Yes, a good point.
As I said, I have a need for this facility occasionally, but not for
production code. Most of my programming is recreational problem-solving.
I think I used the block comment method a week or so ago in solving a numerical
Times Listener crossword, number 4673, 21 August
Ian's idea is even better: don't.
On Tue, Aug 31, 2021 at 11:05 PM Ian Clark wrote:
> Ruda wrote
> > if I load a (longer) script, I sometimes want to ignore a part of it
>
> I take Ruda's use of "sometimes" to mean he wants the proposed feature to
> be switchable.
>
> To anyone wanting to dabble
Ruda wrote
> if I load a (longer) script, I sometimes want to ignore a part of it
I take Ruda's use of "sometimes" to mean he wants the proposed feature to
be switchable.
To anyone wanting to dabble with such a feature, I have a word of advice.
Don't.
If there's part of your script you want to om
I'm starting to like Mike Day's suggestion more, at this point, though it
adds a small layer of necessary software for it to work.
On Tue, Aug 31, 2021 at 3:38 PM Raul Miller wrote:
> Something like
>
> Note''
> }}
> )
>
> would break this approach.
>
> A fix would be to change to something like
Something like
Note''
}}
)
would break this approach.
A fix would be to change to something like
Note''
NB.}}
)
Or, like
Note''
{{
}}
)
Or, like
Note''
'}}'
)
Or, if you are using )n, like
Note''
'}}
)
FYI,
--
Raul
On Tue, Aug 31, 2021 at 2:46 PM Devon McCormick wrote:
>
> > @Devon:
> @Devon: can you get this error in the 9.03 beta? If so, can you shorten
> it to a replicable bug report?
I will take a look. I don't actually know if that was the problem but
something broke when I tried to comment out most of my large script using
double curlies.
On Tue, Aug 31, 2021 at 12:
Actually Raul,
JOD dump scripts see: https://github.com/bakerjd99/joddumps
are very large J scripts that serialize code documentation databases. I've
loaded dump scripts > 100 MBs without problems. Over the years
I've hit various J system limits with these scripts that Roger and now
Henry have
You are right. And I can't produce the problem either.
@Devon: can you get this error in the 9.03 beta? If so, can you shorten
it to a replicable bug report?
Henry Rich
On 8/31/2021 12:46 PM, Raul Miller wrote:
If the word formation rules were changed, hmm...
Indeed, I cannot get an error
If the word formation rules were changed, hmm...
Indeed, I cannot get an error from statements like:
{{'a' 0}}.'this is a test'}}
nor
{{pdn=. ((0{pdn),&.>1{pdn) 0}}.pdn}}
(Even when I spread things out onto multiple lines and use the )n decoration,)
So that reported premature terminatio
I deny it! If I justified {{ }} on grounds of backwards-compatibility,
I was drunk. {{ }} were considered better than the alternatives, (which
were (. ). { } and a Unicode pair) even though they changed the
word-formation rules (to which they now conform).
Henry Rich
On 8/31/2021 12:23 PM
I remember wishing that the delimiters followed existing J word
formation rules, for backwards compatibility reasons. However, I also
remember being told that the new {{ and }} constructs were better for
backwards compatibility reasons.
In any event, replacing the }} in your script with } } would
> .. if you're JQt, it's easy to select all the unwanted lines and switch
on comments with ctrl-/ . You can later uncomment all those lines with
another ctrl-/ on the same selection.
Another way is to use ctrl-shift-/ which toggles on a Note. I often prefer
this as the lines remain uncommented a
So this is open for silly ideas?
Multiline comments ought to be nestable.
Personally, I like the way lua does it.
Both multiline strings and comments work like this:
[[
an example string
spanning two lines
]]
[=[
… and
another one
]=]
[[
as many = signs as you want,
so [[nesting]] is done by i
Could the marker just be an unbalanced
}} that shows up before the initiating {{?
If further marking is required maybe stipulate that it has to start in the
first column
Cheers, bob
> On Aug 31, 2021, at 08:28, Henry Rich wrote:
>
> I have wanted this often. If we could reach consensus on w
I just tried appending together about 2 meg of scripts to do this stress
test. Unfortunately, there is at least one valid instance of "}}" in the
code which prematurely terminates the block:
pdn=. ((0{pdn),&.>1{pdn) 0}}.pdn
On Tue, Aug 31, 2021 at 11:28 AM Henry Rich wrote:
> I have wanted thi
It's tempting to suggest that a bare
return.
would serve.
But maybe it should be a different, invented keyword (to distinguish
0!:0 contexts from 3 :0 contexts)?
Thanks,
--
Raul
On Tue, Aug 31, 2021 at 11:28 AM Henry Rich wrote:
>
> I have wanted this often. If we could reach consensus
I have wanted this often. If we could reach consensus on what the best
end-marker would be, I'll look into putting it in.
Henry Rich
On 8/31/2021 8:27 AM, 'Rudolf Sykora' via Programming wrote:
Dear list,
if I load a (longer) script, I sometimes want to ignore a part of it
beyond a certain
That's good advice. {{ }} does have to scan the lines & handle
recursive definitions.
{{)n
...
}}
creates a noun, similar to 0 : 0, and I don't think it scans the lines.
Henry Rich
On 8/31/2021 11:12 AM, John Baker wrote:
I'd see how abusing {{ .. }} holds up to abuse. Enclose a few hundr
Is anyone loading a J script which is a few hundred megabytes in size?
That would be ... interesting. (I am trying to imagine the purpose of
such a script.)
Thanks,
--
Raul
On Tue, Aug 31, 2021 at 11:13 AM John Baker wrote:
>
> I'd see how abusing {{ .. }} holds up to abuse. Enclose a few hun
I'd see how abusing {{ .. }} holds up to abuse. Enclose a few hundred
megabytes of text in curlies and see if J can absorb it without crashing.
To paraphrase, "spare the rod, spoil the software."
On Tue, Aug 31, 2021 at 9:09 AM Devon McCormick wrote:
> I thought of "0 : 0" too as a solution for
I thought of "0 : 0" too as a solution for this but it gets terminated by
any embedded single closing paren which you would have as part of a
function definition.
I like Raul's idea of abusing the double-curlies notation.
On Tue, Aug 31, 2021 at 11:01 AM Hauke Rehr wrote:
> I usually abuse 0:0 f
I usually abuse 0:0 for block comments.
Am 31.08.21 um 16:21 schrieb Raul Miller:
In recent J implementations (j902, j903), you can wrap the unwanted
section of the script in {{ }}.
This will defer execution of those lines and then discard them (which
is somewhat fortuitous, since execution of
In recent J implementations (j902, j903), you can wrap the unwanted
section of the script in {{ }}.
This will defer execution of those lines and then discard them (which
is somewhat fortuitous, since execution of that kind of definition
would probably be bad).
I hope this helps,
--
Raul
On Tue
I sometimes need this too. If so, and if you're JQt, it's easy to
select all the unwanted
lines and switch on comments with ctrl-/ . You can later uncomment all
those lines with
another ctrl-/ on the same selection.
However, that can be a bit awkward, and thanks to your message I've
just
Dear list,
if I load a (longer) script, I sometimes want to ignore a part of it
beyond a certain line. Is there a way to achieve that gracefully? (I
sometimes just add something syntactically wrong to stop there, on an
error).
Thank you.
Ruda
25 matches
Mail list logo
