)

;2;;========================================================================1EOF2.

Due to the header's format (using an epsilon, which is the font
escape) it throws into the debugger.

In Symbolics 3640 Release 6.1, Release-6-7 6.0, IP-TCP 29.13,
microcode TMC5-IO4-ST506-MIC 336, FEP 127, FEP0:>v127-lisp.flod(44),
FEP0:>v127-loaders.flod(44), FEP0:>v127-debug.flod(31),
FEP0:>v127-info.flod(44), GMD, on GRASPIN 3:


>>Error: Font -43 out of range.
While in the function (:METHOD SI:HARDCOPY-STREAM-FONT-MIXIN :SET-FONT)  SI:HARDCOPY-FROM-STREAM-1  SI:HARDCOPY-FROM-STREAM

(:METHOD SI:HARDCOPY-STREAM-FONT-MIXIN :SET-FONT):  (P.C. = 25)
   Arg 0 (SELF): #<LGP2-PAGE-BUFFERING-STREAM 16002062>
   Arg 1 (SYS:SELF-MAPPING-TABLE): #<Map to flavor SI:HARDCOPY-STREAM-FONT-MIXIN -- 7. IV's, 0. FL's 1200625>
   Arg 2 (SI:OPERATION): :SET-FONT
   Arg 3 (SI:NEW-FONT): -43
   --Defaulted args:--
   Arg 4 (FORMAT): :DEVICE

SI:HARDCOPY-FROM-STREAM-1:  (P.C. = 117)  (from SYS: HARDCOPY; HARDCOPY-46-25)
   Arg 0 (SI:STREAM): #<NFILE-INPUT-CHARACTER-STREAM "G1:>jc>foo.lisp" 16001714>
   Arg 1 (SI:HARDCOPY-STREAM): #<LGP2-PAGE-BUFFERING-STREAM 16002062>
   Arg 2 (SI:INTERPRET-FONT-CHANGES): T

SI:HARDCOPY-FROM-STREAM:  (P.C. = 241)
   Arg 0 (SI:STREAM): #<NFILE-INPUT-CHARACTER-STREAM "G1:>jc>foo.lisp" 16001714>
   Arg 1 (SI:DEVICE): #<LGP2-PRINTER LGP-2 43306403>
   Rest arg (SI:OPTIONS): (:BANNER-FILE-NAME "G1:>jc>foo.lisp.5" :BANNER-CREATION-DATE 2750355146 :PRESS-FONTS (("COURIER" "" 9)) :INTERPRET-FONT-CHANGES T :COPIES 1 ...)

That's  the  way  I really  like it.   It  is annoying  how fonts are
'supported' by Symbolics.  Starting back in the old days with Release
4.x,  where  I even  had to  provide a  reader macro  for the epsilon
escape to allow loading of source files written  with multiple fonts.
Making  hardcopies  of  such  files  was  and  is until  now not well
supported.  (e.g.  mapping from screen fonts to device  fonts:  there
is no LGP(1/2) counterpart for most of the screen fonts.)  

It's sad, but I quiet  accustomed to  it now  and don't  expect it to
change in the near future.  Each  new release  brings a  lot of bells
and whistles (status line time display for example in 7.0), but a lot
of necessary things won't be fixed.  For instance  the Print Spooler:
in 6.x it wouldn't check  that the  printer was  turned off meanwhile
and busy spooled onto a dead  serial line  (at least  it didn't waste
paper :-) and until  now, (including  7.0) it  doesn't recognize when
the  printer  is  switched  off  (or  connected  to  a  MAC) and then
reconnected to the lispm.  OK I'm a hacker and know it's "<select>-P,
get the LW into the debugger, c-9 c-N c-m-R Yes Quit"  and the damned
thing will work again,  but our  'high level'  end users  here.  It's
pretty annoying and the LW is switched a  lot between  our lispms and
our macs.  (Why not AppleTalk; its much faster  (~2.5MBit) than those
19200 bauds we have and it would work with the macs.   Ok,  ok, its a
foreign protocol, I know...)  

Resumee: There's still a lot to be done. And it is the small things
too, not the large ones only!

JC

P.S.  And they've taken hacks away in 7.0!   How  do I  check for new
mail?  I've used #'hacks:background-check-mail, but now it's gone and
I had to write  a function  of my  own.   And all  those visiting our
institute I'll send to  our SUNs  where they  may see  some real nice
demos since #'hacks:demo is gone. Back to work now, it's getting late...