AltME groups: search

One could build a function that catches "face in more than one pane" 
errors that duplicates the face and ties the data values of the faces 
together so that they act like one face but are in fact two faces 
with identical values and therefor allows view to display the face 
in more than one pane...

I noticed a crash with rebol 1.2.107.3.1 on win2k with the script 
logo.r (tet program for AGG)  when I resize the windows ...

btw, i have a question: is there anyone using smtp:// directly?

romano: indeed, that was a bug, and should be fixed. (though, zero-size 
faces are debatable. vid uses them for sensors, so for compatibility 
that must work, however it doesn't seem such a great idea to me.)

anton: if a response does not have "-Gabriele" in it you can assume 
it was written by Carl :)

:)    #3437 is "Built", but on View 1.2.107 it still does not return 
an error.   func [a 'a :a][]

Gabriele that was not a bug, only the new version it is not backward 
compatible.

I have an undebugable bug !! It hapen with 1.2.104 and more, in a 
script that use Uniserve.

It's 100% reproducable, but with precise conditions. It's called 
by a timer (manually, I haven't get any error).

With a timer of 2 mins, it happens the second time. With different 
time it does not happen, or I haven't wait enough (but did not make 
to much test on that)..

Gabriele: #3402 is an outstanding bug when using /CUSTOM refinement. 
 An example is submitting a  form over http using REBOL. You need 
to specify both HEADER and POST. For example:

   read/custom url [HEADER [Content-Type: {multipart/form-data; boundary=---xyz} 
   ] POST {blah}]

The problem is that the handling of POST *overwrites* the Content-Type 
and Referer even though they may have been specified in HEADER. This 
logic bug is within the create-request function of Open in HTTP scheme.

This test generates a test email with the desired number of addresses 
then attempt to import it.

Romano did a lot of work to try to overcome this by trying to take 
recursion out of the code.

Important: 1.2.108 is out. Please, test it. There are a number of 
mezzanine changes that need testing to ensure we are not breaking 
anything.

note that this has been added as a new scheme, esmtp://. Now SEND 
uses esmtp:// and not smtp://. We have left smtp:// for compatibility 
reasons.

header parsing has been changed to support multiple headers (fixes 
#3090). this MIGHT break some code that does not expect getting a 
block instead of a string in the returned object. feedback is welcome.

other changes include: query on http port uses HEAD instead of GET, 
and fixed inform/timeout. please check out RAMBO to see a list of 
all the other bugs that have been fixed.

#3700: great! still doesn't work on images, but works fine with binary! 
and string!

some issues with string! usage in 'and 'xor 'or : 

1) string! op string! -> binary!, should returns a string! instead

2) a) string! op binary! -> error!, a) binary! op string! -> error! 
, 
should be allowed and do the same than:

a) string! op as-string binary! -> string! , b) binary! op as-binary 
string! -> binary!

#3694: seems fixed: my one-face-draw-textlist no longer needs a workaround 
- pen changes are applied.

Oops, talked a little too fast. Testing all datatypes combinations 
for 'and, 'xor, 'or, 

[ binary! | string! ] op [ logic! | number! | char! | tuple! ]  -> 
crash.

Vincent, that's right, we now have name collisions with older betas. 
Luckily the other day I wrote a script to rename all my rebview executables 
unambiguously. For interest, this is how I extracted the version 
number from each executable:

I have one modification wish for the FTP scheme: to add a List-Options 
variable like other FTP client programs have. This would allow LIST 
-la which allows hidden files such as Apache's .htaccess to be transferred 
using the scheme. At present these files cannot be transferred.

Gabrielle: I submitted a bug to Rambo that  was just a mistake in 
my code. Sorry. Please could you delete it - #-233.

Brett: please add that as a wish ticket. if it won't be in 1.3 it 
will be in 1.3.1. :)

Yes, Hooray for Romano's FTP!  Just remember to disable your previous 
Romano FTP patch (if you have one) from your user.r.  I added a version 
check around mine:

 if (system/version * 1.1.1.0.0) < 1.2.110.0.0 [ ; View1.2.110 has 
 Romano's FTP handler built in.

  do load-thru site/patch/ftp-romano-patch/ftp-patch53.r  ; the old 
  patch
	]

I'm not sure it's a good idea to not allow 0-sized faces. Is it a 
speed improvement that makes things so much faster ? The user code 
must HIDE the face instead of sizing it to zero, but HIDE may have 
other unintended side effects that are hard to workaround. eg. time 
events are stopped by HIDE:

If you are making a generic resizer, you can't just go modifying 
the code of the controlled widgets...

request : to really test new AGG/draw, we needs a command list. The 
examples helps, but there is now way to tell the presence and purpose 
of optional parameters. Not a full doc, just a simple one line per 
command list, ie : 'line 1st-point [pair!] 2nd-point [pair!] ... 
n-point [pair!]

#3723 : an old bug (it was at least in /Core 2.2.0) The 'list-dir 
function is missing a 'dir? test on argument. A line like :

if not dir? dir [throw make error! reduce ['access 'cannot-open dir]]

just before "read dir"  would do it (as 'dir? works with ftp:// too.)

not sure it's worth a report : from /View 1.2.10, there is a 'lic 
global word. It's probably a missing /local from a mezzanine related 
to license checking (it's between 'set-license and 'license.key in 
system/words)

hum is it normal I submited a new ticket to rambo and I have this

Yes the system allocates a negative number when you first submit 
a ticket. It seems to get allocated a positive number when it is 
viewed by a human.

I do not know, if this is comparable.

If I work with ODBC, pull out the network connector  and catch the 
error with try [], sooner or later I can observe a consecutive fault 
 either at a different part of the program doing something totally 
unconnected with  that  or a GC fault.
Either way the process is no longer able to work as expected.


In server processes I help myself in this case with automatic closing 
and starting a new instance of the program.

if the crash noted on -258 only happens on 1.2.1 then it's probably 
already fixed (a bug with read/lines was fixed), and the ticket should 
be dismissed.

This looks like a bug in the AGG compositing engine:

At least the black angle in the bottom right corner isn't what I 
would've expected from an uncolored edge ... Is this a known issue?

Ammon, rebol sound system needs overhaul anyway. But rebol implements 
sound/volume by just setting the host os system volume. A bit naughty, 
perhaps.

I think I have a null-result, because I can't reproduce the bug even 
on View 1.2.1

and Giueseppe also previously posted this bug ~15 months ago, which 
is well after View 1.2.1 was released.  Hmm. Maybe there was a buggy 
set of betas that I skipped over...

#3737 (call and file! datatype) : same problem on Win2k. The error 
message is right, it's a path problem - but if it works with string!, 
it should work with file!. A workaround :
call clean-path %prog.exe

Cal found a fun little bug with Select/Skip when using it on a string. 
 Here's the shortest code snippet that we've come up with to reproduce 
the problem.  Only copy and paste this into a console session that 
you don't mind killing cause it is going to lock up...  Tested on 
Win2K with the latest stable release, Command and recent betas...

select/skip {"<a><mm></mmmmf>"} "foo" 2

I think it is related to a bug in FIND:
find/skip "abcdefghij" "azzzz" 5 ; locks up
find/skip "abcdefghij" "abczz" 5 ; locks up
find/skip "abcdefghij" "abcde" 5 ; == "abcdefghij"  OK
find/skip "abcdefghijkl" "zzzzz" 5 ; == none  OK

There are a couple of /skip bugs in RAMBO already... Search for "find/skip".

So before to start to will load htose things you need to know what 
is the library dependencies. On linux you can see the MAKEFILE file 
related to one of the tutorial program integred to the GTK  package 
for example and see the -l*.a entries to have a complete idea of 
the library needed and the order of library dependencies

I'm agree with the idea that loading complicated heavy splitted and 
many dependent library libs GTK  popup up that we have to seek a 
better library loading system. Dependencies of libraries is a very 
weak point. Think of it ... First you have to load the librariries 
in the dependence chain in the right order and make a rebol script 
file to be able to treat in rebol script code the calls of function 
of those libraries (for example GTK sofware unsed commonly lots of 
function that are builded into many different librairies in the dependencie 
chain (like g_malloc() instead of malloc(), gint type instead  int 
type, gchar type intead of char, g_thread**() instead of  thread**() 
etc...)

I make a discussion group on this world called DLL.SO to collect 
all the ideas and organise a working group around the loading library 
topic

the point is that making bridge  betwen rebol script and librari 
in the way it is actually done is good for tiny simple library but 
very a tremendous work when it touch to heavy complicated library 
that intent to abstract from the os consept and give the same way 
to code on any OS/material. Mostly Opengl and GTK for example. Both 
libraries are heavy (lot of libs lot of struct lot of types lot of 
dependencies that needs a bridge too). For OPENGL you have two way 
to work or you make a OS based I/O and windowing system example gdi32.dll 
user32.dll for windows or xlib.so for Linux and  then exploite gl* 
function that are stored in the openGL.so/dll  or you use the related 
to opengl portable set of libraries to handle window drawing and 
Mouse/Keyboard events glut.dll/so. The fact that a librarie portable 
must be a library the abstract from the OS dependencies make them 
very complicated to handle. AS we don't have the same coding effort 
on library bridge coding than other language because many reasons 
in witch the fact that library loading is a Pro functionnality and 
maybe too because the system is not enought developped. It's easier 
to make a library bridge for a language when this language allows 
type creation and have based type in this language that feets with 
the one in C/C++

making library bridges is a very complicated thing for rebol coders...

If this fonctionnality remains payant we at most need to rethink 
internally the way to work or build an intermediate library loading 
system script that allows people to load the library without the 
needs of wrtiting by their howns the bridge or a tools that take 
a makefile of a sample program and the include file and generate 
a rebol script bridge for any library that must be loaded to use 
the wanted libraries thos bridges script files must then be supplyed 
as the libraries and the program script in rebol that exploit those 
libraries

A sutch tool will makes easier the verson change in library handling

Max was working on a generic Library loader..
But the project appears to be abandoned.
Last known URL is 404ing: www.rebol.it/~steel/libraries/

I found a very critical bug in the old and the new rebols

write/append does not give back an error when writing to a disk without 
enough space

just do this

insert/dup mem-st: make string! 100'000 "1" 100'000
counter: 0
forever [
    write/append %/A/mem-tst mem-st
    probe counter: counter + 1
]

then you will see that it runs without writing

I did add the write/append problem.

This just caused me some big problems, as I had 
some Rebol programs running     
since Sept. 24 2004 without interruption 
on a Win2000 server .(

But OSes often signal the program of such an error after they detect 
it. Do the REBOL internals react to such a signal?

From what I can determine the bug where select/skip on a string locks 
up the interpreter is related to bug #3327  My guess is that because 
/skip isn't skipping properly then it sometimes finds itself in a 
forever loop trying skip at the same point repeatedly.  Shall I submit 
this as a new bug references the expected relationship between these 
two or do we want to extend the description of #3327 to include this 
information as well?

#3755 : not a bug - it's just a syntax problem: "file: read %//server/share/file" 
doesn't work but "file: read %/server/share/file" does. {to-local-file 
%/server/share} and {to-rebol-file "\\server\share"} gives the expected 
results.

Ammon: i'd say submit a new one, and write that it may be related 
to 3327.

Vincent's description same for me on XP - I'll add that I get a "Problem 
uninstalling..." message when I try to uninstall. The rebol directory 
in docs and settings gets deleted but the registry key remains.

Will 'create-link be included in final 1.3? If yes, there's a number 
of issues:

- create-link %/c/foo.txt "d:\bar.lnk" will create a link to "current-path\c\foo.txt"
- create-link %/c/foo.txt %/d/bar.lnk don't work
and the big security hole:
    secure [file ask]

    create-link %anything.dat "c:\my-file-to-overwrite.exe" ; bypass 
    sandbox
RAMBO?

I mention this because I just noticed that 1.2.125 is considered 
a final beta, short of major bugs. This is one of those.

As the file handling has been fixed, someone can write an external 
installer that can do the trick if necessary, but at least moving 
the appropriate registry key would help a great deal.

This is a complex request - because all earlier versions of REBOL 
were installed the same way, as well as many other apps.

Installation code (and especially the self-installing kind like in 
REBOL) require a lot of testing -- not only by us in the dev team, 
but by all the beta users.

sounds like a bug to me.

ok, I'm writing a report.

maybe that double-slash should be allowed then? would ibreak something, 
like splitting a path?

They work in pairs.

One reads data from files, does some parsing, does some odbc and 
sends data via tcp.

the other reads data from tcp, writes some files and sends back an 
acknowledgement. 


one reads data from files, does some parsing and sends data via tcp.

The other reads data from tcp, does some parsing, does more odbc 
and sends the more data back as a reply.

Due to change in 'forall, yes, but some people (on French forum) 
are in trouble with that.
Does it worth a RAMBO ticket ???

(split-path mucks up when its target argument is at a series offset.)

About the ticket for 'forall. Cocinelle said it's just a missing 
[catch throw] args to the 'forall function!

Gabriele: I let you check that as I'm not "confortable" with this 
sort of things and maybe add the proposal  correction code to the 
ticket.

yes, it's a missing [throw] both in forall and in throw-on-error

lad: is that from the console? try secure quero from a launcyhed 
script (i.e. a script launched by the desktop)

launched scripts, instead, are kept into a sandbox

while I am posting this I notice the foreign characters under Linux 
is not yet solved. Using Debian Ubuntu with brazilian ABNT2 keyboard 
layout I am not able to input some usual characters in Portuguese 
such as ´a ´e ´i ´o `u `a ^o ~a .

Not sure if it goes here.. but Rebol 1.3 doesn't seem to like the 
bit on a Draw animation in the cookbook: http://www.rebol.net/cookbook/recipes/0047.html
 (specifically, the adding and subtracting in the draw block...)

Izkata, right, submitted a rambo ticket with fixed code for that 
cookbook entry.

Oh, I just wanted to make a suggestion for the green "RAMBO Statistics" 
box. I would like each number to be a link that is the rambo search 
query which gives the tickets in question. ie. "New tickets: 1"  
 - when you click on the "1",  you should get a page with the summary 
of the new ticket.

But before I did that, I wanted to see if I could help by writing 
the code to implement the change.

it is strange, that a set-path behaves this way

should behave like a set-word!?

i think that a path! with only 1 slot is a bad formed path

it already behaves like a set-word in case its length is higher than 
one, that is why I think, that the behaviour is natural

I needed to handle this:


foreach [path value] [a 1 o/b 2] [do reduce [to set-path! path value]]

The problem i think is the first slot has a different meaning from 
second, third and so on

It is a surprise for me that a set-path! == set-word!

that 'a in your example is a global word?

in my example yes, but I had a more complicated code originally

where it wasn't a global word

OTOH, it seems logical to allow expression:

    to set-path 'a-word

I think that give too importance to the fact that a set-path with 
1 slot is molded like a set-word (anc cannot be loaded as a set-path). 
But the first word of a set-path is not the same of a word! of a 
set-word!: the set-path word must be get and then selected.

The set-word word is like a second word in a path, the first hidden, 
implicit word! in a set-word is the context bound to the word.

but i agree: your proposal can  be useful, it is like to have to-set-word 
which does not change the context of a word

i do not think too much, but it is true it needs a separate test 
and handling

another experiment:

>> a: func [/b] [print b]
>> a/b: "OK"
true
== "OK"
>> get second second :a
== true

yarg... 'try was changed.. I just spent a half hour trying to figure 
out why a -short- peice of code would work!

>> try [to-time {Hi}]
== none
>> ? try
USAGE:
    TRY block

DESCRIPTION:

     Tries to DO a block and returns its value or an error. ;What happend 
     to the error?
     TRY is a native value.

to-time {hi} has returned 'none at least as far back as 1.2.1.
Maybe that is a bug :-)

IIRC depending on the desktop manager you use, you can open icons 
with a shell automatically.  and can ask the manager to hide the 
shell until the app bums out.  look into the icon's properties... 
This might also only be an option for .r file associated icons themselves 
(not rebol, but for script with an icon).

I used to do so for python scripts... it was really cool... no one 
saw the shell, until a crash occured, in which case all the prints 
where available for copy/paste.

This is a tad frustrating .. my emr is a single script ( from prerebol 
) and runs under windows.  Under Fedora, it does nothing :(

Just downloaded the lastest sdk ( I think for fedora ).  Try and 
run enface,and get a permission denied message :(

exports, I believe. type 'export' in a shell and see what happens.