[ANN] Lua 5.3.1 (final) now available

classic Classic list List threaded Threaded
30 messages Options
12
Reply | Threaded
Open this post in threaded view
|

[ANN] Lua 5.3.1 (final) now available

Luiz Henrique de Figueiredo
Lua 5.3.1 has been frozen and is now available at
        http://www.lua.org/ftp/lua-5.3.1.tar.gz

MD5 a1b0a7e92d0c85bbff7a8d27bf29f8af
SHA1 1c46d1c78c44039939e820126b86a6ae12dadfba

Lua 5.3.1 is now the current release of Lua 5.3.

A test suite is available at
        http://www.lua.org/tests/

We thank everyone for their feedback on Lua 5.3 till now.

All feedback welcome. Thanks.
--lhf

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Ignacio Burgueño-2

On Wed, Jun 17, 2015 at 8:23 AM, Luiz Henrique de Figueiredo <[hidden email]> wrote:
Lua 5.3.1 has been frozen and is now available at
        http://www.lua.org/ftp/lua-5.3.1.tar.gz


Congrats on the release!
Just a little reminder. The sources at http://www.lua.org/source/5.3/ still point to 5.3.0

Regards,
Ignacio
Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Luiz Henrique de Figueiredo
> Just a little reminder. The sources at http://www.lua.org/source/5.3/ still
> point to 5.3.0

It'll be fixed by tomorrow. Thanks.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Tom N Harris-2
In reply to this post by Luiz Henrique de Figueiredo
I see you put a check to not save 'return' in readline history. Cool, thanks.
But it still doesn't handle expressions in parenthesis.

Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
> (1 +
>> 1)
>>

or this should print XY

Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
> ("x"..
>> "y"):upper()
>

The patch I had previously suggested does not have this flaw.

http://lua-users.org/lists/lua-l/2014-12/msg00336.html

--
tom <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Brigham Toskin
On Wed, Jun 17, 2015 at 3:23 PM, Tom N Harris <[hidden email]> wrote:
I see you put a check to not save 'return' in readline history. Cool, thanks.
But it still doesn't handle expressions in parenthesis.

Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
> (1 +
>> 1)
>>

I'm seeing it in 5.3.0 as well, but doesn't  happen if you prepend '='. Might be a clue.

    Lua 5.3.0  Copyright (C) 1994-2015 Lua.org, PUC-Rio
    > =(1+
    >> 1)
    2
 
or this should print XY

Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
> ("x"..
>> "y"):upper()
>

Same. The explicit "print returned values" works.
 

The patch I had previously suggested does not have this flaw.

http://lua-users.org/lists/lua-l/2014-12/msg00336.html

Haven't looked at this in enough depth to have an opinion. Thanks for submitting it though!

--
Brigham Toskin
Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Dirk Laurie-2
In reply to this post by Tom N Harris-2
2015-06-18 0:23 GMT+02:00 Tom N Harris <[hidden email]>:
> I see you put a check to not save 'return' in readline history. Cool, thanks.
> But it still doesn't handle expressions in parenthesis.

Correctly so, Open parenthesis as a way to claim the next line as
a continuation is a Python convention. Not Lua.

>
> Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
>> (1 +
>>> 1)
>>>
>
> or this should print XY

Not "should". At most "might by some be expected to".

>
> Lua 5.3.1  Copyright (C) 1994-2015 Lua.org, PUC-Rio
>> ("x"..
>>> "y"):upper()
>>

The manual says:

   In interactive mode, Lua repeatedly prompts and waits for a line.
   After reading a line, Lua first try to interpret the line as an expression.
   If it succeeds, it prints its value. Otherwise, it interprets the line as
   a statement. If you write an incomplete statement, the interpreter
   waits for its completion by issuing a different prompt.

It's quite clear: printing the result is only available for a one-line
expression.

In the first example, what you have is an incomplete statement.
Enter a semicolon to force completion and get, correctly, an error
message for the dangling expression.

In the second example, the statement is indeed complete, but
since it is not on one line, the result iss not printed.

> The patch I had previously suggested does not have this flaw.

It is not a flaw. It is a deliberate design decision, since for most
people, if the parentheses on the first line they type in an
interactive session are unbalanced, it is a mistake.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Tom N Harris-2
On Thursday, June 18, 2015 01:12:27 PM Dirk Laurie wrote:
> Correctly so, Open parenthesis as a way to claim the next line as
> a continuation is a Python convention. Not Lua.
>

A Lua expression may be enclosed in a parenthesis and it is still an
expression. There's nothing pythonic about that. And the Lua interpreter has
also made use of it.

> x = (1
>> + 1)
>

If the interpreter is going to accept expressions then it should do so without
quirky corner-cases.

> 1 + math.random()
1.8401877167635
> 1 + math.random(
stdin:1: unexpected symbol near '1'
> math.random()
0.39438292663544
> math.random(
>> )
>

--
tom <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Roberto Ierusalimschy
> If the interpreter is going to accept expressions then it should do so
> without quirky corner-cases.

Maybe it is only me, but this "it should do" sounds quite demanding
(mainly coming from a non-paying customer :-)

-- Roberto

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Soni "They/Them" L.


On 18/06/15 04:04 PM, Roberto Ierusalimschy wrote:
>> If the interpreter is going to accept expressions then it should do so
>> without quirky corner-cases.
> Maybe it is only me, but this "it should do" sounds quite demanding
> (mainly coming from a non-paying customer :-)
Sure, what's your bitcoin address?
>
> -- Roberto
>

--
Disclaimer: these emails are public and can be accessed from <TODO: get a non-DHCP IP and put it here>. If you do not agree with this, DO NOT REPLY.


Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Tom N Harris-2
In reply to this post by Roberto Ierusalimschy
On Thursday, June 18, 2015 04:04:22 PM Roberto Ierusalimschy wrote:
> Maybe it is only me, but this "it should do" sounds quite demanding
> (mainly coming from a non-paying customer :-)

By RFC 2119 "must" is a demand, "should" is a recommendation, and "may" is an
optional suggestion.

I think the interpreter would be improved by this and I'm defending my
position. If a wheel wants grease it must squeak, as the idiom would go.

--
tom <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Dirk Laurie-2
In reply to this post by Tom N Harris-2
2015-06-18 20:43 GMT+02:00 Tom N Harris <[hidden email]>:
> On Thursday, June 18, 2015 01:12:27 PM Dirk Laurie wrote:

> If the interpreter is going to accept expressions then it should do so without
> quirky corner-cases.

None of these is quirky or a corner case. All behave exactly as documented.

>> 1 + math.random()
> 1.8401877167635

Complete expression on one line --> evaluated and printed.

>> 1 + math.random(
> stdin:1: unexpected symbol near '1'

Code that cannot possibly be a prefix to a valid function body --> error

>> math.random()
> 0.39438292663544

Complete expression on one line --> evaluated and printed.

>> math.random(
>>> )
>>

First line is not an expression but could be a prefix to a valid function
body --> another line is added.
Second line completes it. It is a function call --> performed but
return values are discarded.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Tom N Harris-2
On Thursday, June 18, 2015 11:08:39 PM Dirk Laurie wrote:
> None of these is quirky or a corner case. All behave exactly as documented.

The description isn't even strictly correct. Not only expressions are printed.
"math.random();" is a statement that is printed. (Because of course it's not
checking if the input is an expression but rather can the input be compiled
with "return " prepended.)

When a statement has an open parenthesis or long-string or table constructor,
you can fix it by adding the closing token and be on your way. Why not do the
same for expressions? What utility is gained by not allowing multi-line
expressions? What utility is lost by having them? All I see are shortcomings,
such as not being able to type a string with a newline in it.

If "because that's not what the manual says" were a good enough reason we'd
never have expressions in interactive mode. It was simpler then, even if our
pinky fingers would be exercised pressing '=' so often. But I never had a
problem with that. Lone expressions are not valid syntax so it made sense not
to accept them.

Having expressions lets Lua act more like a desk calculator. But then the
legacy restrictions on statement-versus-expression act as a stumbling block
when you're typing an expression and either forget a parenthesis or want to
break it into multiple lines to be easier to read.

You can still use the '=' shortcut, but that has been omitted from the 5.3
manual, I presume to indicate that it is deprecated. But if Lua is never going
to allow multi-line expressions in interactive mode, then I'd prefer the '='
shortcut remain as a supported feature.

Or go back to the old statements-only interpreter and if you want a full-
featured interactive mode you can install luaprompt, which does handle multi-
line expressions. (Well, almost. It won't continue "1 +".)

--
tom <[hidden email]>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Dirk Laurie-2
2015-06-19 23:45 GMT+02:00 Tom N Harris <[hidden email]>:

> You can still use the '=' shortcut, but that has been omitted from the 5.3
> manual, I presume to indicate that it is deprecated. But if Lua is never going
> to allow multi-line expressions in interactive mode, then I'd prefer the '='
> shortcut remain as a supported feature.

I'll support that.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Lorenzo Donati-3
In reply to this post by Luiz Henrique de Figueiredo
Hi all!
Sorry to revive an old thread...

On 17/06/2015 13:23, Luiz Henrique de Figueiredo wrote:
> Lua 5.3.1 has been frozen and is now available at
> http://www.lua.org/ftp/lua-5.3.1.tar.gz
>
> MD5 a1b0a7e92d0c85bbff7a8d27bf29f8af
> SHA1 1c46d1c78c44039939e820126b86a6ae12dadfba
>

I just downloaded that package and checked its signatures. I get a
mismatch for both:

MD5: 797adacada8d85761c079390ff1d9961
SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee

I checked with several programs I have and I got the same results. I
also tried to re-download the package, fearing a corrupted download.
Nope, just the same results.

Is it just me or what else? Thanks!

-- Lorenzo


> Lua 5.3.1 is now the current release of Lua 5.3.
>
> A test suite is available at
> http://www.lua.org/tests/
>
> We thank everyone for their feedback on Lua 5.3 till now.
>
> All feedback welcome. Thanks.
> --lhf
>
>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Dirk Laurie-2
2015-07-19 19:17 GMT+02:00 Lorenzo Donati <[hidden email]>:

> On 17/06/2015 13:23, Luiz Henrique de Figueiredo wrote:
>>
>> Lua 5.3.1 has been frozen and is now available at
>>         http://www.lua.org/ftp/lua-5.3.1.tar.gz
>>
>> MD5     a1b0a7e92d0c85bbff7a8d27bf29f8af
>> SHA1    1c46d1c78c44039939e820126b86a6ae12dadfba
>>
>
> I just downloaded that package and checked its signatures. I get a mismatch
> for both:
>
> MD5: 797adacada8d85761c079390ff1d9961
> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee
>
> I checked with several programs I have and I got the same results. I also
> tried to re-download the package, fearing a corrupted download. Nope, just
> the same results.
>
> Is it just me or what else? Thanks!

Probably just you.

I get the correct checksums, having downloaded the
package on June 17 and again now. I used wget.
How did you download it?

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Ashwin Hirschi
In reply to this post by Lorenzo Donati-3

>> http://www.lua.org/ftp/lua-5.3.1.tar.gz
>
> I just downloaded that package and checked its signatures. I get a  
> mismatch for both:
>
> MD5: 797adacada8d85761c079390ff1d9961
> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee

According to http://www.lua.org/download.html there is no mismatch.

Your checksums match exactly (and are the same ones I get).

Ashwin.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Lorenzo Donati-3


On 19/07/2015 20:17, Ashwin Hirschi wrote:

>
>>>     http://www.lua.org/ftp/lua-5.3.1.tar.gz
>>
>> I just downloaded that package and checked its signatures. I get a
>> mismatch for both:
>>
>> MD5: 797adacada8d85761c079390ff1d9961
>> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee
>
> According to http://www.lua.org/download.html there is no mismatch.

Yep! Just checked. You're right!

So the signatures in the message of lhf I replied to are wrong. But
@Dirk seems to have downloaded a package that matches these older
signatures so there is a problem somewhere. Was the release repackaged
for some reason? I haven't seen any announce in the list about that.
Maybe I missed a message (I'm not too active on the list lately).

Cheers!

-- Lorenzo




>
> Your checksums match exactly (and are the same ones I get).
>
> Ashwin.
>
>

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Jonathan Goble
On Sun, Jul 19, 2015 at 2:47 PM, Lorenzo Donati
<[hidden email]> wrote:

> On 19/07/2015 20:17, Ashwin Hirschi wrote:
>>
>>
>>>>     http://www.lua.org/ftp/lua-5.3.1.tar.gz
>>>
>>>
>>> I just downloaded that package and checked its signatures. I get a
>>> mismatch for both:
>>>
>>> MD5: 797adacada8d85761c079390ff1d9961
>>> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee
>>
>>
>> According to http://www.lua.org/download.html there is no mismatch.
>
>
> Yep! Just checked. You're right!
>
> So the signatures in the message of lhf I replied to are wrong. But @Dirk
> seems to have downloaded a package that matches these older signatures so
> there is a problem somewhere. Was the release repackaged for some reason? I
> haven't seen any announce in the list about that. Maybe I missed a message
> (I'm not too active on the list lately).
>
> Cheers!

The 797ada... and 1676c6... checksums match the Lua 5.3.0 gzipped
tarball that I downloaded in early February.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Dirk Laurie-2
In reply to this post by Lorenzo Donati-3
2015-07-19 20:47 GMT+02:00 Lorenzo Donati <[hidden email]>:

>
>
> On 19/07/2015 20:17, Ashwin Hirschi wrote:
>>
>>
>>>>     http://www.lua.org/ftp/lua-5.3.1.tar.gz
>>>
>>>
>>> I just downloaded that package and checked its signatures. I get a
>>> mismatch for both:
>>>
>>> MD5: 797adacada8d85761c079390ff1d9961
>>> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee
>>
>>
>> According to http://www.lua.org/download.html there is no mismatch.
>
>
> Yep! Just checked. You're right!
>
> So the signatures in the message of lhf I replied to are wrong. But @Dirk
> seems to have downloaded a package that matches these older signatures so
> there is a problem somewhere. Was the release repackaged for some reason? I
> haven't seen any announce in the list about that. Maybe I missed a message
> (I'm not too active on the list lately).

Sorry! I calculated checksums, saw them in your message, and
only now on rereading do I notice that they are your new checksums,
not the old ones.

So the webpage (and LHF's email) give the wrong checksums.
Jonathan's post leads one to suspect a possible mechanism for
the error.

Reply | Threaded
Open this post in threaded view
|

Re: [ANN] Lua 5.3.1 (final) now available

Lorenzo Donati-3
In reply to this post by Jonathan Goble


On 19/07/2015 20:56, Jonathan Goble wrote:

> On Sun, Jul 19, 2015 at 2:47 PM, Lorenzo Donati
> <[hidden email]> wrote:
>> On 19/07/2015 20:17, Ashwin Hirschi wrote:
>>>
>>>
>>>>>      http://www.lua.org/ftp/lua-5.3.1.tar.gz
>>>>
>>>>
>>>> I just downloaded that package and checked its signatures. I get a
>>>> mismatch for both:
>>>>
>>>> MD5: 797adacada8d85761c079390ff1d9961
>>>> SHA1: 1676c6a041d90b6982db8cef1e5fb26000ab6dee
>>>
>>>
>>> According to http://www.lua.org/download.html there is no mismatch.
>>
>>
>> Yep! Just checked. You're right!
>>
>> So the signatures in the message of lhf I replied to are wrong. But @Dirk
>> seems to have downloaded a package that matches these older signatures so
>> there is a problem somewhere. Was the release repackaged for some reason? I
>> haven't seen any announce in the list about that. Maybe I missed a message
>> (I'm not too active on the list lately).
>>
>> Cheers!
>
> The 797ada... and 1676c6... checksums match the Lua 5.3.0 gzipped
> tarball that I downloaded in early February.
>
>

Yep! Just checked!
So it seems Luiz put the wrong signatures in the original post. Funny
no-one came up with the issue before! :-)


-- Lorenzo

12