New REP
[1/17] from: lmecir:mbox:vol:cz at: 20-Jan-2003 19:20
Hi all,
I wrote a new version of my REP and posted it to http://www.rebolforces.com/~ladislav/rep.html
any comments/objections?
Regards
-L
[2/17] from: andreas::bolka::gmx::net at: 20-Jan-2003 22:28
Monday, January 20, 2003, 6:20:47 PM, Ladislav wrote:
> I wrote a new version of my REP and posted it to
> http://www.rebolforces.com/~ladislav/rep.html
> any comments/objections?
It looks like suggestion 21. is missing some words - it finishes with
an incomplete sentence.
Apart from this, it looks like good work and I can only wholeheartedly
agree and hope that some day, better sooner than later, those EPs will
be tackled by RT.
--
Best regards,
Andreas mailto:[andreas--bolka--gmx--net]
[3/17] from: petr:krenzelok:trz:cz at: 21-Jan-2003 8:29
Andreas Bolka wrote:
>Monday, January 20, 2003, 6:20:47 PM, Ladislav wrote:
>>I wrote a new version of my REP and posted it to
<<quoted lines omitted: 9>>
>agree and hope that some day, better sooner than later, those EPs will
>be tackled by RT.
yes, some of them seem so natural to me (well, the others I may not
understand properly :-), that I wonder, what is Carl's take on this, as
it seems to me, that it would make language implementation "cleaner"?
Ladislav, have you send it to feedback? Feedback accepts also
enhancement proposals IIRC
-pekr-
[4/17] from: lmecir:mbox:vol:cz at: 21-Jan-2003 8:53
Hi Andreas,
> It looks like suggestion 21. is missing some words - it finishes with
> an incomplete sentence.
>
> Apart from this, it looks like good work and I can only wholeheartedly
> agree and hope that some day, better sooner than later, those EPs will
> be tackled by RT.
Thanks, it looks that the upload didn't succeed completely. I will try to
upload again when I get home.
Regards
-L
[5/17] from: lmecir:mbox:vol:cz at: 22-Jan-2003 9:35
The corrected version submitted to
http://www.rebolforces.com/~ladislav/rep.html
A copy of this sent to feedback.
Regards
-L
[6/17] from: lmecir:mbox:vol:cz at: 22-Jan-2003 9:42
Hi Pekr,
you wrote:
> well, the others I may not
> understand properly :-)
If you woul need any explanations, just ask. I may have not explained all
matters as clearly as they should have been explained.
Regards
-L
[7/17] from: petr:krenzelok:trz:cz at: 22-Jan-2003 10:33
Ladislav Mecir wrote:
>Hi Pekr,
>you wrote:
<<quoted lines omitted: 6>>
>Regards
>-L
No, that was kind of an joke. Look - I can always look at stuff you and
Joel and Romano discuss here, I can understand what it eventually does,
but I would hardly produce any such code myself :-)
-pekr-
[8/17] from: rotenca:telvia:it at: 22-Jan-2003 14:28
Hi Ladislav,
i'll ad a thing:
out-of-range series! should not exist
One reason is that usual debug tools fail with a out-of-range series.
---
Ciao
Romano
[9/17] from: lmecir:mbox:vol:cz at: 22-Jan-2003 15:42
Hi Romano,
I am afraid, that it is too hard to implement such a feature.
----- Original Message -----
From: "Romano Paolo Tenca"
> i'll ad a thing:
>
> out-of-range series! should not exist
>
> One reason is that usual debug tools fail with a out-of-range series.
Which tools do you have in mind? I think, that it would be much easier to
correct the debugging tools.
Regards
-L
[10/17] from: g:santilli:tiscalinet:it at: 22-Jan-2003 20:39
Hi Ladislav,
On Wednesday, January 22, 2003, 3:42:05 PM, you wrote:
LM> I am afraid, that it is too hard to implement such a feature.
Actually I don't think so. Currently we have this:
>> a: [1 2 3]
== [1 2 3]
>> b: tail a
== []
>> clear a
== []
>> b
== ** Script Error: Out of range or past end
My (and Romano's) request (already sent to feedback) is that when
the interpreter encounters an out of range series, instead of
creating an error, it should reset the series to its tail.
This is possible now too:
>> b: tail b
== []
so I don't think it would be a problem.
Regards,
Gabriele.
--
Gabriele Santilli <[g--santilli--tiscalinet--it]> -- REBOL Programmer
Amigan -- AGI L'Aquila -- REB: http://web.tiscali.it/rebol/index.r
[11/17] from: rotenca:telvia:it at: 22-Jan-2003 23:45
Hi Ladislav,
> > One reason is that usual debug tools fail with a out-of-range series.
>
> Which tools do you have in mind? I think, that it would be much easier to
> correct the debugging tools.
>> clear head b: next [1]
== []
>> b
== ** Script Error: Out of range or past end
>> probe b
** Script Error: Out of range or past end
** Where: probe
** Near: print mold :value :value
>> mold b
** Script Error: Out of range or past end
** Near: mold b
>> trace on
Result: (unset)
>> b
Trace: b (word)
== ** Script Error: Out of range or past end
** Near: mold b
>> ? b
B is a block of value: ** Script Error: Out of range or past end
** Where: ?
** Near: mold :value
>> ?? b
** Script Error: Out of range or past end
** Where: rejoin
** Near: mold name: get name
If you remember the old thread on this subject, they have also not uniform
behaviour:
1) safe
2) error!
3) rebol crash
All this lead to very hard to find and debug bugs.
Uh! i discovered this:
>> mold/all b
== "#[block![]2]"
useful for sameness?
---
Ciao
Romano
[12/17] from: g:santilli:tiscalinet:it at: 23-Jan-2003 0:05
Hi Romano,
On Wednesday, January 22, 2003, 11:45:13 PM, you wrote:
>>> mold/all b
RPT> == "#[block![]2]"
Interesting!
>> #[block! [] 10]
== ** Script Error: Out of range or past end
>> #[block! [] -10]
== ** Script Error: Out of range or past end
Regards,
Gabriele.
--
Gabriele Santilli <[g--santilli--tiscalinet--it]> -- REBOL Programmer
Amigan -- AGI L'Aquila -- REB: http://web.tiscali.it/rebol/index.r
[13/17] from: lmecir:mbox:vol:cz at: 23-Jan-2003 14:27
Hi Romano,
> Uh! i discovered this:
>
> >> mold/all b
> == "#[block![]2]"
>
> useful for sameness?
this is a non-invasive method as opposed to my MINDEX? function,
nevertheless, I don't know, whether it is reliable?
Ciao
-L
[14/17] from: lmecir:mbox:vol:cz at: 23-Jan-2003 13:54
Hi,
<<Romano wrote>>
out-of-range series! should not exist
<</Romano>>
> LM> I am afraid, that it is too hard to implement such a feature.
<<Gabriele>>
> Actually I don't think so. Currently we have this:
> >> a: [1 2 3]
<<quoted lines omitted: 12>>
> == []
> so I don't think it would be a problem.
<</Gabriele>>
Our different POV's seem to be mainly a terminological difference. After
executing:
a: [1 2 3] ; == [1 2 3]
b: tail a ; == []
clear a ; == []
B will be an out-of-range SERIES! and the interpreter has got no means to
prevent that.
There is a possibility to "correct" B later, but the existence of B cannot
be prevented without too much work for the interpreter.
Moreover, there is a question that needs an answer: When should the
interpreter do the corrections?
See e.g. another code:
a: [1 2 3] ; == [1 2 3]
b: tail a ; == []
clear a ; == []
; b is out-of-range now
append a [4 5 6]
; b is "in range" again
I would not expect the interpreter to prevent B from being temporarily
out-of-range in this case, because it really might mean "too much work".
Regards
-L
[15/17] from: rotenca:telvia:it at: 23-Jan-2003 17:30
Hi Ladislav,
> Moreover, there is a question that needs an answer: When should the
> interpreter do the corrections?
<<quoted lines omitted: 7>>
> I would not expect the interpreter to prevent B from being temporarily
> out-of-range in this case, because it really might mean "too much work".
I propose that the INTERNAL and external Rebol function to get the index of a
series should be:
proposed-index?: func [s][min index? s 1 + length? head s]
so b is never out of range for any access to the series.
I don't think it is a big work for the interpreter.
Do you see some problems in this?
---
Ciao
Romano
[16/17] from: g:santilli:tiscalinet:it at: 23-Jan-2003 17:14
Hi Ladislav,
On Thursday, January 23, 2003, 1:54:05 PM, you wrote:
LM> Our different POV's seem to be mainly a terminological difference. After
Yes, indeed. I should have said: the "out of range series" error
should not exist. (I.e. out of range series can exist, but they
should not produce an error, but rather should be corrected.)
LM> a: [1 2 3] ; == [1 2 3]
LM> b: tail a ; == []
LM> clear a ; == []
LM> ; b is out-of-range now
LM> append a [4 5 6]
LM> ; b is "in range" again
LM> I would not expect the interpreter to prevent B from being temporarily
LM> out-of-range in this case, because it really might mean "too much work".
I agree.
Regards,
Gabriele.
--
Gabriele Santilli <[g--santilli--tiscalinet--it]> -- REBOL Programmer
Amigan -- AGI L'Aquila -- REB: http://web.tiscali.it/rebol/index.r
[17/17] from: lmecir:mbox:vol:cz at: 25-Jan-2003 11:58
Hi all,
a response from feedback:
Ok, thanks for posting that link. The timing is good, as we are assembling
the bug list for fixes. We might need to contact you about it later.
-REBOL Support
-------------------------------------------------
Feedback: #31m029440
From: [lmecir--mbox--vol--cz]
Contents:
I wrote a REP, that is available at
http://www.rebolforces.com/~ladislav/rep.html
It contains descriptions of some GC bugs, PARSE bugs and some enhancement
ideas.
Notes
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted