World: r3wp

Pity .. I was wondering what the status of my submissions was.

Bug #3099: looks much better although I noticed this:
>> to-integer -2147483648.0
** Math Error: Math or number overflow
** Where: to-integer
** Near: to integer! :value

Bug #3403 Looks good to me on View 1.2.103

Bug #3436 looks good and solid by View 1.2.102

don't know if it's really a bug ...

>> greater-or-equal? 1.2.48.3.1 system/version
== false
>> system/version
== 1.2.102.3.1
>>

if it's a bug it could be cool to patch it too ...

>> help greater-or-equal?
USAGE:
    GREATER-OR-EQUAL? value1 value2

DESCRIPTION:

     Returns TRUE if the first value is greater than or equal to the second 
     value.
     GREATER-OR-EQUAL? is an action value.

ARGUMENTS:
     value1 -- (Type: any)
     value2 -- (Type: any)

So you have to invert arguments :
greater-or-equal? system/version 1.2.48.3.1

RE #3099:

>> to integer! #{80000000}
== -2147483648
>> - to integer! #{80000000}
== -2147483648

about: http://www.rebol.net/cgi-bin/rambo.r?id=3363&do you think 
FIND should stay as it is for compatibility? or it's better to change 
it?

Carl: "On negate bug: looks like a C compiler problem. I may be able 
to fix that with an explicit check for that value, but it will slow 
down the function by a lot. Is that worth it?"

what is that negate-bug? Ladislavs example? thats an overflow. 0 
is "positive", so there is one negative number more. and negative 
numbers count downward, all bits on is highest possible. the conversion 
is: invert all bits, add 1. gives #{7FF..}, + 1 #{800..}

find - hmm. till 'as-string it was handy with big data. now we can 
rewrite it, but detecting that use is tricky. but its surely a surprise. 
maybe we can forbid find without /case for a while with binary? gibing 
some usefull warning.

i have a wish: integrating 'comment in some more dialects. parse, 
layout come to mind. could then be used for embedded documentation. 
what do you think?

I second the request for comment!

Pun aside, of course.

a lot of bugs fixed: http://www.rebol.net/cgi-bin/rambo.r?sort=1&limit=2&cmd=Search&id=&pattern=&subject-only=true

3662 parse/lines second arg: how to split into lines and apply the 
rule to each line?

 a grep:   parse/lines string-with-list-of-file-names [ set file thru 
 ".r" end ( ?? file ) ]

btw, can we discuss auto-dequoting again? maybe together with /all? 
i don't like such surprises:
!>> parse {
a
b
} "^/" 
== "a" "b"

correction, result is == ["a" "b"]
my "do from editor molded /only..

#3659 to-idate: i would like that fix native. that seconds are always 
included, even if 0. makes little clocks much nicer. and if one really 
dont want seconds, copy/part "00:00:00" 5 is easy.

(means i would fix form/mold, not to-idate)

Tom: can you please check if 3676 still applies to 1.2.104?

#3256: it doesn't crash here. does it crash for any of you with 1.2.104?

Brett: do you have the example code for #3420? is it still broken?

#3463 : still crash with 1.2.104 - the problem is not construct/with, 
it's using an object containing a recursive block as prototype :
b: copy [] append/only b b 
an-object: make object! [a: b]
make an-object [] ; crash

like
  b: copy [] append/only b b copy/deep b

make object makes deep copies of all blocks. sure that crashes. crash 
should should be a softer?

Gabriele: Yes it stills seems wrong to me. Compare the last two columns 
of the display given by this:

 http://www.codeconscious.com/rebol/mlscripts/multiply-effect-test.r
(Tested on REBOL/View 1.2.105.3.1 14-May-2005 Core 2.6.1)

thanks brett. added link to ticket.

Anamonitor fails on the last beta, i do not know why:
** Script Error: Invalid graphics face object
** Where: view
** Near: show scr-face
if new [do-events]

could it be regarded a bug in interpreter parser?
>> ; test
>> ; test "aaaa
** Syntax Error: Missing " at ; test "aaaa

Gabriele: where might I find  1.2.104 for Solaris  (sans X11)

favourite bug - the one where an encapped application is unable to 
launch another instance after it does a system call

But for 1.3 .. I have to think about this.

Particles & Bubbles demo are a true CPU eater under linux... refresh 
of the GUI desktop under linux drops very quickly when 2 times running 
Particles demo...

rebview1210242 does not kill child processes when executed like 
$ ./rebview1210242  &
[1] 9203
$ 
[1]+  Stopped                 ./rebview1210242

$ ps a
  PID TTY      STAT   TIME COMMAND

 9203 pts/2    T      0:00 ./rebview1210242
 9204 pts/2    T      0:00 ./rebview1210242
 9205 pts/2    T      0:00 ./rebview1210242

$ kill -9 9203 9204 9205

9229  0.5  0.8  7780 5316 pts/2    T    10:29   0:00 ./rebview1210242

     9230  0.0  0.0     0    0 pts/2    Z    10:29   0:00 [rebview1210242] 
     <defunct>

     9231  0.0  0.0     0    0 pts/2    Z    10:29   0:00 [rebview1210242] 
     <defunct>

that was after doing a "fg job-id" and kill parent...

View Test folder
 : I guess it's under IOS/Developper, isn't it ?

No its a new folder in View/Desktop

Oh, OK. Did not run Desktop this last days (months ?) ;-)

Romano: that should be fixed, that error came out whenever there 
was a face with size = 0x0

pekr: that happens in 1.2.1 too, and i think it's a console quirk.

tom: i meant, if you could repeat the test on windows. if it's solaris 
only, we'll need to wait for that then.

yeah, I know ... just curious .... I tried to paste some script, 
and those comments did not allow me to, as it contained " char, which 
was not closed, but it was comment section, so it should be imo ignored 
...

in such case rebol console should look after line-break imo?

#3656 "copy/part open/direct/read/binary what-dir 100000 ;=CRASH"
Looks nice and stable from View 1.2.104.3.1

Graham: i found out that you can search any field with the "Search 
for:" input box. so if you input "Graham" there, it will show all 
tickets from you.

With 1.2.107 (and may be others), when I launch a script from the 
desktop (clicking its icon) the new interpreter open a security request 
to read %rebol.r in the directory where the interpreter is (ie: %/c/rebol/view/).
Something wrong with 'secure setting ?