Re: function hoisting like var
Yuh-Ruey Chen wrote:
FWIW, I definitely like this proposal more than any convoluted
var-scoping block proposal.
I just don't see the value of adding var-scoping blocks for the added
parser complexity, and problematic return/break/continue.
Unfortunately it seems people dislike the notation too much.
But the complexity that you discuss must be supported anyway.
Regardless whether it's written {let} or {{var}}, they both
say the same thing, and what they say is supported in ES4.
My code looks convoluted only because it's an odd trick.
Not only does it add {let} functionality to ES3.1, it also
strives very hard to make the tiniest possible change.
That tiny code is also conclusive proof that {{var}} is
extremely simple and has no capture problems, despite many
vague claims to the contrary.
Unfortunately it seems people dislike the notation so much,
they prefer having no {let} functionality at all in ES3.1.
--
Ingvar von Schoultz
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
It's not that your code is convoluted (although it is), it's that the added
complexity to parse {{...}}, backwards incompatibility of that syntax (code
generators), and the questionable aesthetics (it really is subjective) just
make it infeasible. All just to provide a hoisted-let-like functionality.
Plus, it's a dead horse that's been beaten one too many times :)
I would like ES3.x to support let functionality, but I'd rather ES3.x just
implement let statements/declarations, rather than {{...}}. I think others
would agree with me here.
I also prefer Igor's proposal more than {{...}}. And even if you solve
problematic continuations (return/continue/break), I would like them to be
generalized into first-class blocks.
-Yuh-Ruey Chen
On Wed, Aug 6, 2008 at 3:30 PM, Ingvar von Schoultz [EMAIL PROTECTED]wrote:
Yuh-Ruey Chen wrote:
FWIW, I definitely like this proposal more than any convoluted var-scoping
block proposal.
I just don't see the value of adding var-scoping blocks for the added
parser complexity, and problematic return/break/continue.
Unfortunately it seems people dislike the notation too much.
But the complexity that you discuss must be supported anyway.
Regardless whether it's written {let} or {{var}}, they both
say the same thing, and what they say is supported in ES4.
My code looks convoluted only because it's an odd trick.
Not only does it add {let} functionality to ES3.1, it also
strives very hard to make the tiniest possible change.
That tiny code is also conclusive proof that {{var}} is
extremely simple and has no capture problems, despite many
vague claims to the contrary.
Unfortunately it seems people dislike the notation so much,
they prefer having no {let} functionality at all in ES3.1.
--
Ingvar von Schoultz
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Okay then, even shorter. Brendan Eich wrote: But it is not what you proposed. In what way? Please be more specific, because I don't know what this supposed proposal of mine is. For example, those scope questions mention rebinding in a visible scope, midway through its code block, but that's unrelated to anything I've proposed and unrelated to how JavaScript works. I've tried to guess what you think I've proposed from the scope questions, but the guessing has made my answers long. Stupid mistake, I should have asked instead, sorry. So please say something specific that I can address. Waldemar wrote a while back: Keep in mind that function assignments hoist to the beginning of the scope in which the function is defined, so your proposal won't work. It works perfectly well. The initial |undefined| is fully intentional, it protects against capture problems. It's a minor limitation, certainly not a showstopper. -- Ingvar von Schoultz ___ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Regarding my explanations quoted below, did they clarify
things?
It would be a relief to know that things are now clear and
understandable, if they are. If there are still things
that seem unclear, strange, incompatible or impractical,
please ask about them!
After some thinking I've found that I could have explained
some of the things below better, making them clearer. Should
I post better explanations?
I'm really very curious about that rebinding problem! Both
what it means, and where I seem to propose it (or actually
propose it, inadvertently).
Ingvar
Ingvar von Schoultz wrote:
Sorry about the length of this, but I'm trying to cover the
unclear things, and often I don't know which things are unclear.
Brendan Eich wrote:
On Jul 26, 2008, at 2:07 PM, Ingvar von Schoultz wrote:
You can't get away from supporting this:
{
function a(){}
var b = a;
}
ES4 is planning to support function declarations locally
bound in blocks, so the above is valid ES4 code.
What you see above is function b() hoisting like var.
(I said b, not a.)
What you said does not make sense. It's true that var b is hoisted to
the top of the program or function body. But it is not initialized
until control flows through the assignment b = a that is part of the
var declaration. So there is no capture problem.
That's my point! There isn't any capture problem. That's exactly
what I'm showing here. And, more importantly, you can't insert a
capture problem while keeping the structure intact. The arrangement
is inherently well-behaved.
I'm trying to show that you can support functions that hoist like
var in a well-behaving way, and that this is /not/ complicated.
And the snippet is intended as proof.
But perhaps I should interpret what you said somewhat differently.
The snippet, as shown here, does not have a capture problem.
But if you improve the hoisting carelessly, there's a threat
of a capture problem lurking within it. (That's assuming that
I understand capture problem correctly -- this is a rare
case of Wikipedia not explaining a computing term.)
Somebody might want to improve the hoisting by making it so
that the function is assigned (is callable) from before we enter
the global scope.
Let's assume, for the argument, that the above code magically
starts to behave that way. Then it will work fine as long as it
stays the same. But later the programmer might decide to add a
local variable:
{
function a(){}
var b = a;
let c = 3;
}
c is now visible from inside a(), but exists only when we enter
the block. So in this situation a() must not be called before
we enter the block. The function must no longer be assigned
(callable) at the top of the global scope.
I would consider it extremely surprising semantics if I can call
a() above the block, but only before I add that local variable,
and this suddenly changes just because I add c. It's a huge change,
it's far away, and it's unrelated.
So for the sake of consistency and predictability we must assign
|undefined| whenever the code structure allows a capture problem
to be inserted as a side effect of doing something unrelated,
like adding |let c|.
That's why I said this, in the email where I first showed this
snippet:
,---
| Assigning |undefined| is correct for any function whose
| assignment depends on sequential code. The above is such a
| sequential dependency, even though it may not look that way.
`---
It may not look sequential, but that's just because I left out
a lot of details, in an attempt to keep it as brief as ever
possible, to minimize any misunderstandings.
As you can see, in my opinion, what I'm saying does make perfect
sense!
The claim that just because of this limitation my proposal
doesn't work is in my opinion quite mistaken. I consider
this functionality fully acceptable. It's simple, understandable
and predictable. Sure it's a limitation, but a minor one. It's
/far/ better and more useful and intuitive than having functions
not hoisting out at all.
The usual use case will be an if() or some such. Then the
programmer fully expects the function to be available only
afterward. Very useful. Using it in a bare block like the
above will be unusual, but if used, the rules are simple.
There is no far-too-complicated split-scope complexity. There
is no capturing of variables that haven't been declared yet.
It's simple, intuitive, well-defined and well-behaved.
Thanks, I agree. But it is not what you proposed.
Sorry, I don't understand. What is not what I proposed, and
in which one of my proposals did I not propose it?
Again, from
Waldemar's original reply, but with your proposed {{}} interpolated
and the elided code amended to say what the consequence is:
// outer scope
function c() ...;
// inner scope
{{
if (foo) {
const c = 37;
}
... c in
Re: A simple translation of the scoping-block syntax sugar -- Was: Re: function hoisting like var
Ingvar von Schoultz wrote:
Any variable
that you declare anywhere will simply splash out of any { } and
attach itself to the nearest surrounding {{ }}, becoming visible
in the entire space between them.
Int this splashing out of { }, function declarations are
a special case. Although implementations differ, I think
{{ }} could serve as an opt-in where they become specified
to follow clear rules.
In my opinion the most useful rules would be:
The name is visible in the entire space between the nearest
surrounding {{ }}. Usually it's also callable in that entire
space, but there are a few exceptions.
If you put the declaration in a sequential construct,
for example under if(), then at the {{ you can't call it
(it's |undefined|), and it becomes callable only from the
sequential construct and onward. If there is a { at the
beginning of the sequential construct, the function becomes
callable from the { and onward.
If you make more than one declaration with the same name,
this makes all the declarations of that name sequential.
Then at {{ you can't call any of the versions of the
function (again |undefined|). Each version becomes callable
at the spot of declaration.
In all other cases the function is callable in the entire
space between {{ }}.
(This is not intended as an exhaustively detailed description,
just an overview.)
(The rules will have to be changed a little if let declarations
are allowed between {{ }}, or if allowing them there in a
future version should be possible. And reserving the possibility
might be a good idea -- who knows what people will want or
will invent. To allow that, the rules become slightly more
limiting.)
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
On Jul 30, 2008, at 3:13 PM, Ingvar von Schoultz wrote: Regarding my explanations quoted below, did they clarify things? No, and I don't have time right now to deal with the great number of words you have dedicated to promoting your ideas. This is a shame, since you could have a point, but I simply can't expend the effort to try to find it given other priorities. Sorry, this is not something I'm happy about, but it's not entirely due to my being busy (i.e., it's not just me -- it's you :-/). May I suggest using fewer words when replying than you have? Generally proportionate responses would be best. I'm assuming fair play, i.e., that people are not resorting to too short and dismissive a style of replying. I'm not doing that here, I'm just letting you know I don't have time to plow through what you have written. Maybe someone else on the list has the time. /be ___ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
A simple translation of the scoping-block syntax sugar -- Was: Re: function hoisting like var
The simple translation of {{ }} that I posted yesterday needs
an improvement. It must pass |this| and |arguments| to the scoping
function.
I still think it looks quite small and simple, and that it would
be a very useful addition to the language.
try
{ var _ReturnValue = (function (arguments)
{
// Any governing for(), while() or switch() goes here.
{
// The code between {{ }} goes here.
}
throw _NormalEndSignal;
}).call (this, arguments);
return _ReturnValue;
}
catch (_Sig)
{
if (_Sig != _NormalEndSignal)
throw _Sig;
}
We must use .call(), not .apply(), so that we pass the arguments
object as such, with the outer function as arguments.callee.
In the global scope, |arguments| will be undefined, unless the
programmer has defined it. It should remain undefined inside the
scoping function.
As an alternative to the above parameter passing, I suppose the
implementation can simply remove the scope function's arguments
object, thereby making the outer arguments object visible.
Initial underscore means hidden, invisible.
For a description of the translation, see the quoted text below.
Ingvar
Ingvar von Schoultz wrote:
Ingvar von Schoultz wrote:
In theory {{ code }} could be converted to a function that
returns information about whatever break/continue/return was
reached.
I now think this could be made very simple.
The solution is to make this version slightly limited. Don't
support break/continue statements that mention any label
outside {{ }}. Leave that for a future version.
This simplifies things tremendously! And you don't need that
kind of break/continue all that often. (But a future version
must include it.)
This limited solution only needs break/continue functionality
that is already supported. Even the necessary label checking
is already supported. An unsupported break such as the following
gives a syntax error if you translate {{ }} to a one-shot
function:
Outer:
for (var OutName in OutThing)
for (var InName in InThing)
{{
break Outer;
}}
The error message that I get says undefined label, which
is misleading, but apart from that it works perfectly.
If people find the above limited support for break/continue
acceptable for this version, all that remains is the return
statement.
It seems easiest to use a solution that leaves all the return
statements in the original code intact. This makes the translation
from {{ }} to function a little elaborate and odd, but everything
is nicely contained at one spot.
Within the created function, if control reaches the final }},
throw a special error that isn't really an error, rather the
opposite, it's a signal that indicates that this is a normal
termination. When this signal is received, continue below }}.
If instead, within the created function, control reaches a
return statement that appears in the original code, the result
is a regular return, without the above thrown signal. This
indicates that whatever value was returned should be passed
along as return value of the enclosing function.
This means that {{ }} is essentially translated into this:
try
{ var ReturnValue = (function()
{ // The code that was written between {{ }} goes here.
throw NormalEndSignal;
})();
return ReturnValue;
}
catch (Sig)
{ if (Sig != NormalEndSignal)
throw Sig;
}
It seems to me that this becomes very nicely contained,
very manageable.
Implementations will probably want to optimize away this
entire elaborate arrangement and replace it with plain and
simple scoping blocks. But they can take their time. The
above provides the functionality in a simple way.
That is, unless I've missed something, and the above solution
doesn't work the way I think it does.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: A simple translation of the scoping-block syntax sugar -- Was: Re: function hoisting like var
If the simple translation of {{ }} is used, any governing
for(), while() or switch() must be moved inside the scoping
function.
A very simple, very minimalistic approach may be enough.
for (var Name in Thing)
{{
}}
Name is locally contained inside {{ }} while Thing is outside.
But maybe the compiler doesn't have to make this distinction.
Personally I'd never use the same name with different meanings
so close together, like this:
var Thing = ...;
for (var Name in Thing)
{{ var Thing = ...;
}}
I find that obscure and error-prone.
If the limitation is acceptable, a very minimalistic translation
of {{ }} could simply move the entire for (...) inside the
scoping function, and leave it at that. As long as Thing isn't
redeclared inside, the outer Thing is visible and it works.
Does this make it simple enough for ES3.1?
As an added bonus, the functionality becomes very easy to explain.
The rules become very plain and simple.
I don't know if a let statement with this limitation would be
useful. Every description of let that I can find re-uses names.
But if it's useful, it could be trivially implemented as syntax
sugar:
let (a = x, b = y)
{{
}}
would be syntax sugar for
{{ var a = x, b = y;
}}
However let has different semantics in at least one existing
implementation (it does distinguish between same-name variables
in the two scopes), so if a minimalistic let is introduced, it
should probably use a different keyword.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: A simple translation of the scoping-block syntax sugar -- Was: Re: function hoisting like var
I'd like to summarize some clarifications and arguments.
I'm thinking of what it might be like if {{ }} should become
available in ES3.1.
=== Intuitive ===
The proposed {{ }} mean exactly the same thing as { } in C, Java etc.
If switching between languages causes you any difficulties, simply
use {{ }} everywhere in your ES3.1 code. This way you get the braces
that you're used to from C and Java, apart from a slightly different
notation.
Think of it as JavaScript having a small difference in brace notation
while the functionality is the same.
=== Control flow ===
In ES3 you can use { } for plain control flow, without any scoping.
This is very useful for small projects.
If this is not a good fit for your projects, it's still very useful
elsewhere. Please don't deprecate this functionality.
For more on the usefulness of this plain control flow, see:
https://mail.mozilla.org/pipermail/es4-discuss/2008-July/003353.html
=== Difference ===
If you want the full flexibility of using both {{ }} and { },
first decide which type of brace you want to use most in your
program. Stick with that almost everywhere. Make an exception
only when you have a clear and specific reason, because you
need the functionality of the other type.
This way your code will be consistent and readable.
You can view the {{ }} as containers for variables. Any variable
that you declare anywhere will simply splash out of any { } and
attach itself to the nearest surrounding {{ }}, becoming visible
in the entire space between them.
If you prefer a more low-level view, { } are plain jumps and
branches, and only {{ }} affect the scope chain.
=== Ugly ===
The proposed {{ }} will inevitably look horrible in emails if
you use a proportional font such as Times New Roman. Real code
is never displayed in such a font. Try a programmer's editor
with a suitable font and syntax coloring. It looks completely
different!
If people still find it ugly, I hope a good alternative can be
found, because the functionality would be very useful.
=== Let declarations ===
There are existing implementations that support let.
My proposal is that /let declarations/ should be completely disabled
between {{ }}. This of course includes all { } that are nested between
{{ }}.
The only reason is simplicity, mainly for the script programmer.
Outside {{ }} they should be enabled, for compatibility and to satisfy
those who prefer the let and var combination.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: A simple translation of the scoping-block syntax sugar -- Was: Re: function hoisting like var
The translation can be made simpler.
First, define a unique value, to be returned when the end of
the scoping block is reached. If done in JavaScript:
var _EndOfBlock = {};
Then each {{ }} can be translated to this:
var _ReturnValue = (function (arguments)
{
// Any governing for(), while() or switch() goes here.
{
// The code between {{ }} goes here.
}
return _EndOfBlock;
}).call (this, arguments);
if (_ReturnValue != _EndOfBlock)
return _ReturnValue;
Ingvar
Ingvar von Schoultz wrote:
The simple translation of {{ }} that I posted yesterday needs
an improvement. It must pass |this| and |arguments| to the scoping
function.
I still think it looks quite small and simple, and that it would
be a very useful addition to the language.
try
{ var _ReturnValue = (function (arguments)
{
// Any governing for(), while() or switch() goes here.
{
// The code between {{ }} goes here.
}
throw _NormalEndSignal;
}).call (this, arguments);
return _ReturnValue;
}
catch (_Sig)
{
if (_Sig != _NormalEndSignal)
throw _Sig;
}
We must use .call(), not .apply(), so that we pass the arguments
object as such, with the outer function as arguments.callee.
In the global scope, |arguments| will be undefined, unless the
programmer has defined it. It should remain undefined inside the
scoping function.
As an alternative to the above parameter passing, I suppose the
implementation can simply remove the scope function's arguments
object, thereby making the outer arguments object visible.
Initial underscore means hidden, invisible.
For a description of the translation, see the quoted text below.
Ingvar
Ingvar von Schoultz wrote:
Ingvar von Schoultz wrote:
In theory {{ code }} could be converted to a function that
returns information about whatever break/continue/return was
reached.
I now think this could be made very simple.
The solution is to make this version slightly limited. Don't
support break/continue statements that mention any label
outside {{ }}. Leave that for a future version.
This simplifies things tremendously! And you don't need that
kind of break/continue all that often. (But a future version
must include it.)
This limited solution only needs break/continue functionality
that is already supported. Even the necessary label checking
is already supported. An unsupported break such as the following
gives a syntax error if you translate {{ }} to a one-shot
function:
Outer:
for (var OutName in OutThing)
for (var InName in InThing)
{{
break Outer;
}}
The error message that I get says undefined label, which
is misleading, but apart from that it works perfectly.
If people find the above limited support for break/continue
acceptable for this version, all that remains is the return
statement.
It seems easiest to use a solution that leaves all the return
statements in the original code intact. This makes the translation
from {{ }} to function a little elaborate and odd, but everything
is nicely contained at one spot.
Within the created function, if control reaches the final }},
throw a special error that isn't really an error, rather the
opposite, it's a signal that indicates that this is a normal
termination. When this signal is received, continue below }}.
If instead, within the created function, control reaches a
return statement that appears in the original code, the result
is a regular return, without the above thrown signal. This
indicates that whatever value was returned should be passed
along as return value of the enclosing function.
This means that {{ }} is essentially translated into this:
try
{ var ReturnValue = (function()
{ // The code that was written between {{ }} goes here.
throw NormalEndSignal;
})();
return ReturnValue;
}
catch (Sig)
{ if (Sig != NormalEndSignal)
throw Sig;
}
It seems to me that this becomes very nicely contained,
very manageable.
Implementations will probably want to optimize away this
entire elaborate arrangement and replace it with plain and
simple scoping blocks. But they can take their time. The
above provides the functionality in a simple way.
That is, unless I've missed something, and the above solution
doesn't work the way I think it does.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
Re: function hoisting like var
Ingvar von Schoultz wrote:
In theory {{ code }} could be converted to a function that
returns information about whatever break/continue/return was
reached. But I'd be quite surprised if that is easy.
Something like this. Written with syntax sugar:
Outer:
for (var OutName in OutThing)
for (var InName in InThing)
{{
break Outer;
}}
Translation:
Outer:
for (var OutName in OutThing)
{
var _Result = (function (_InThing)
{
for (var InName in _InThing)
{
return ({JumpSpot: 'break Outer'})
}
})(InThing);
if (_Result.JumpSpot == 'break Outer')
break Outer;
}
The inner for() is part of the scoping block, so it belongs
inside, even though the original code has it above.
We must make sure the value of InThing is available inside
even if the name is declared for a different variable inside.
I use initial underscore to indicate something internal and
invisible.
It doesn't look complicated here! Unfortunately these things
have a terrible tendency to grow in complexity...
I now think this could be made very simple.
The solution is to make this version slightly limited. Don't
support break/continue statements that mention any label
outside {{ }}. Leave that for a future version.
This simplifies things tremendously! And you don't need that
kind of break/continue all that often. (But a future version
must include it.)
This limited solution only needs break/continue functionality
that is already supported. Even the necessary label checking
is already supported. An unsupported break such as the following
gives a syntax error if you translate {{ }} to a one-shot
function:
Outer:
for (var OutName in OutThing)
for (var InName in InThing)
{{
break Outer;
}}
The error message that I get says undefined label, which
is misleading, but apart from that it works perfectly.
If people find the above limited support for break/continue
acceptable for this version, all that remains is the return
statement.
It seems easiest to use a solution that leaves all the return
statements in the original code intact. This makes the translation
from {{ }} to function a little elaborate and odd, but everything
is nicely contained at one spot.
Within the created function, if control reaches the final }},
throw a special error that isn't really an error, rather the
opposite, it's a signal that indicates that this is a normal
termination. When this signal is received, continue below }}.
If instead, within the created function, control reaches a
return statement that appears in the original code, the result
is a regular return, without the above thrown signal. This
indicates that whatever value was returned should be passed
along as return value of the enclosing function.
This means that {{ }} is essentially translated into this:
try
{ var ReturnValue = (function()
{ // The code that was written between {{ }} goes here.
throw NormalEndSignal;
})();
return ReturnValue;
}
catch (Sig)
{ if (Sig != NormalEndSignal)
throw Sig;
}
It seems to me that this becomes very nicely contained,
very manageable.
Implementations will probably want to optimize away this
entire elaborate arrangement and replace it with plain and
simple scoping blocks. But they can take their time. The
above provides the functionality in a simple way.
That is, unless I've missed something, and the above solution
doesn't work the way I think it does.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Brendan Eich wrote:
On Jul 26, 2008, at 2:06 PM, Ingvar von Schoultz wrote:
How sad! It seemed such a simple and intuitive notation!
Opinions vary, but all the ones I heard at the Ecma TC39 meeting
found it neither simple nor intuitive, and some abhorred it on
aesthetic grounds to boot.
Not simple? How is that possible?
There have been several misunderstandings. Did they spread
that far?
{{ }} is just the same as a scoping function used on ES3:
(function() { code })()
Everything works just the way I meant if you take ES3, or any
platform that does /not/ support let declarations, and make
it so that {{ }} becomes syntax sugar for the above one-shot
scoping function.
Of course, the way I think about {{ }}, I don't see it as a
function, I see it as a name-binding scope. But the scoping
effect is just that.
For me, having a single scoping block is simpler than having
two, and having a single visibility keyword is simpler than
having two. And I think my notation is /much/ simpler than
the one-shot function!
{{ code }} (function() { code })()
I do find my notation slightly ugly. But the one-shot function
is worse in my view. Charming but very kludgey.
(Thanks Igor for noticing that {{ }} are synonymous with one-shot
scoping functions. Why didn't I think of that!)
I think all of these would be unambiguous:
{. code .}
{: code :}
{| code |}
{[ code ]}
[[ code ]]
[ code ]
These are either syntax errors without opt-in versioning,
Yes, and this should guarantee that they are unique.
I now have the impression that, on the client side, opt-in
by version would be necessary in any case, with any notation.
If old platforms accept the syntax, it changes existing
semantics; if they don't, it requires opt-in to avoid errors.
or (the
last three) do create incompatible ambiguity (consider array
initialisers).
It seems to me that they could be disambiguated easily. I
comment on this in a reply to Igor.
What's more, as Waldemar pointed out many threads (and too many
words) ago, they create capture problems.
There were misunderstandings.
If there are fundamental differences in this regard between {{ }}
and one-shot scoping functions, I can't find them.
Please work through the
last mail I sent before replying; if some vocabulary or infelicitous
word choice is causing any confusion, feel free to mail me privately
and ask pointed questions. Thanks,
Thanks for the invitation. I'll certainly take you up on it
if the need or urge arises.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
2008/7/28 Ingvar von Schoultz [EMAIL PROTECTED]:
I wonder if people would like using them.
Anything that is available in a browser will be used and abused ;).
{[ code ]}
[[ code ]]
[ code ]
Any of this can be a valid ES3 code as [] always means an array literal.
I get the impression that disambiguation would be easy.
eval('[]') is a valid and, in fact, useful ES3 code. Similarly eval('{[]}').
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
2008/7/28 Ingvar von Schoultz [EMAIL PROTECTED]:
{{ }} is just the same as a scoping function used on ES3:
(function() { code })()
As Lars Hansen has pointed out any proposal for a shorthand for
(function() { code })() has to deal with break/continue/return inside
the code. This is the the reason if anything I would prefer in ES3.1
just shorthands for function definitions like
function() expr - equivalent to function () { return expr; }. This is
already in ES4 and is implemented by at least on implementation
(SpiderMonkey).
or
function optional_name { code } - equivalent to function
optional_name() { code } - a hypothetical shortcut that would allow to
write a pseudo-blocks like
function {
}();
with clear emphasis that this is a lambda with usual rules for
break/continue/return.
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Mike Shaver wrote:
On Sun, Jul 27, 2008 at 7:51 PM, Ingvar von Schoultz
[EMAIL PROTECTED] wrote:
Igor Bukanov wrote:
eval('[]') is a valid and, in fact, useful ES3 code. Similarly eval('{[]}').
Yes, and by my rules they both create and return a new, empty
array, which is intuitively expected and compatible.
And the latter only does because the { is parsed as starting a block;
if it were to start out assuming a literal, you would get an illegal
object initializer.
I meant that [ would follow those rules, not {. The rules I
sketched are definitely not usable for {.
I get the impression that the parser always must start out
assuming it's a literal, and backtrack if the syntax doesn't
fit. The only exception would be at a spot where a literal
isn't possible while a block is possible, but I don't think
any such spot is possible.
if { block; }
would seem to be such a spot, no?
These are valid literals (immediately discarded, but valid
syntax):
if (true) [10, 20];
if (true) { [10, 20] };
Likewise valid:
if (true) 10;
Regardless, I'm pretty sure ES3
compat requires that a statement starting with a { be taken as a block
and not a literal, with no backtracking.
js {a:5, b:6}
typein:4: SyntaxError: invalid label:
typein:4: {a:5, b:6}
typein:4: ..^
Mike
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Ingvar von Schoultz wrote:
Igor Bukanov wrote:
2008/7/28 Ingvar von Schoultz wrote:
snip
Any of this can be a valid ES3 code as [] always means an
array literal.
I get the impression that disambiguation would be easy.
eval('[]') is a valid and, in fact, useful ES3 code. Similarly
eval('{[]}').
Yes, and by my rules they both create and return a new, empty
array, which is intuitively expected and compatible.
Are you saying that a block then may never be empty as a syntax
rule? ES 3 blocks are allowed to be empty and I bet there are
examples where blocks contain nothing but an IE/JScript
conditional comment and so are going to be interpreted as
being empty by other ES 3 implementations. (Comments are
allowed inside array literal definitions at present).
When I said that an array is a comma-separated list of values,
the fact that I left out the empty and the single-value cases
didn't mean that they should be seen as blocks! I just didn't
want to write a lengthy, exhaustive full grammar...
snip
{
1,2,3,4,5
}
- is a valid ES 3 Program (even if a pointless one); A Block
statement containing an Expression statement (with automatic
semi-colon insertion making the Expression statement
into - 1,2,3,4,5; -).
eval('{1,2,3,4,5}') - results in the value 5, while -
eval('[1,2,3,4,5]') - returns a 5 element array.
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Igor Bukanov wrote: 2008/7/28 Ingvar von Schoultz [EMAIL PROTECTED]: If [...] is preceded by =, or enclosed in ( ), or in some other position where you can't have a block, it's a literal. If it's a literal it must contain a comma-separated list of values, so if the syntax doesn't match this, it's a block. Such rules require an arbitrary look ahead in the parser so it can distinguish that in cases like if (x) [arbitrary_expression;] [] means a block while in if (x) [arbitrarily_expression]; [] would mean a literal. This would require a mayor change in most if not all current ES parser implementations. There were some proposals for ES4 syntax that would require such look ahead, but they were rejected not only technical grounds but also on the grounds that such look ahead poses comprehension problem for a human brain! And here are we talking about a minimalistic sugar for ES3.1. Regards, Igor ___ Es3.x-discuss mailing list [EMAIL PROTECTED] https://mail.mozilla.org/listinfo/es3.x-discuss I wasn't aware that backtracking was a difficulty. That makes quite a difference! In theory this could be overcome. Often the ambiguity doesn't matter: if (x) [a(), b(), c()]; Block or throwaway array, the end result is the same. So in principle you could start out building an array, and on the first nonmatching syntax you discard the half-built array and build a scope block instead: if (x) [a(), b(), c(), d(), e(), let f = g;] Of course in practice such a solution is too messy, unless you have some extremely thorny problem that just can't be solved in any other way. And this certainly isn't that kind of problem. So [ ] are out. Unless they're reincarnated as: [: code :] Ah, now I know! I have the perfect solution, guaranteed to be unambiguous. Happy code! (-: Scoping blocks are delimited by smileys! :-) -- Ingvar von Schoultz --- (My quirky use of capitals in code comes from my opinion that reserved and predefined words should all start with lowercase, and user-defined should all start with uppercase, because this will easily and elegantly prevent a host of name-collision problems when things like programming languages are upgraded with new labels.) ___ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Richard Cornford wrote:
Ingvar von Schoultz wrote:
Igor Bukanov wrote:
2008/7/28 Ingvar von Schoultz wrote:
snip
Any of this can be a valid ES3 code as [] always means an
array literal.
I get the impression that disambiguation would be easy.
eval('[]') is a valid and, in fact, useful ES3 code. Similarly
eval('{[]}').
Yes, and by my rules they both create and return a new, empty
array, which is intuitively expected and compatible.
Are you saying that a block then may never be empty as a syntax
rule? ES 3 blocks are allowed to be empty and I bet there are
examples where blocks contain nothing but an IE/JScript
conditional comment and so are going to be interpreted as
being empty by other ES 3 implementations. (Comments are
allowed inside array literal definitions at present).
Does it matter whether it's interpreted as an empty block or
an empty discarded array?
When I said that an array is a comma-separated list of values,
the fact that I left out the empty and the single-value cases
didn't mean that they should be seen as blocks! I just didn't
want to write a lengthy, exhaustive full grammar...
snip
{
1,2,3,4,5
}
- is a valid ES 3 Program (even if a pointless one); A Block
statement containing an Expression statement (with automatic
semi-colon insertion making the Expression statement
into - 1,2,3,4,5; -).
eval('{1,2,3,4,5}') - results in the value 5, while -
eval('[1,2,3,4,5]') - returns a 5 element array.
Yes. And if the rules for { remain unchanged, and [ starts out
assuming it's a literal, you get the same result with my rules.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
On Sun, Jul 27, 2008 at 9:15 PM, Ingvar von Schoultz [EMAIL PROTECTED] wrote: Does it matter whether it's interpreted as an empty block or an empty discarded array? It's not always discarded: js var x = 1 js eval (if (x) [1, 2, 3]) 1,2,3 Mike ___ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Mike Shaver wrote: On Sun, Jul 27, 2008 at 9:15 PM, Ingvar von Schoultz [EMAIL PROTECTED] wrote: Does it matter whether it's interpreted as an empty block or an empty discarded array? It's not always discarded: js var x = 1 js eval (if (x) [1, 2, 3]) 1,2,3 That becomes an array, no problem. My Does it matter was about empty blocks only, in reply to your question: Are you saying that a block then may never be empty as a syntax rule? Does it matter whether it's interpreted as an empty block or an empty discarded array? Array is the default. An empty block gets misinterpreted as an empty array. So a code generator that produces an eval that returns a scoping block must deal with the special case that the block may be empty or contain a comma-separated list of values. It's the generator's own fault for producing such silly code, so I don't feel sorry for that generator. :-) -- Ingvar von Schoultz --- (My quirky use of capitals in code comes from my opinion that reserved and predefined words should all start with lowercase, and user-defined should all start with uppercase, because this will easily and elegantly prevent a host of name-collision problems when things like programming languages are upgraded with new labels.) ___ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Ingvar von Schoultz wrote:
Richard Cornford wrote:
snip
{
1,2,3,4,5
}
- is a valid ES 3 Program (even if a pointless one); A Block
statement containing an Expression statement (with automatic
semi-colon insertion making the Expression statement
into - 1,2,3,4,5; -).
eval('{1,2,3,4,5}') - results in the value 5, while -
eval('[1,2,3,4,5]') - returns a 5 element array.
Yes. And if the rules for { remain unchanged, and [ starts
out assuming it's a literal, you get the same result with my rules.
And if I wanted that block scope this -
eval({[\n eval('var a = 1'),eval('var b = 2'),eval('a + b')\n]});
- will still give me an array (and side effects in the containing scope)?
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Richard Cornford wrote:
Ingvar von Schoultz wrote:
Richard Cornford wrote:
snip
{
1,2,3,4,5
}
- is a valid ES 3 Program (even if a pointless one); A Block
statement containing an Expression statement (with automatic
semi-colon insertion making the Expression statement
into - 1,2,3,4,5; -).
eval('{1,2,3,4,5}') - results in the value 5, while -
eval('[1,2,3,4,5]') - returns a 5 element array.
Yes. And if the rules for { remain unchanged, and [ starts
out assuming it's a literal, you get the same result with my rules.
And if I wanted that block scope this -
eval({[\n eval('var a = 1'),eval('var b = 2'),eval('a + b')\n]});
- will still give me an array (and side effects in the containing scope)?
Goodness, if the code generator writes that, it deserves to melt!
I'd say if you're aware of scoping blocks, and want to have a
scoping block, and despite this insist on separating with commas
rather than semicolons or bare newlines, well, then you deserve
the array that you get!
So the solution is to use semicolon or newline as separator,
preferably somewhere near the beginning.
Of course the real solution to all this is to use my original
{{ }} instead. You get used to them. And they're very easy to
type!
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Igor Bukanov wrote:
2008/7/28 Ingvar von Schoultz [EMAIL PROTECTED]:
{{ }} is just the same as a scoping function used on ES3:
(function() { code })()
As Lars Hansen has pointed out any proposal for a shorthand for
(function() { code })() has to deal with break/continue/return inside
the code.
Indeed I wasn't thinking of that in my description.
This is the the reason if anything I would prefer in ES3.1
just shorthands for function definitions like
function() expr - equivalent to function () { return expr; }. This is
already in ES4 and is implemented by at least on implementation
(SpiderMonkey).
I rarely feel a need for such tiny scoping blocks, but often
long for big scoping blocks with intuitive syntax.
function optional_name { code } - equivalent to function
optional_name() { code } - a hypothetical shortcut that would allow to
write a pseudo-blocks like
function {
}();
with clear emphasis that this is a lambda with usual rules for
break/continue/return.
Yes, we can't have syntax that looks like a block but forbids
or misunderstands break/continue/return.
In theory {{ code }} could be converted to a function that
returns information about whatever break/continue/return was
reached. But I'd be quite surprised if that is easy.
Something like this. Written with syntax sugar:
Outer:
for (var OutName in OutThing)
for (var InName in InThing)
{{
break Outer;
}}
Translation:
Outer:
for (var OutName in OutThing)
{
var _Result = (function (_InThing)
{
for (var InName in _InThing)
{
return ({JumpSpot: 'break Outer'})
}
})(InThing);
if (_Result.JumpSpot == 'break Outer')
break Outer;
}
The inner for() is part of the scoping block, so it belongs
inside, even though the original code has it above.
We must make sure the value of InThing is available inside
even if the name is declared for a different variable inside.
I use initial underscore to indicate something internal and
invisible.
It doesn't look complicated here! Unfortunately these things
have a terrible tendency to grow in complexity...
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
On Jul 26, 2008, at 4:03 AM, Ingvar von Schoultz wrote:
[EMAIL PROTECTED] wrote:
Ingvar von Schoultz wrote:
[EMAIL PROTECTED] wrote:
I'm trying to keep the language relatively simple.
You can't get away from supporting this:
{
function a(){}
var b = a;
}
What do you mean? This is a syntax error in both ES3 and ES3.1.
It works fine in Firefox 2, Konqueror 3, Opera 9, Internet
Explorer 6, and server-side Rhino with JavaScript 1.6.
Waldemar meant precisely what he wrote: ES3 and draft ES3.1 -- the
specifications, not random JS implementations.
Five platforms out of five. Can you throw a syntax error here
and claim to be compatible?
The implementations are not compatible. Please see the earlier es4-
discuss thread with subject Function declarations in statements at:
https://mail.mozilla.org/pipermail/es4-discuss/2007-March/
thread.html#527
It does not already exist in ES3 or ES3.1.
It exists on platforms as described above. I assumed that ES4
would be compatible.
No, because it is impossible to be compatible with conflicting
extensions to ES3 that browsers have implemented. The conflicts and
undesirable intersection semantics are why ES4 proposes, and ES3.1
considered but deferred, block-scoped functions that must be direct
children of braced blocks. This requires opt-in versioning, which is
why ES3.1 deferred it.
/be
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Igor Bukanov wrote:
2008/7/26 Ingvar von Schoultz [EMAIL PROTECTED]:
Do you really think {{ }} appears in existing code, correctly
enclosing statement blocks, with the {{ and the }} placed tightly
together both at the beginning and at the end?
Yes: I have seen the code like
if (x) {
{
code
}
}
Now pass through a JS compressor that removes redundant whitespace and
you get precizelly {{ }}.
In addition, those {{ can happen in a code generator.
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
How sad! It seemed such a simple and intuitive notation!
I think all of these would be unambiguous:
{. code .}
{: code :}
{| code |}
{[ code ]}
[[ code ]]
[ code ]
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Brendan Eich wrote:
Waldemar meant precisely what he wrote: ES3 and draft ES3.1 -- the
specifications, not random JS implementations.
Oops, I got lost in details and strayed far away from the
point that I wanted to make. In fact I should have said
this from the beginning:
You can't get away from supporting this:
{
function a(){}
var b = a;
}
ES4 is planning to support function declarations locally
bound in blocks, so the above is valid ES4 code.
What you see above is function b() hoisting like var.
(I said b, not a.)
There is no far-too-complicated split-scope complexity. There
is no capturing of variables that haven't been declared yet.
It's simple, intuitive, well-defined and well-behaved.
The above is the /exact/ functionality of function hoisting
like var, apart from using two names. You can refuse the
clearer syntax, but you can't refuse the above code and
functionality.
In other words, complexity is not a problem. ES4 can easily
choose whatever semantics people prefer.
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
2008/7/26 Ingvar von Schoultz [EMAIL PROTECTED]:
I think all of these would be unambiguous:
{. code .}
{. not work since {.0; } is a valid ES3
{: code :}
{| code |}
These 2 cases are indeed invalid ES3.
{[ code ]}
[[ code ]]
[ code ]
Any of this can be a valid ES3 code as [] always means an array literal.
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
On Jul 26, 2008, at 2:07 PM, Ingvar von Schoultz wrote:
You can't get away from supporting this:
{
function a(){}
var b = a;
}
ES4 is planning to support function declarations locally
bound in blocks, so the above is valid ES4 code.
What you see above is function b() hoisting like var.
(I said b, not a.)
What you said does not make sense. It's true that var b is hoisted to
the top of the program or function body. But it is not initialized
until control flows through the assignment b = a that is part of the
var declaration. So there is no capture problem.
There is no far-too-complicated split-scope complexity. There
is no capturing of variables that haven't been declared yet.
It's simple, intuitive, well-defined and well-behaved.
Thanks, I agree. But it is not what you proposed. Again, from
Waldemar's original reply, but with your proposed {{}} interpolated
and the elided code amended to say what the consequence is:
// outer scope
function c() ...;
// inner scope
{{
if (foo) {
const c = 37;
}
... c in your proposal must be hoisted to the {{,
so it can't be function c -- yet it can't be
initialized to 37 if foo is falsy ...
}}
You could reply that const is new (sort of -- two browsers already
implement it one way, another treats it as var) and therefore should
always scope to { or {{, whichever is closer. But the point stands if
you replace const with function or var and hoist to the {{. Repeating
the next counter-example, with {{}} changes again, to track your
proposal since the original exchange with Waldemar:
// outer scope
function c() ...;
// inner scope
{{
function f() {
return c;
}
a = f();
if (foo) {
const c = 37;
}
b = f();
... just what do a and b hold here? Was f's captured
variable rebound by the if statement? ...
}}
And so on.
The above is the /exact/ functionality of function hoisting
like var, apart from using two names. You can refuse the
clearer syntax, but you can't refuse the above code and
functionality.
I think I see the confusion now. Do you believe that in the var b =
a; code you wrote, both the binding of the var named b *and* its
initialization with the value of the function object denoted a are
hoisted? Hoisted up to what point?
Waldemar wrote a while back: Keep in mind that function assignments
hoist to the beginning of the scope in which the function is defined,
so your proposal won't work.
The word assignment where definition was perhaps more precise
(function definitions replace extant properties of the same name in
the variable object, they are not equivalent to assignment
expressions) may have misled you. From the context and the long-
standing spec and implementation behavior with functions not in
blocks or any other sub-statement position, it was clear (I think)
what was meant, but I can see how this could be confusing.
Assignment expressions and initializers in var statements do not
hoist or otherwise move in the flow of control.
/be
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
On Jul 26, 2008, at 2:06 PM, Ingvar von Schoultz wrote:
How sad! It seemed such a simple and intuitive notation!
Opinions vary, but all the ones I heard at the Ecma TC39 meeting
found it neither simple nor intuitive, and some abhorred it on
aesthetic grounds to boot.
I think all of these would be unambiguous:
{. code .}
{: code :}
{| code |}
{[ code ]}
[[ code ]]
[ code ]
These are either syntax errors without opt-in versioning, or (the
last three) do create incompatible ambiguity (consider array
initialisers).
What's more, as Waldemar pointed out many threads (and too many
words) ago, they create capture problems. Please work through the
last mail I sent before replying; if some vocabulary or infelicitous
word choice is causing any confusion, feel free to mail me privately
and ask pointed questions. Thanks,
/be
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Sorry about the length of this, but I'm trying to cover the
unclear things, and often I don't know which things are unclear.
Brendan Eich wrote:
On Jul 26, 2008, at 2:07 PM, Ingvar von Schoultz wrote:
You can't get away from supporting this:
{
function a(){}
var b = a;
}
ES4 is planning to support function declarations locally
bound in blocks, so the above is valid ES4 code.
What you see above is function b() hoisting like var.
(I said b, not a.)
What you said does not make sense. It's true that var b is hoisted to
the top of the program or function body. But it is not initialized
until control flows through the assignment b = a that is part of the
var declaration. So there is no capture problem.
That's my point! There isn't any capture problem. That's exactly
what I'm showing here. And, more importantly, you can't insert a
capture problem while keeping the structure intact. The arrangement
is inherently well-behaved.
I'm trying to show that you can support functions that hoist like
var in a well-behaving way, and that this is /not/ complicated.
And the snippet is intended as proof.
But perhaps I should interpret what you said somewhat differently.
The snippet, as shown here, does not have a capture problem.
But if you improve the hoisting carelessly, there's a threat
of a capture problem lurking within it. (That's assuming that
I understand capture problem correctly -- this is a rare
case of Wikipedia not explaining a computing term.)
Somebody might want to improve the hoisting by making it so
that the function is assigned (is callable) from before we enter
the global scope.
Let's assume, for the argument, that the above code magically
starts to behave that way. Then it will work fine as long as it
stays the same. But later the programmer might decide to add a
local variable:
{
function a(){}
var b = a;
let c = 3;
}
c is now visible from inside a(), but exists only when we enter
the block. So in this situation a() must not be called before
we enter the block. The function must no longer be assigned
(callable) at the top of the global scope.
I would consider it extremely surprising semantics if I can call
a() above the block, but only before I add that local variable,
and this suddenly changes just because I add c. It's a huge change,
it's far away, and it's unrelated.
So for the sake of consistency and predictability we must assign
|undefined| whenever the code structure allows a capture problem
to be inserted as a side effect of doing something unrelated,
like adding |let c|.
That's why I said this, in the email where I first showed this
snippet:
,---
| Assigning |undefined| is correct for any function whose
| assignment depends on sequential code. The above is such a
| sequential dependency, even though it may not look that way.
`---
It may not look sequential, but that's just because I left out
a lot of details, in an attempt to keep it as brief as ever
possible, to minimize any misunderstandings.
As you can see, in my opinion, what I'm saying does make perfect
sense!
The claim that just because of this limitation my proposal
doesn't work is in my opinion quite mistaken. I consider
this functionality fully acceptable. It's simple, understandable
and predictable. Sure it's a limitation, but a minor one. It's
/far/ better and more useful and intuitive than having functions
not hoisting out at all.
The usual use case will be an if() or some such. Then the
programmer fully expects the function to be available only
afterward. Very useful. Using it in a bare block like the
above will be unusual, but if used, the rules are simple.
There is no far-too-complicated split-scope complexity. There
is no capturing of variables that haven't been declared yet.
It's simple, intuitive, well-defined and well-behaved.
Thanks, I agree. But it is not what you proposed.
Sorry, I don't understand. What is not what I proposed, and
in which one of my proposals did I not propose it?
Again, from
Waldemar's original reply, but with your proposed {{}} interpolated
and the elided code amended to say what the consequence is:
// outer scope
function c() ...;
// inner scope
{{
if (foo) {
const c = 37;
}
... c in your proposal must be hoisted to the {{,
so it can't be function c -- yet it can't be
initialized to 37 if foo is falsy ...
}}
Yes, c exists in the inner scope {{ }}. It exists there from
before you enter the scope and throughout. It shadows the outer
c() throughout. If foo is false, the constant c is never initialized.
You could reply that const is new (sort of -- two browsers already
implement it one way, another treats it as var) and therefore should
always scope to { or {{, whichever is closer. But the point stands if
you replace const with function or var and hoist to the {{.
Which point? Forgive me, I didn't see any point. I
Re: function hoisting like var
Ingvar von Schoultz wrote:
[EMAIL PROTECTED] wrote:
I'm trying to keep the language relatively simple.
You can't get away from supporting this:
{
function a(){}
var b = a;
}
What do you mean? This is a syntax error in both ES3 and ES3.1.
On the contrary, the functionality already exists, as shown
above.
It does not already exist in ES3 or ES3.1.
Keep in mind that function assignments hoist to the beginning of the
scope
in which the function is defined, so your proposal won't work.
When the programmer explicitly says that the assignment depends
on sequential code, then do what the programmer says. Anything
else is an error. Do it by assigning |undefined| before scope
entry. This is the only correct thing to do.
I don't understand.
You're
trying to do a complex split-scope approach where each function
definition
has *two* scopes, one in which it is declared and one in which it is
defined, but even that won't work with const, typed functions and
variables, etc.
Are you saying that the function /body/ gets into trouble,
or the function /name/?
Both.
The function /body/ stays where it is. The hoisting doesn't
affect it in any way. Moving the body would change its context
and meaning catastrophically. Don't touch it.
That doesn't answer the question. The problem is that the body can
capture variables that haven't been declared yet.
The /name/ becomes a var. Treat it like any other var. Hoist
it and assign |undefined|, exactly like you do with other vars.
That's incompatible with how functions are defined in ES3.
Are you saying that because this var is related to a function
it can't be treated like other vars? Is this var fundamentally
different from other vars? At least above it isn't.
Huh? Do you understand how ES3 works?
That email is about some wildly unworkable dynamic scoping.
It has nothing to do with anything I ever said. You jumped
to that conclusion.
Please stop insisting that I'm proposing that nonsense. I'm
not. I never did.
I haven't seen a sensible and compatible proposal yet.
You'd then have to
introduce extra rules about some definitions only being possible within
{{}} blocks, which would then affect the behavior of existing definitions
like var if one of the other definitions within the same block was a
const
or function, which would snowball into a complex mess.
Unrelated, I believe.
So you're saying that your {{}} proposal has nothing to do with this?
Waldemar
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
I guess that proposal can be summarized in a very short form:
make
{{ code }}
a syntax sugar for
(function() { code })()
On a few occasions I have used the latter form in ES3 programs to get
the benefits of the let locals without using the let keyword.
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
RE: function hoisting like var
Is
{{ return }}
syntactic sugar for
(function() {{ return }})()
too?
--lars
-Original Message-
From: [EMAIL PROTECTED] [mailto:es4-discuss-
[EMAIL PROTECTED] On Behalf Of Igor Bukanov
Sent: 25. juli 2008 13:34
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]; es4-discuss@mozilla.org; Ingvar von
Schoultz
Subject: Re: function hoisting like var
I guess that proposal can be summarized in a very short form:
make
{{ code }}
a syntax sugar for
(function() { code })()
On a few occasions I have used the latter form in ES3 programs to get
the benefits of the let locals without using the let keyword.
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
2008/7/25 Lars Hansen [EMAIL PROTECTED]:
Is
{{ return }}
syntactic sugar for
(function() {{ return }})()
too?
I have forgot about the return or break/continue. But this is not a
point since a syntax sugar for (function() { code })() can require
that code must not contain return or break/continue outside the code
effectively making it a let expression on steroids. The point is that
is it worth to provide a shortcut for (function() { code })() since
that is used on web. Of cause, if desired, such shortcut should be
pure syntax extension over ES3, not {{ }} as that silently changes the
meaning of the existing code.
One way to get that extension comes from an observation that ES4
allows to drop the braces around function body and the return keyword
if the body is the single return exp. If ES3.1 would support it, then
to simulate an effect of let expression like
let (a = arg1, b = arg2) expression
one could write
(function(a, b) expression)(arg1, arg2)
or
(function(a, b) a = arg1, b = arg2, expression)()
If, in addition to this shortcut, ES4 would allow to drop an empty
argument list from a function with a requirement that function { }
always means an expression and ES3.1 would also support that, then in
place of a block with let variables one can write:
function {
}();
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Igor Bukanov wrote:
I guess that proposal can be summarized in a very short form:
make
{{ code }}
a syntax sugar for
(function() { code })()
Indeed you're right.
When I proposed {{ }} my intent was simplicity, but I didn't
realize that implementing it could be made that simple, the
change that small.
By the way, your translation
(function() { code })()
can also be expressed as
new function() { code }
Less parentheses!
(But I haven't explored possible drawbacks.)
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
Igor Bukanov wrote:
Of cause, if desired, such shortcut should be
pure syntax extension over ES3, not {{ }} as that silently changes the
meaning of the existing code.
Do you really think {{ }} appears in existing code, correctly
enclosing statement blocks, with the {{ and the }} placed tightly
together both at the beginning and at the end?
I find this somewhat unlikely, since it's a redundant repetition
that means exactly the same as { }.
Ingvar
Igor Bukanov wrote:
2008/7/25 Lars Hansen [EMAIL PROTECTED]:
Is
{{ return }}
syntactic sugar for
(function() {{ return }})()
too?
I have forgot about the return or break/continue. But this is not a
point since a syntax sugar for (function() { code })() can require
that code must not contain return or break/continue outside the code
effectively making it a let expression on steroids. The point is that
is it worth to provide a shortcut for (function() { code })() since
that is used on web. Of cause, if desired, such shortcut should be
pure syntax extension over ES3, not {{ }} as that silently changes the
meaning of the existing code.
One way to get that extension comes from an observation that ES4
allows to drop the braces around function body and the return keyword
if the body is the single return exp. If ES3.1 would support it, then
to simulate an effect of let expression like
let (a = arg1, b = arg2) expression
one could write
(function(a, b) expression)(arg1, arg2)
or
(function(a, b) a = arg1, b = arg2, expression)()
If, in addition to this shortcut, ES4 would allow to drop an empty
argument list from a function with a requirement that function { }
always means an expression and ES3.1 would also support that, then in
place of a block with let variables one can write:
function {
}();
Regards, Igor
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
--
Ingvar von Schoultz
--- (My quirky use of capitals in code comes from my opinion that
reserved and predefined words should all start with lowercase, and
user-defined should all start with uppercase, because this will easily
and elegantly prevent a host of name-collision problems when things
like programming languages are upgraded with new labels.)
___
Es4-discuss mailing list
Es4-discuss@mozilla.org
https://mail.mozilla.org/listinfo/es4-discuss
Re: function hoisting like var
[EMAIL PROTECTED] wrote:
Ingvar von Schoultz wrote:
[EMAIL PROTECTED] wrote:
I'm trying to keep the language relatively simple.
You can't get away from supporting this:
{
function a(){}
var b = a;
}
What do you mean? This is a syntax error in both ES3 and ES3.1.
It works fine in Firefox 2, Konqueror 3, Opera 9, Internet
Explorer 6, and server-side Rhino with JavaScript 1.6.
Five platforms out of five. Can you throw a syntax error here
and claim to be compatible?
You can test it in browsers at http://test-a-1.appjet.net/
Below is a more detailed example that also includes let. It
uses JavaScript 1.7 and works in Firefox 2. You can test it
at http://test-a-2.appjet.net/
alert (b);
{
function a(){}
var b = a;
let c = 3;
alert (c);
}
alert (b);
alert (typeof c);
On the contrary, the functionality already exists, as shown
above.
It does not already exist in ES3 or ES3.1.
It exists on platforms as described above. I assumed that ES4
would be compatible.
Keep in mind that function assignments hoist to the beginning of the
scope
in which the function is defined, so your proposal won't work.
When the programmer explicitly says that the assignment depends
on sequential code, then do what the programmer says. Anything
else is an error. Do it by assigning |undefined| before scope
entry. This is the only correct thing to do.
I don't understand.
In all normal cases the function is assigned at the beginning
of the scope. However, there are some special cases where this
would be an incorrect interpretation of the source code.
If the programmer clearly says that the assignment is /unknown/
at the beginning of the scope, then assigning prematurely would
be wrong.
For example, the programmer might say:
if (Math.random() 0.5)
function f() {return 1}
else
function f() {return 2}
Implementations vary here, but let's assume that it is defined
to mean what it says.
Then f will be assigned either the first or the second body,
depending on the random number. This decision comes /after/
you enter the scope. The assignment depends on something that
happens /inside/ the scope. Therefore, here the programmer
is saying quite clearly that the assignment is /unknown/ at
the beginning of the scope.
It would be wrong for the compiler to guess. The solution is
to assign |undefined| at the beginning of the scope.
This situation is unusual. It happens only when the programmer
says that the assignment is unknown at the beginning of the
scope, as above.
The function /body/ stays where it is. The hoisting doesn't
affect it in any way. Moving the body would change its context
and meaning catastrophically. Don't touch it.
That doesn't answer the question. The problem is that the body can
capture variables that haven't been declared yet.
With suitable arrangements this becomes impossible.
Of course if you /insist/ on setting everything up and assigning
to the function name before the body's context exists, then you
can't expect it to work. Then you get an incorrect program that
will act as you say. That's to be expected if you allow calling
a function whose context doesn't exist.
But if you accept that when the context doesn't exist you should
assign |undefined| at the beginning of the scope, and only assign
the body later, when its context does exist, then you get a well-
defined and well-behaved program.
Then any premature call to the function will throw an error as
a call to undefined(). This will work in ways that the programmer
can understand well.
As far as I can tell, this is correct by definition. Can you find
an example where it breaks?
The /name/ becomes a var. Treat it like any other var. Hoist
it and assign |undefined|, exactly like you do with other vars.
That's incompatible with how functions are defined in ES3.
Incompatible? Your comment on my tiny two-line program says
that you want to throw a syntax error in these cases. Breaking
five platforms out of five is less compatible than what I
describe.
It certainly doesn't break JavaScript 1.7 in Firefox. It's
almost exactly what Firefox does in this special situation.
According to at least one discussion in the ES3.1 archives,
ES3 does not specify how you should treat a function in this
special situation. Here's one of these comments:
https://mail.mozilla.org/pipermail/es4-discuss/2008-July/003142.html
Are you saying that because this var is related to a function
it can't be treated like other vars? Is this var fundamentally
different from other vars? At least above it isn't.
Huh? Do you understand how ES3 works?
Yes, I know it quite well. I've been using JavaScript near
full-time on the client since Netscape 3, and on both server
and client since Microsoft ASP with JScript first appeared.
But why do you refer to ES3? ES3 doesn't have any scoping
blocks with var hoisting out of them.
Could you
