Loading ...
Sorry, an error occurred while loading the content.

[jslint] Re: Can we please lose the trailing whitespace and "this" restriction?

Expand Messages
  • douglascrockford
    ... JSLint was written for me. JavaScript is a language full of traps and sharp edges. I do not dare write in JavaScript without JSLint s assistance. I have
    Message 1 of 22 , Nov 14, 2011
    • 0 Attachment
      --- In jslint_com@yahoogroups.com, Erik Eckhardt <erik@...> wrote:

      > Who is jslint for? Who is the customer? Why is it being produced? What is
      > its actual purpose? Does the owner care if some portion of the user base is
      > alienated by unnanounced beraking changes? If we had these things clearly
      > stated, perhaps some of the customers might decide not to be so any more.
      > If the expectation is that jslint can and will be changed without notice in
      > a way that breaks actual money-making operations (that is, the day-to-day
      > business need of users to produce code without unneeded & costly roadblocks
      > randomly showing up), perhaps it is too risky to integrate it into our
      > businesses. Wait... we can just use an old version, right? Unless we find
      > some other tool that is also growing but doesn't have the same risks...


      JSLint was written for me. JavaScript is a language full of traps and sharp edges. I do not dare write in JavaScript without JSLint's assistance. I have made it available to the world for free. If it helps you too, then great. If it doesn't help you, then don't use it. I get paid the same either way.

      I have been improving JSLint with the assistance of the people who use it. As our understanding of the hazards in JavaScript improve, JSLint improves, and the subset of JavaScript that it recommends changes.

      Regrettably, this necessarily causes programs with previously undiscovered faults to, as you would say, berak. When that happens, the best advice is to follow JSLint's advice and fix your code.

      With respect to spaces at the end of lines: From a code hygiene perspective, there is no good argument for keeping them. If you are arguing that there is no one in your organization with the programming skill to remove them, then it seems your organization is suffering from problems that JSLint cannot help you solve. But in any case, the "Tolerate messy white space" option will ignore those spaces. But my advice is to not use that option and write smart looking code instead.
    • paulcrowder1979
      ... I agree with some of your points. JSLint is a valuable tool, and I do not write JavaScript without it. I also see the value in adding new rules as new
      Message 2 of 22 , Nov 16, 2011
      • 0 Attachment
        --- In jslint_com@yahoogroups.com, "douglascrockford" <douglas@...> wrote:
        >
        > JSLint was written for me. JavaScript is a language full of traps and sharp edges. I do not dare write in JavaScript without JSLint's assistance. I have made it available to the world for free. If it helps you too, then great. If it doesn't help you, then don't use it. I get paid the same either way.
        >
        > I have been improving JSLint with the assistance of the people who use it. As our understanding of the hazards in JavaScript improve, JSLint improves, and the subset of JavaScript that it recommends changes.
        >
        > Regrettably, this necessarily causes programs with previously undiscovered faults to, as you would say, berak. When that happens, the best advice is to follow JSLint's advice and fix your code.
        >
        > With respect to spaces at the end of lines: From a code hygiene perspective, there is no good argument for keeping them. If you are arguing that there is no one in your organization with the programming skill to remove them, then it seems your organization is suffering from problems that JSLint cannot help you solve. But in any case, the "Tolerate messy white space" option will ignore those spaces. But my advice is to not use that option and write smart looking code instead.
        >

        I agree with some of your points. JSLint is a valuable tool, and I do not write JavaScript without it. I also see the value in adding new rules as new hazards are discovered. What I don't understand is adding arbitrary rules that don't add to the quality of the code. You say there's no argument for keeping trailing whitespace, but I disagree. The best argument is that the whitespace is already there, it's doing no harm, and to remove it and enforce that all new code not have it would require a non-trivial amount of effort. That is effort that could be spent adding to the quality of our product, such as adding new features and fixing real bugs, rather than fiddling with code (wasn't it you who said "I highly recommend you don't fiddle with code"? https://developer.yahoo.com/yui/theater/video.php?v=crockonjs-5). I'm not sure why you jumped to the conclusion that no one in my organization has the skill to write a utility to remove trailing whitespace rather than perhaps considering the fact that writing the utility costs time and money; I think that says more about you than it does my organization.
      • Michael Mikowski
        Removing trailing whitespace IS trivial: sed -i -e s/ s*$// $(find ./ -type f) Of course if you are using notepad, it might take a few months.
        Message 3 of 22 , Nov 16, 2011
        • 0 Attachment
          Removing trailing whitespace IS trivial:

          sed -i -e 's/\s*$//' $(find ./ -type f)


          Of course if you are using notepad, it might take a few months.



          ________________________________
          From: paulcrowder1979 <paul@...>
          To: jslint_com@yahoogroups.com
          Sent: Wednesday, November 16, 2011 12:44 PM
          Subject: [jslint] Re: Can we please lose the trailing whitespace and "this" restriction?


           


          --- In jslint_com@yahoogroups.com, "douglascrockford" <douglas@...> wrote:
          >
          > JSLint was written for me. JavaScript is a language full of traps and sharp edges. I do not dare write in JavaScript without JSLint's assistance. I have made it available to the world for free. If it helps you too, then great. If it doesn't help you, then don't use it. I get paid the same either way.
          >
          > I have been improving JSLint with the assistance of the people who use it. As our understanding of the hazards in JavaScript improve, JSLint improves, and the subset of JavaScript that it recommends changes.
          >
          > Regrettably, this necessarily causes programs with previously undiscovered faults to, as you would say, berak. When that happens, the best advice is to follow JSLint's advice and fix your code.
          >
          > With respect to spaces at the end of lines: From a code hygiene perspective, there is no good argument for keeping them. If you are arguing that there is no one in your organization with the programming skill to remove them, then it seems your organization is suffering from problems that JSLint cannot help you solve. But in any case, the "Tolerate messy white space" option will ignore those spaces. But my advice is to not use that option and write smart looking code instead.
          >

          I agree with some of your points. JSLint is a valuable tool, and I do not write JavaScript without it. I also see the value in adding new rules as new hazards are discovered. What I don't understand is adding arbitrary rules that don't add to the quality of the code. You say there's no argument for keeping trailing whitespace, but I disagree. The best argument is that the whitespace is already there, it's doing no harm, and to remove it and enforce that all new code not have it would require a non-trivial amount of effort. That is effort that could be spent adding to the quality of our product, such as adding new features and fixing real bugs, rather than fiddling with code (wasn't it you who said "I highly recommend you don't fiddle with code"? https://developer.yahoo.com/yui/theater/video.php?v=crockonjs-5). I'm not sure why you jumped to the conclusion that no one in my organization has the skill to write a utility to remove trailing whitespace
          rather than perhaps considering the fact that writing the utility costs time and money; I think that says more about you than it does my organization.




          [Non-text portions of this message have been removed]
        • Michael Mikowski
          You want to do this independent of all other code work.  So the following workflow is recommended: git pull sed -i -e s/ s*$// $(find ./ -type f) // ensure
          Message 4 of 22 , Nov 16, 2011
          • 0 Attachment
            You want to do this independent of all other code work.  So the following workflow is recommended:

            git pull
            sed -i -e 's/\s*$//' $(find ./ -type f)
            // ensure nothing has been broken (run your executable against regression tests!):
            git diff
            git difftool

            // and now save the changes ye hath wrought:
            git commit -a -m 'Removed trailing whitespace in all files in repo'


            ________________________________
            From: Michael Mikowski <z_mikowski@...>
            To: "jslint_com@yahoogroups.com" <jslint_com@yahoogroups.com>
            Cc: paulcrowder1979 <paul@...>
            Sent: Wednesday, November 16, 2011 1:27 PM
            Subject: Re: [jslint] Re: Can we please lose the trailing whitespace and "this" restriction?


            Removing trailing whitespace IS trivial:

            sed -i -e 's/\s*$//' $(find ./ -type f)


            Of course if you are using notepad, it might take a few months.



            ________________________________
            From: paulcrowder1979 <paul@...>
            To: jslint_com@yahoogroups.com
            Sent: Wednesday, November 16, 2011 12:44 PM
            Subject: [jslint] Re: Can we please lose the trailing whitespace and "this" restriction?


             


            --- In jslint_com@yahoogroups.com, "douglascrockford" <douglas@...> wrote:
            >
            > JSLint was written for me. JavaScript is a language full of traps and sharp edges. I do not dare write in JavaScript without JSLint's assistance. I have made it available to the world for free. If it helps you too, then great. If it doesn't help you, then don't use it. I get paid the same either way.
            >
            > I have been improving JSLint with the assistance of the people who use it. As our understanding of the hazards in JavaScript improve, JSLint improves, and the subset of JavaScript that it recommends changes.
            >
            > Regrettably, this necessarily causes programs with previously undiscovered faults to, as you would say, berak. When that happens, the best advice is to follow JSLint's advice and fix your code.
            >
            > With respect to spaces at the end of lines: From a code hygiene perspective, there is no good argument for keeping them. If you are arguing that there is no one in your organization with the programming skill to remove them, then it seems your organization is suffering from problems that JSLint cannot help you solve. But in any case, the "Tolerate messy white space" option will ignore those spaces. But my advice is to not use that option and write smart looking code instead.
            >

            I agree with some of your points. JSLint is a valuable tool, and I do not write JavaScript without it. I also see the value in adding new rules as new hazards are discovered. What I don't understand is adding arbitrary rules that don't add to the quality of the code. You say there's no argument for keeping trailing whitespace, but I disagree. The best argument is that the whitespace is already there, it's doing no harm, and to remove it and enforce that all new code not have it would require a non-trivial amount of effort. That is effort that could be spent adding to the quality of our product, such as adding new features and fixing real bugs, rather than fiddling with code (wasn't it you who said "I highly recommend you don't fiddle with code"? https://developer.yahoo.com/yui/theater/video.php?v=crockonjs-5). I'm not sure why you jumped to the conclusion that no one in my organization has the skill to write a utility to remove trailing whitespace
            rather than perhaps considering the fact that writing the utility costs time and money; I think that says more about you than it does my organization.




            [Non-text portions of this message have been removed]
          • paulcrowder1979
            ... Writing the code is trivial. Pushing out an IDE plug-in to 100+ developers that will run the code when they save their files, waiting for the plug-in to
            Message 5 of 22 , Nov 17, 2011
            • 0 Attachment
              --- In jslint_com@yahoogroups.com, Michael Mikowski <z_mikowski@...> wrote:
              >
              > Removing trailing whitespace IS trivial:
              >
              > sed -i -e 's/\s*$//' $(find ./ -type f)
              >
              >
              > Of course if you are using notepad, it might take a few months.
              >

              Writing the code is trivial. Pushing out an IDE plug-in to 100+ developers that will run the code when they save their files, waiting for the plug-in to propagate to every developer, then updating our check-in policy to enforce the new rule is not trivial. I don't mind going through that effort when the pay-off is significant, but removing trailing whitespace has zero impact on the shipping product so it's not worth the effort.

              It looks like my only recourse at this point is to rip out the two lines of JSLint that enforce the trailing whitespace rule and the "this" strict violation rule each time I upgrade JSLint. I actually have more of a problem with the "this" restriction for reasons I mentioned earlier, but I have no confidence that it will be addressed after having read the opinion of his users that Douglas Crockford stated in his previous post.
            • Michael Mikowski
              Hey Paul: I feel your pain.  I don t know your process, environment, and restrictions.  So all this is speculation. If you are supporting 100+ developers,
              Message 6 of 22 , Nov 17, 2011
              • 0 Attachment
                Hey Paul:

                I feel your pain.  I don't know your process, environment, and restrictions.  So all this is speculation.

                If you are supporting 100+ developers, you certainly seem to have the critical mass to justify creating a forked version for your purposes.  You could then distribute to the developers via your SCMS, and using it in your build process.  Our official JSLint utility is kept in SCMS, and changes about once per quarter.  It is almost always adjusted to meet our standards that deviate from the JSLint way.

                I am very happy with that solution, and am very greatfull that I don't have to write 99% of our Lint utility, and that Mr. Crockford is kind enough to share this wonderful software with us, even if his idea of well formatted code doesn't agree with mine :)

                Sincerely, Mike

                ps here is my vim "plugin" for removing whitespace:

                "=====[ ;removes trailing space ]=====

                map ;k :%s?\s\+$??<CR>
                 
                pss

                If you wanted to enforce no-trailing-whitespace, you could update your entire repository as below, check-in, and then make developers aware that trailing whitespace will no longer be acceptable under JSLint.



                ________________________________
                From: paulcrowder1979 <paul@...>
                To: jslint_com@yahoogroups.com
                Sent: Thursday, November 17, 2011 7:31 AM
                Subject: [jslint] Re: Can we please lose the trailing whitespace and "this" restriction?


                 


                --- In jslint_com@yahoogroups.com, Michael Mikowski <z_mikowski@...> wrote:
                >
                > Removing trailing whitespace IS trivial:
                >
                > sed -i -e 's/\s*$//' $(find ./ -type f)
                >
                >
                > Of course if you are using notepad, it might take a few months.
                >

                Writing the code is trivial. Pushing out an IDE plug-in to 100+ developers that will run the code when they save their files, waiting for the plug-in to propagate to every developer, then updating our check-in policy to enforce the new rule is not trivial. I don't mind going through that effort when the pay-off is significant, but removing trailing whitespace has zero impact on the shipping product so it's not worth the effort.

                It looks like my only recourse at this point is to rip out the two lines of JSLint that enforce the trailing whitespace rule and the "this" strict violation rule each time I upgrade JSLint. I actually have more of a problem with the "this" restriction for reasons I mentioned earlier, but I have no confidence that it will be addressed after having read the opinion of his users that Douglas Crockford stated in his previous post.




                [Non-text portions of this message have been removed]
              • Jean-Charles Meyrignac
                ... In my company, we use SVN+CCNet+Nant/MsBuild If you have 100+ developers, you necessarily have a continuous build, and I m 100% sure that you can script
                Message 7 of 22 , Nov 17, 2011
                • 0 Attachment
                  On Thu, Nov 17, 2011 at 11:48 PM, Michael Mikowski <z_mikowski@...>wrote:

                  > **
                  >
                  >
                  > Our official JSLint utility is kept in SCMS, and changes about once per
                  > quarter. It is almost always adjusted to meet our standards that deviate
                  > from the JSLint way.
                  >
                  > +1
                  In my company, we use SVN+CCNet+Nant/MsBuild

                  If you have 100+ developers, you necessarily have a continuous build, and
                  I'm 100% sure that you can script your own build automation.
                  When people commit their source, the source could be automatically cleaned
                  (or you can clean it once, and force the developers to follow the rules).
                  Then, your build process should parse JS sources with JSLint, and block the
                  build if there are errors.
                  The goal of breaking the build is to force developers to acquire good
                  programming habits. It's very difficult to change habits, because nobody
                  wants to follow new rules.
                  Once the sources have been checked against your coding rules, the process
                  should run some unit tests, to validate that the basic functions are not
                  broken.

                  I agree that this solution is quite "brute", but you should see JsLint as
                  an ally instead of an enemy.
                  Personally, I tended to fight against software until I stopped trying to
                  force my wishes on it, and everything went smoothly afterwards.

                  Alternatively, you can use a modified version of JsLint in your build
                  process, and upgrade your local version every 6 months, as the parent post
                  explained.

                  JC


                  [Non-text portions of this message have been removed]
                • Nagy Endre
                  Hi everyone, maybe it s a little offtopic sorry, but can anyone explain why this inside the o object is refer to window ? Why it s not refer to the o
                  Message 8 of 22 , Nov 18, 2011
                  • 0 Attachment
                    Hi everyone, maybe it's a little offtopic sorry, but can anyone explain why "this" inside the "o" object is refer to window ?
                    Why it's not refer to the "o" object if we are inside there? I want to access from the returned object in function "f" the "o.p" property without using "o". Thanks!

                    var p = 2,
                        o  =  {
                            p: "a",
                            f: (function (m) {
                                return {
                                    l: m.p
                                };
                            }(this))
                        };

                    p = o.f.l;

                    Endre


                    [Non-text portions of this message have been removed]
                  • z_mikowski@yahoo.com
                    Javascript does not provide block scope, only functional scope. Therefore the this in your code takes the value as provided to the enclosing function. You
                    Message 9 of 22 , Nov 18, 2011
                    • 0 Attachment
                      Javascript does not provide block scope, only functional scope. Therefore the 'this' in your code takes the value as provided to the enclosing function. You could use the module pattern to get the results you are seeking, e.g.

                      var o = function (){
                      // private stuff
                      return {
                      p : "a",
                      // rest of object here
                      };
                      };



                      Nagy Endre <forewer2000@...> wrote:

                      >
                      >Hi everyone, maybe it's a little offtopic sorry, but can anyone explain why "this" inside the "o" object is refer to window ?
                      >Why it's not refer to the "o" object if we are inside there? I want to access from the returned object in function "f" the "o.p" property without using "o". Thanks!
                      >
                      >var p = 2,
                      >    o  =  {
                      >        p: "a",
                      >        f: (function (m) {
                      >            return {
                      >                l: m.p
                      >            };
                      >        }(this))
                      >    };
                      >
                      >p = o.f.l;
                      >
                      >Endre
                      >
                      >
                      >[Non-text portions of this message have been removed]
                      >


                      [Non-text portions of this message have been removed]
                    • Rob Richardson
                      Try this: var p = 2, o = { p: a }; o.f = (function (m) { return { l: m.p }; }(o)); p = o.f.l; Or this: var p = 2, o = { p: a }; o.prototype.f =
                      Message 10 of 22 , Nov 18, 2011
                      • 0 Attachment
                        Try this:

                        var p = 2,
                        o = {
                        p: "a"
                        };
                        o.f = (function (m) {
                        return {
                        l: m.p
                        };
                        }(o));
                        p = o.f.l;

                        Or this:

                        var p = 2,
                        o = {
                        p: "a"
                        };
                        o.prototype.f = function () {
                        return {
                        l: this.p
                        };
                        };
                        p = o.f.l;

                        Rob


                        -----Original Message-----
                        From: jslint_com@yahoogroups.com [mailto:jslint_com@yahoogroups.com] On
                        Behalf Of Nagy Endre
                        Sent: Friday, November 18, 2011 7:13 AM
                        To: jslint_com@yahoogroups.com
                        Subject: [jslint] this question




                        Hi everyone, maybe it's a little offtopic sorry, but can anyone explain why
                        "this" inside the "o" object is refer to window ?
                        Why it's not refer to the "o" object if we are inside there? I want to
                        access from the returned object in function "f" the "o.p" property without
                        using "o". Thanks!

                        var p = 2,
                        o = {
                        p: "a",
                        f: (function (m) {
                        return {
                        l: m.p
                        };
                        }(this))
                        };

                        p = o.f.l;

                        Endre

                        [Non-text portions of this message have been removed]
                      • stauren
                        I don t think this is about block scope or functional sope. This is probably because the function is executed when the script is being parsed, at which time
                        Message 11 of 22 , Nov 20, 2011
                        • 0 Attachment
                          I don't think this is about block scope or functional sope. This is
                          probably because the function is executed when the script is being parsed,
                          at which time the outer object is not fully established, so the the keyword
                          'this' points to the window object.

                          Another funny aspect of javascript.

                          ---
                          stauren (畅)
                          blog : http://stauren.net
                          email : stauren@...


                          On Sat, Nov 19, 2011 at 12:19 AM, <z_mikowski@...> wrote:

                          > **
                          >
                          >
                          > Javascript does not provide block scope, only functional scope. Therefore
                          > the 'this' in your code takes the value as provided to the enclosing
                          > function. You could use the module pattern to get the results you are
                          > seeking, e.g.
                          >
                          > var o = function (){
                          > // private stuff
                          > return {
                          > p : "a",
                          > // rest of object here
                          > };
                          > };
                          >
                          >
                          > Nagy Endre <forewer2000@...> wrote:
                          >
                          > >
                          > >Hi everyone, maybe it's a little offtopic sorry, but can anyone explain
                          > why "this" inside the "o" object is refer to window ?
                          > >Why it's not refer to the "o" object if we are inside there? I want to
                          > access from the returned object in function "f" the "o.p" property without
                          > using "o". Thanks!
                          > >
                          > >var p = 2,
                          > > o = {
                          > > p: "a",
                          > > f: (function (m) {
                          > > return {
                          > > l: m.p
                          > > };
                          > > }(this))
                          > > };
                          > >
                          > >p = o.f.l;
                          > >
                          > >Endre
                          > >
                          > >
                          > >[Non-text portions of this message have been removed]
                          > >
                          >
                          > [Non-text portions of this message have been removed]
                          >
                          >
                          >


                          [Non-text portions of this message have been removed]
                        • Nagy Endre
                          Yes, i realized that in {a: this} the value of a is the window value, not what we expect, because at runtime probably the object is not created yet, so
                          Message 12 of 22 , Nov 21, 2011
                          • 0 Attachment
                            Yes, i realized that in {a: this} the value of "a" is the "window" value, not what we expect, because at runtime probably the object is not created yet,
                            so only a method that is called later could have the correct "this" value.

                            I think it would be great if JSLINT could display an error in this situation like:
                            var o = {
                                a : this;
                            };

                            Because i think is there a little confusion here what will be the value of "this".



                            ________________________________
                            From: stauren <liuch.pku@...>
                            To: jslint_com@yahoogroups.com
                            Sent: Monday, November 21, 2011 9:49 AM
                            Subject: Re: [jslint] this question


                             
                            I don't think this is about block scope or functional sope. This is
                            probably because the function is executed when the script is being parsed,
                            at which time the outer object is not fully established, so the the keyword
                            'this' points to the window object.

                            Another funny aspect of javascript.

                            ---
                            stauren (畅)
                            blog : http://stauren.net
                            email : stauren@...

                            On Sat, Nov 19, 2011 at 12:19 AM, <z_mikowski@...> wrote:

                            > **
                            >
                            >
                            > Javascript does not provide block scope, only functional scope. Therefore
                            > the 'this' in your code takes the value as provided to the enclosing
                            > function. You could use the module pattern to get the results you are
                            > seeking, e.g.
                            >
                            > var o = function (){
                            > // private stuff
                            > return {
                            > p : "a",
                            > // rest of object here
                            > };
                            > };
                            >
                            >
                            > Nagy Endre <forewer2000@...> wrote:
                            >
                            > >
                            > >Hi everyone, maybe it's a little offtopic sorry, but can anyone explain
                            > why "this" inside the "o" object is refer to window ?
                            > >Why it's not refer to the "o" object if we are inside there? I want to
                            > access from the returned object in function "f" the "o.p" property without
                            > using "o". Thanks!
                            > >
                            > >var p = 2,
                            > > o = {
                            > > p: "a",
                            > > f: (function (m) {
                            > > return {
                            > > l: m.p
                            > > };
                            > > }(this))
                            > > };
                            > >
                            > >p = o.f.l;
                            > >
                            > >Endre
                            > >
                            > >
                            > >[Non-text portions of this message have been removed]
                            > >
                            >
                            > [Non-text portions of this message have been removed]
                            >
                            >
                            >

                            [Non-text portions of this message have been removed]




                            [Non-text portions of this message have been removed]
                          Your message has been successfully submitted and would be delivered to recipients shortly.