Incorrect handling of precedence of >> and & bitwise operators

[cohler]

According to LUA 5.3 manual, << and >> are HIGHER precedence operators than &.

But luamin takes the following expression,

• (a & b) >> c

removes the parens, and minifies it to

• a&b>>c

which is NOT syntactically equivalent and therefore gives different results.

PLEASE FIX.

[lgfrbcsgo]

The AST returned by luaparse seems to be correct.

x = (a & b) >> c

{
  "type": "Chunk",
  "body": [
    {
      "type": "AssignmentStatement",
      "variables": [
        {
          "type": "Identifier",
          "name": "x",
          "isLocal": false
        }
      ],
      "init": [
        {
          "type": "BinaryExpression",
          "operator": ">>",
          "left": {
            "type": "BinaryExpression",
            "operator": "&",
            "left": {
              "type": "Identifier",
              "name": "a",
              "isLocal": false
            },
            "right": {
              "type": "Identifier",
              "name": "b",
              "isLocal": false
            },
            "inParens": true
          },
          "right": {
            "type": "Identifier",
            "name": "c",
            "isLocal": false
          }
        }
      ]
    }
  ],
  "globals": [
    {
      "type": "Identifier",
      "name": "x",
      "isLocal": false
    },
    {
      "type": "Identifier",
      "name": "a",
      "isLocal": false
    },
    {
      "type": "Identifier",
      "name": "b",
      "isLocal": false
    },
    {
      "type": "Identifier",
      "name": "c",
      "isLocal": false
    }
  ]
}

[cohler]

Please READ what I wrote:

But luamin takes the following expression,

(a & b) >> c
removes the parens, and minifies it to

a&b>>c
which is NOT syntactically equivalent and therefore gives different results.

PLEASE FIX.

[lgfrbcsgo]

I am sorry for the misunderstanding. I am not one of the maintainers. I've also encountered that bug. That's why I did some tests to see if it is a problem with luamin or luaparse.

[cohler]

This is a bug report filed under "mathiasbynens/luamin" and it refers to "luamin" in the report.

NOWHERE did I even mention "luaparse", a product with with which I am not even familiar.

So as I stated luamin has a bug, precisely as I reported. Is there ANYONE out there who knows something about this software or cares about it anymore? Clearly Matthias does not care and has not responded to anything in years...

[cs8898]

I Also encountered that problem and like to change it,

but now i ran into a problem (ok i just started reading the code, so any clue can help)

In the PRECEDENCE map, there are only the left prioritys listed, ist that sufficient?
Seems correct, because everything is right associative or?

[cohler]

As I have said multiple times now:

(a&b)>>c is NOT the same as a&b>>c

'>>' is a higher precedence operator than '&'

And NO everything is NOT right associative. Binary operators in LUA are all LEFT associative except for concatenation ('..') and exponentiation ('^').

So for the THIRD TIME....

There is a MAJOR BUG in luamin.

luamin removes the parentheses from (a&b)>>c WHICH IS INCORRECT and changes the results.

[cs8898]

https://github.com/FATH-Mechatronics/luamin

ITs working ;D

[cs8898]

Its not a bug ;D

Its juts Lua 5.2 conform

Binary ops where added in 5.3 and the minifier does not interprete these prioritys, so by adding the missing, and shifting it got working, im happy with that ;D

30min fix, for the stuff i needed

[cohler]

I'm using the luamin for Mac. And IT IS A BUG. And IT DOES NOT WORK CORRECTLY. Did you create a new build for Mac? Jonathan Cohler cohler@jonathancohler.com http://jonathancohler.com/ <http://jonathancohler.com/> <cohler@jonathancohler.com> <https://facebook.com/J.Cohler> <https://twitter.com/cohler> <https://youtube.com/cohler59> <https://instagram.com/JonathanCohler> Video: Cohler • Weber <https://www.youtube.com/watch?v=rUVBVn8S1N4> Video: Bozza Bucolique <https://www.youtube.com/watch?v=ptnkmhfw5sw> Video: Schumann Intermezzo <https://www.youtube.com/watch?v=m3uaW0D5tFk> Video: McAllister Black Dog <https://www.youtube.com/watch?v=v4XVow4gT-0> *This message (including any attachments) may contain confidential, proprietary, privileged and/or private information. The information is intended to be for the private use of the individual or entity designated above. If you are not the intended recipient of this message, please notify the sender immediately, and delete the message and any attachments. Any disclosure, reproduction, distribution or other use of this message or any attachments by an individual or entity other than the intended recipient is prohibited. *

…

On Mon, Feb 22, 2021 at 3:11 PM Chris ***@***.***> wrote: Its not a bug ;D Its juts Lua 5.2 conform Binary ops where added in 5.3 and the minifier does not interprete these prioritys, so by adding the missing, and shifting it got working, im happy with that ;D 30min fix, for the stuff i needed — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#76 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACEVYVBHPHB5PVABUSL2EPTTAK25FANCNFSM4KRRODAQ> .

[cs8898]

For me it worked, i use the Fork i created directly in yarn, because i just needed a single docker container i just told yarn to use that repo instead of the NPM one...

The brackets get minified correctl and aren't droppen anymore.

Just by adding the "New" lua 5.3 (or 5.4) precedences the minifier got working correctly for lua 5.4 source codes...

This minifier uses the precedences of 5.2 an in 5.2 there are no bitweis operations, they got added in 5.3.

And because the SRC tells 5.2 precedences i wouldnt say its a bug... Its just an oder standard of lua...

I havent put mutch effort into thinking about assoziative operations, i just applied the "logic" that the old src used.

With that single commit it got working for my 5.4 code...

The only reason why i modified the code is, that i dont wanted to use tmp variables to mimic the brackets...
It would have blown up the code and would be some kind of unreadable...

[cohler]

Did you create a new build for Mac? Jonathan Cohler cohler@jonathancohler.com http://jonathancohler.com/ <http://jonathancohler.com/> <cohler@jonathancohler.com> <https://facebook.com/J.Cohler> <https://twitter.com/cohler> <https://youtube.com/cohler59> <https://instagram.com/JonathanCohler> Video: Cohler • Weber <https://www.youtube.com/watch?v=rUVBVn8S1N4> Video: Bozza Bucolique <https://www.youtube.com/watch?v=ptnkmhfw5sw> Video: Schumann Intermezzo <https://www.youtube.com/watch?v=m3uaW0D5tFk> Video: McAllister Black Dog <https://www.youtube.com/watch?v=v4XVow4gT-0> *This message (including any attachments) may contain confidential, proprietary, privileged and/or private information. The information is intended to be for the private use of the individual or entity designated above. If you are not the intended recipient of this message, please notify the sender immediately, and delete the message and any attachments. Any disclosure, reproduction, distribution or other use of this message or any attachments by an individual or entity other than the intended recipient is prohibited. *

…

On Mon, Feb 22, 2021 at 5:45 PM Chris ***@***.***> wrote: For me it worked, i use the Fork i created directly in yarn, because i just needed a single docker container i just told yarn to use that repo instead of the NPM one... The brackets get minified correctl and aren't droppen anymore. Just by adding the "New" lua 5.3 (or 5.4) precedences the minifier got working correctly for lua 5.4 source codes... This minifier uses the precedences of 5.2 an in 5.2 there are no bitweis operations, they got added in 5.3. And because the SRC tells 5.2 precedences i wouldnt say its a bug... Its just an oder standard of lua... I havent put mutch effort into thinking about assoziative operations, i just applied the "logic" that the old src used. With that single commit it got working for my 5.4 code... The only reason why i modified the code is, that i dont wanted to use tmp variables to mimic the brackets... It would have blown up the code and would be some kind of unreadable... — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#76 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACEVYVGXFSR4H3S646IOZYTTALM7FANCNFSM4KRRODAQ> .

[cs8898]

Ok i started to write a roman,
but now i convinced myself not to argue with you.

Here are some links, you will find out how it works...
https://stackoverflow.com/questions/43411864/how-to-install-package-from-github-repo-in-yarn
https://stackoverflow.com/questions/17509669/how-to-install-an-npm-package-from-github-directly

And you probably will understand why there is no need to "compile" it specially for the Mac OperatingSystem.

Be aware that also the fork i made is just for lua 5.3!
I don't want to hear anything about constants or stuff introduced in lua 5.4 or above

[cohler]

So if I do (as the repository says): npm install -g luamin Will it install the new supposedly working version with the precedence bug corrected? Jonathan Cohler cohler@jonathancohler.com http://jonathancohler.com/ <http://jonathancohler.com/> <cohler@jonathancohler.com> <https://facebook.com/J.Cohler> <https://twitter.com/cohler> <https://youtube.com/cohler59> <https://instagram.com/JonathanCohler> Video: Cohler • Weber <https://www.youtube.com/watch?v=rUVBVn8S1N4> Video: Bozza Bucolique <https://www.youtube.com/watch?v=ptnkmhfw5sw> Video: Schumann Intermezzo <https://www.youtube.com/watch?v=m3uaW0D5tFk> Video: McAllister Black Dog <https://www.youtube.com/watch?v=v4XVow4gT-0> *This message (including any attachments) may contain confidential, proprietary, privileged and/or private information. The information is intended to be for the private use of the individual or entity designated above. If you are not the intended recipient of this message, please notify the sender immediately, and delete the message and any attachments. Any disclosure, reproduction, distribution or other use of this message or any attachments by an individual or entity other than the intended recipient is prohibited. *

…

On Tue, Feb 23, 2021 at 11:29 AM Chris ***@***.***> wrote: Ok i started to write a roman, but now i convinced myself not to argue with you. Here are some links, you will find out how it works... https://stackoverflow.com/questions/43411864/how-to-install-package-from-github-repo-in-yarn https://stackoverflow.com/questions/17509669/how-to-install-an-npm-package-from-github-directly And you probably will understand why there is no need to "compile" it specially for the Mac OperatingSystem. Be aware that also the fork i made is just for lua 5.3! I don't want to hear anything about constants or stuff introduced in lua 5.4 or above — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#76 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACEVYVAGHXSOZPPSLUF25MDTAPJXDANCNFSM4KRRODAQ> .

[Jerakin]

He is telling you to install his fork. I don't use npm but I think it would be like this.

npm install git+https://github.com/FATH-Mechatronics/luamin.git