File:
[gforth] /
gforth /
stuff.fs
Revision
1.29:
download - view:
text,
annotated -
select for diffs
Sun Jan 26 20:56:37 2003 UTC (21 years, 3 months ago) by
anton
Branches:
MAIN
CVS tags:
HEAD
use AC_SYS_LARGEFILE instead of declaring this stuff unconditionally
moved memcasecmp() into support.c
eliminated -TEXT (all uses replaced, except within prim)
converted CAPSCOMP and -TRAILING into high-level words
folded C code for more complex primitives into support.c
Fliterals are now decompiled ok
f.rdp etc. documentation changes
added test cases for SEARCH
\ miscelleneous words
\ Copyright (C) 1996,1997,1998,2000 Free Software Foundation, Inc.
\ This file is part of Gforth.
\ Gforth is free software; you can redistribute it and/or
\ modify it under the terms of the GNU General Public License
\ as published by the Free Software Foundation; either version 2
\ of the License, or (at your option) any later version.
\ This program is distributed in the hope that it will be useful,
\ but WITHOUT ANY WARRANTY; without even the implied warranty of
\ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
\ GNU General Public License for more details.
\ You should have received a copy of the GNU General Public License
\ along with this program; if not, write to the Free Software
\ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
require glocals.fs
' require alias needs ( ... "name" -- ... ) \ gforth
\G An alias for @code{require}; exists on other systems (e.g., Win32Forth).
\ needs is an F-PC name. we will probably switch to 'needs' in the future
\ a little more compiler security
\ currently not used by Gforth, but maybe by add-ons e.g., the 486asm
AUser CSP
: !CSP ( -- )
sp@ csp ! ;
: ?CSP ( -- )
sp@ csp @ <> -22 and throw ;
\ DMIN and DMAX
: dmin ( d1 d2 -- d ) \ double d-min
2over 2over d> IF 2swap THEN 2drop ;
: dmax ( d1 d2 -- d ) \ double d-max
2over 2over d< IF 2swap THEN 2drop ;
\ shell commands
0 Value $? ( -- n ) \ gforth dollar-question
\G @code{Value} -- the exit status returned by the most recently executed
\G @code{system} command.
: system ( c-addr u -- ) \ gforth
\G Pass the string specified by @var{c-addr u} to the host operating system
\G for execution in a sub-shell.
(system) throw TO $? ;
: sh ( "..." -- ) \ gforth
\G Parse a string and use @code{system} to pass it to the host
\G operating system for execution in a sub-shell.
'# parse cr system ;
\ stuff
: ]L ( compilation: n -- ; run-time: -- n ) \ gforth
\G equivalent to @code{] literal}
] postpone literal ;
[ifundef] in-dictionary?
: in-dictionary? ( x -- f )
forthstart dictionary-end within ;
[endif]
: in-return-stack? ( addr -- f )
rp0 @ swap - [ forthstart 6 cells + ]L @ u< ;
\ const-does>
: compile-literals ( w*u u -- ; run-time: -- w*u ) recursive
\ compile u literals, starting with the bottommost one
?dup-if
swap >r 1- compile-literals
r> POSTPONE literal
endif ;
: compile-fliterals ( r*u u -- ; run-time: -- w*u ) recursive
\ compile u fliterals, starting with the bottommost one
?dup-if
{ F: r } 1- compile-fliterals
r POSTPONE fliteral
endif ;
: (const-does>) ( w*uw r*ur uw ur target "name" -- )
\ define a colon definition "name" containing w*uw r*ur as
\ literals and a call to target.
{ uw ur target }
header docol: cfa, \ start colon def without stack junk
ur compile-fliterals uw compile-literals
target compile, POSTPONE exit reveal ;
: const-does> ( run-time: w*uw r*ur uw ur "name" -- )
\G Defines @var{name} and returns.@sp 0
\G @var{name} execution: pushes @var{w*uw r*ur}, then performs the
\G code following the @code{const-does>}.
here >r 0 POSTPONE literal
POSTPONE (const-does>)
POSTPONE ;
noname : POSTPONE rdrop
lastxt r> cell+ ! \ patch the literal
; immediate
\ !! rewrite slurp-file using slurp-fid
: slurp-file ( c-addr1 u1 -- c-addr2 u2 )
\G @var{c-addr1 u1} is the filename, @var{c-addr2 u2} is the file's contents
r/o bin open-file throw >r
r@ file-size throw abort" file too large"
dup allocate throw swap
2dup r@ read-file throw over <> abort" could not read whole file"
r> close-file throw ;
: slurp-fid { fid -- addr u }
\G @var{addr u} is the content of the file @var{fid}
0 0 begin ( awhole uwhole )
dup 1024 + dup >r extend-mem ( anew awhole uwhole R: unew )
rot r@ fid read-file throw ( awhole uwhole uread R: unew )
r> 2dup =
while ( awhole uwhole uread unew )
2drop
repeat
- + dup >r resize throw r> ;
\ ]] ... [[
: compile-literal ( n -- )
postpone literal ;
: [[ ( -- )
\G switch from postpone state to compile state
\ this is only a marker; it is never really interpreted
compile-only-error ; immediate
: postponer ( c-addr u -- )
2dup find-name dup if ( c-addr u nt )
nip nip name>comp
2dup [comp'] [[ d= if
2drop ['] compiler is parser
else
postpone,
endif
else
drop
2dup snumber? dup if
0> IF
swap postpone literal postpone compile-literal
THEN
postpone Literal postpone compile-literal
2drop
ELSE
drop no.extensions
THEN
then ;
: ]] ( -- )
\ switch into postpone state
['] postponer is parser state on ; immediate restrict
\ f.rdp
: push-right ( c-addr u1 u2 cfill -- )
\ move string at c-addr u1 right by u2 chars (without exceeding
\ the original bound); fill the gap with cfill
>r over min dup >r rot dup >r ( u1 u2 c-addr R: cfill u2 c-addr )
dup 2swap /string cmove>
r> r> r> fill ;
: f>buf-rdp-try { f: rf c-addr ur nd up um1 -- um2 }
\ um1 is the mantissa length to try, um2 is the actual mantissa length
c-addr ur um1 /string '0 fill
rf c-addr um1 represent if { nexp fsign }
nd nexp + up >=
ur nd - 1- dup { beforep } fsign + nexp 0 max >= and if
\ fixed-point notation
c-addr ur beforep nexp - dup { befored } '0 push-right
c-addr beforep 1- befored min dup { beforez } 0 max bl fill
fsign if
'- c-addr beforez 1- 0 max + c!
endif
c-addr ur beforep /string 1 '. push-right
nexp nd +
else \ exponential notation
c-addr ur 1 /string 1 '. push-right
fsign if
c-addr ur 1 '- push-right
endif
nexp 1- s>d tuck dabs <<# #s rot sign 'E hold #> { explen }
ur explen - 1- fsign + { mantlen }
mantlen 0< if \ exponent too large
drop c-addr ur '* fill
else
c-addr ur + 0 explen negate /string move
endif
#>> mantlen
endif
else \ inf or nan
if \ negative
c-addr ur 1 '- push-right
endif
drop ur
\ !! align in some way?
endif
1 max ur min ;
: f>buf-rdp ( rf c-addr +nr +nd +np -- ) \ gforth
\G Convert @i{rf} into a string at @i{c-addr nr}. The conversion
\G rules and the meanings of @i{nr nd np} are the same as for
\G @code{f.rdp}.
\ first, get the mantissa length, then convert for real. The
\ mantissa length is wrong in a few cases because of different
\ rounding; In most cases this does not matter, because the
\ mantissa is shorter than expected and the final digits are 0;
\ but in a few cases the mantissa gets longer. Then it is
\ conceivable that you will see a result that is rounded too much.
\ However, I have not been able to construct an example where this
\ leads to an unexpected result.
swap 0 max swap 0 max
fdup 2over 2over 2 pick f>buf-rdp-try f>buf-rdp-try drop ;
: f>str-rdp ( rf +nr +nd +np -- c-addr nr ) \ gforth
\G Convert @i{rf} into a string at @i{c-addr nr}. The conversion
\G rules and the meanings of @i{nr +nd np} are the same as for
\G @code{f.rdp}. The result in in the pictured numeric output buffer
\G and will be destroyed by anything destroying that buffer.
rot holdptr @ 1- 0 rot negate /string ( rf +nd np c-addr nr )
over holdbuf u< -&17 and throw
2tuck 2>r f>buf-rdp 2r> ;
: f.rdp ( rf +nr +nd +np -- ) \ gforth
\G Print float @i{rf} formatted. The total width of the output is
\G @i{nr}, the number of digits after the decimal point is @i{+nd},
\G the minimum number of significant digits for fixed-point notation
\G is @i{np}. @code{Set-precision} has no effect on @code{f.rdp}.
\G Fixed-point notation is used if the number of siginicant digits
\G would be at least @i{np} and if the number of digits before the
\G decimal point would fit. If fixed-point notation is not used,
\G exponential notation is used, and if that does not fit, asterisks
\G are printed. We recommend using @i{nr}>=7 to avoid the risk of
\G numbers not fitting at all. We recommend @i{nr}>=@i{np}+5 to avoid
\G cases where @code{f.rdp} switches to exponential notation because
\G fixed-point notation would have too few significant digits, yet
\G exponential notation offers fewer significant digits. We recommend
\G @i{nr}>=@i{nd}+2, if you want to have fixed-point notation for some
\G numbers. We recommend @i{np}>@i{nr}, if you want to have
\G exponential notation for all numbers.
f>str-rdp type ;
0 [if]
: testx ( rf ur nd up -- )
'| emit f.rdp ;
: test ( -- )
-0.123456789123456789e-20
40 0 ?do
cr
fdup 7 3 1 testx
fdup 7 3 4 testx
fdup 7 3 0 testx
fdup 7 7 1 testx
fdup 7 5 1 testx
fdup 7 0 2 testx
fdup 5 2 1 testx
fdup 4 2 1 testx
fdup 18 8 5 testx
'| emit
10e f*
loop ;
[then]
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>