geomol / world Goto Github PK

[btiffin@home World]$ ./world
Loading Cortex... Done
World/Cortex 0  Copyright (c) 2009-2011 John Niclasen, NicomSoft
build: Dec 23 2011 10:09:47
w> system/options/format/real/digits: 32
== 32
w> 1 / 3
== 0.33333333333333331482961625624739
w> system/options/format/real/digits: 48
== 48
w> 1 / 3
*** stack smashing detected ***: ./world terminated
======= Backtrace: =========

...

== 0.333333333333333314829616256247390992939472198486Aborted (core dumped)

Somewhere between 32 and 48, my GNU/Linux world cracks. Don't care about digit support that high, but access to the real/digits field may require some fencing.

CTRL-C crashes World on Windows in some cases

Example:
w> wait 10 ;then press CTRL-C
or
w> p: open udp://:1234 first p ;then press CTRL-C

sqlab: the automatic closing of open brackets may be handy when typing, but it gets in the way copying from the clipboard

== -5:00
w>n: now
== -5:00
w> n/time
== 4:03:45
w> call "set"
...
LANG=en_US.utf8
LANGUAGE=
...

I don't see any other LC_ or locale settings

Currently, starting the world binary with a script as an argument just ignores the script and gets you to the prompt:

$ ./world_linux hello.w 
Loading Cortex... Done
World/Cortex 0  Copyright (c) 2009-2011 John Niclasen, NicomSoft
build: Dec  4 2011 09:53:36
w>

It would be nice if a world script could be executed directly that way.

Implicit compilation of blocks (via do) works fine:

w> x: 42 
== 42
w> c1: [x]
== [x]
w> do c1
== 42

Explicit compilation, however, does not:

w> c2: [x]
== [x]
w> compile c2
== [x]
w> compiled? c2 
== true
w> do c2
== [x] ;; expected: 42

Latest git pull, 2011-12-22 try.w is missing from test/functions/control

w> help system system is a context! of value: ** Script error: attempt to compile a value of type string! ** Near: length? series

Currently pressing ctrl-d inputs the EOT character (u+0004). It would be nice if ctrl-d just exits the interpreter instead (as if calling q).

For scripts to be invoked from the command-line, a method to access command-line parameters is required (such as wrapping argv (where available) as a block! in system/options/args).

Geomol, Can you add "words" to list out the dictionary? I know REBOL uses "what" but being a Forther, words makes far more sense and I never got with the what. Having said that, what would suffice if you are going for compatibility.

$ ./world_linux -q hello.w 
Loading Cortex... Done
hello, world!

In quiet mode, the "Loading Cortex... Done" output should be suppressed.

I guess that would require exposing the information that World was started in quiet mode somewhere. REBOL's system/options/quiet would be one possibility. (A script-local argument passed from the interpreter to cortex, another.)

Currently, World prints the whole gamut of startup banners even when executing a script:

$ ./world_linux hello.w 
Loading Cortex... Done
World/Cortex 0  Copyright (c) 2009-2011 John Niclasen, NicomSoft
build: Dec 22 2011 09:52:46
hello, world!
$

For executing scripts, this excessive verbosity is rarely useful. Instead, executing a script should default to quiet mode (i.e. as if -q was passed to the World interpreter).

In rebol:

>> use [x][x: 5 y: func [r][r + x]]
>> y 6
== 11

In World:

Loading Cortex... Done
World/Cortex 0  (c) John Niclasen, NicomSoft
build: Apr  3 2015 19:29:22
w> do %rebol.w
Loading REBOL Extension... Done
w> use [x][x: 5 y: func [r][r + x]]
w> y 5
** Script error: y has no value
** Near: y 5

The Red programming language(https://github.com/red/red) alse inspired by Rebol with native-code compiler,can derectly generates EXE

w> help 'abc ** Script error: abc has no value ** Near: path! <> type? :word w> a: 'abc == abc w> help a a is a word! of value: abc w>

Loading Cortex... Done
World/Cortex 0 Copyright (c) 2009-2011 John Niclasen, NicomSoft
build: Dec 11 2011 21:28:07

git pull on Dec 13, 22:00 Ottawa time.

Missing the second colon borks.

w> lp: open tcp://8080
Segmentation fault (core dumped)
[btiffin@home World]$ ./world
Loading Cortex... Done
World/Cortex 0  Copyright (c) 2009-2011 John Niclasen, NicomSoft
build: Dec 11 2011 21:28:07
w> lp: open tcp://:8080
w> call "uname -a"
Linux home 2.6.35.14-106.fc14.x86_64 #1 SMP Wed Nov 23 13:07:52 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux
== 0
w>

Let's start with a simple code block:

w> x: 99 
== 99
w> c: [x]
== [x]
w> do c
== 99 ;; Fine
w> compiled? c
== true

If we now change the block's content, the compiled flag is not reset:

w> append c [* 2]
== [x * 2]
w> c
== [x * 2]
w> compiled? c
== true

Now it could be that it's just marked as "dirty" and recompiled the next time it's evaluated. But no, evaluating the block again returns an unexpected result:

w> do c
== 99 ;; expected: 198