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

Re: [ydn-javascript] Re: YUI Compresser could produce shorter output

Expand Messages
  • Philip Tellis
    ... Adding braces isn t purely a style convention. They re also used to ensure that the machine does what the programmer intended. Taking them out in
    Message 1 of 15 , Nov 2, 2008
    • 0 Attachment
      c dropped bits saying:

      > Remember, we're talking about compressed code, intended to be read
      > by browsers, not humans. Style conventions don't apply here.

      Adding braces isn't purely a style convention. They're also used to
      ensure that the machine does what the programmer intended. Taking them
      out in compressed code makes for problems which are very hard to debug
      because the code works in development, but not in production.
    • dav.glass@yahoo.com
      Conrad -- I used the wrong term (syntax error), to me code that doesn t run as expected is an error ;) So not putting the braces in and compressing the line
      Message 2 of 15 , Nov 2, 2008
      • 0 Attachment
        Conrad --

        I used the wrong term (syntax error), to me code that doesn't run as expected is an error ;)

        So not putting the braces in and compressing the line could result in code that doesn't run as you want it to.

        Dav

        Dav Glass
        dav.glass@...
        blog.davglass.com




        + Windows: n. - The most successful computer virus, ever. +
        + A computer without a Microsoft operating system is like a dog
        without bricks tied to its head +
        + A Microsoft Certified Systems Engineer is to computing what a
        McDonalds Certified Food Specialist is to fine cuisine +



        ----- Original Message ----
        > From: crlender <crlender@...>
        > To: ydn-javascript@yahoogroups.com
        > Sent: Sunday, November 2, 2008 3:51:36 PM
        > Subject: [ydn-javascript] Re: YUI Compresser could produce shorter output
        >
        > --- In ydn-javascript@yahoogroups.com, dav.glass@... wrote:
        > > You also need to remember that the Compressor is removing the line
        > breaks, so your example:
        > >
        > > for (var k in map)
        > > if (map.hasOwnProperty(k))
        > > if (!(k in target))
        > > target[k] = map[k];
        > >
        > > Would result in a syntax error when line breaks are removed.
        >
        > No, it wouldn't; it's perfectly legal ECMAScript. The curlies really
        > are optional (just like in C, Java, and other languages with C-like
        > syntax).
        >
        >
        > - Conrad
        >
        >
        > ------------------------------------
        >
        > Yahoo! Groups Links
        >
        >
        >
      • Philip Tellis
        ... Try this code: if (map.hasOwnProperty(k)) { if (!(k in target)) target[k] = map[k]; else target[k] = null; } It s still one statement inside the outer if,
        Message 3 of 15 , Nov 2, 2008
        • 0 Attachment
          c dropped bits saying:

          > --- In ydn-javascript@yahoogroups.com, dav.glass@... wrote:
          >> You also need to remember that the Compressor is removing the line
          > breaks, so your example:
          >>
          >> for (var k in map)
          >> if (map.hasOwnProperty(k))
          >> if (!(k in target))
          >> target[k] = map[k];
          >>
          >> Would result in a syntax error when line breaks are removed.
          >
          > No, it wouldn't; it's perfectly legal ECMAScript. The curlies really
          > are optional (just like in C, Java, and other languages with C-like
          > syntax).


          Try this code:

          if (map.hasOwnProperty(k)) {
          if (!(k in target))
          target[k] = map[k];
          else
          target[k] = null;
          }

          It's still one statement inside the outer if, but if you remove the
          braces, you confuse the engine. Take out the semi colons for more fun.


          In any event, what you're stating is a micro-optimisation that buys you
          nothing if you also gzip your javascript before serving. All those
          braces will compress into a single bit or something of that order.
        • Randall R Schulz
          ... Grotesque! Redundant punctuation does not accentuate readability, and any programmer who would make mistakes that result from reading code without braces
          Message 4 of 15 , Nov 2, 2008
          • 0 Attachment
            On Sunday 02 November 2008 15:30, D. Wayne Fincher wrote:
            > That would be great if it weren't bad programming technique.

            Grotesque! Redundant punctuation does not accentuate readability, and
            any programmer who would make mistakes that result from reading code
            without braces as if it did should either:

            1) Find a new career.
            2) Use Python.


            > I presume you're not suggesting that the compressor take out curly
            > braces if the programmer did use them.

            Redundant is redundant. Why not remove them? Besides, the compressed
            code is not meant for human consumption, it's mean to load ASAP.


            > In which case, we're left with cases where the programmer doesn't use
            > them. But I don't speak for myself here. Every "code conventions" or
            > "programming standards" I've ever read suggests that curly braces
            > always be used.

            Gross. Grotesque! Hideous!!

            Programming is about thinking. Auto-pilot "coding conventions" that try
            to protect the programmer from his own inattention are not ultimately
            productive.


            Randall Schulz
          • Randall R Schulz
            ... Clearly a source change that alters the code s semantics is not acceptable for the compressor. Randall Schulz
            Message 5 of 15 , Nov 2, 2008
            • 0 Attachment
              On Sunday 02 November 2008 15:52, Philip Tellis wrote:
              > c dropped bits saying:
              > > Remember, we're talking about compressed code, intended to be read
              > > by browsers, not humans. Style conventions don't apply here.
              >
              > Adding braces isn't purely a style convention. They're also used to
              > ensure that the machine does what the programmer intended. Taking
              > them out in compressed code makes for problems which are very hard to
              > debug because the code works in development, but not in production.

              Clearly a source change that alters the code's semantics is not
              acceptable for the compressor.


              Randall Schulz
            • Philip Tellis
              ... When 900 years old you reach, look as good, you will not. Hmm?
              Message 6 of 15 , Nov 2, 2008
              • 0 Attachment
                RRS dropped bits saying:

                > On Sunday 02 November 2008 15:30, D. Wayne Fincher wrote:
                >> That would be great if it weren't bad programming technique.
                >
                > Grotesque! Redundant punctuation does not accentuate readability, and


                When 900 years old you reach, look as good, you will not. Hmm?

                :)
              • crlender
                ... Just tested it, and it works just as expected. ... No, it s *two* statements inside the outer if . And those semicolons are not optional, so taking them
                Message 7 of 15 , Nov 2, 2008
                • 0 Attachment
                  --- In ydn-javascript@yahoogroups.com, Philip Tellis <philip@...> wrote:
                  > Try this code:
                  >
                  > if (map.hasOwnProperty(k)) {
                  > if (!(k in target))
                  > target[k] = map[k];
                  > else
                  > target[k] = null;
                  > }

                  Just tested it, and it works just as expected.

                  > It's still one statement inside the outer if, but if you remove the
                  > braces, you confuse the engine. Take out the semi colons for more fun.

                  No, it's *two* statements inside the outer 'if'. And those semicolons
                  are not optional, so taking them out *would* be a syntax error without
                  line breaks.

                  I'm not sure what you mean by the engine getting confused... I've
                  never seen a JavaScript implementation that couldn't correctly parse
                  control structures with or without redundant curlies. This is basic
                  stuff. Or maybe did you mean the compressor engine could get confused?

                  > In any event, what you're stating is a micro-optimisation that buys you
                  > nothing if you also gzip your javascript before serving. All those
                  > braces will compress into a single bit or something of that order.

                  Sure, it's not that big a deal, but the compressor's output *could*
                  be improved (even if only by a few hundred bytes per script) without
                  sacrificing code integrity. Whether the developers choose to implement
                  it that way is of course their choice.

                  Regarding gzip: if we could just assume that gzip were universally
                  available and supported, the compressor wouldn't have to go to all
                  the trouble... just strip out comments and let gzip do the rest.


                  - Conrad
                • Randall R Schulz
                  ... Umm... Yes? Unless, perhaps, you re an acolyte of Ray Kurzweil. However, I fail to apprehend the significance of the assertion... RRS
                  Message 8 of 15 , Nov 2, 2008
                  • 0 Attachment
                    On Sunday 02 November 2008 16:11, Philip Tellis wrote:
                    > RRS dropped bits saying:
                    > > On Sunday 02 November 2008 15:30, D. Wayne Fincher wrote:
                    > >> That would be great if it weren't bad programming technique.
                    > >
                    > > Grotesque! Redundant punctuation does not accentuate readability,
                    > > and
                    >
                    > When 900 years old you reach, look as good, you will not. Hmm?
                    >
                    > :)


                    Umm... Yes?

                    Unless, perhaps, you're an acolyte of Ray Kurzweil.


                    However, I fail to apprehend the significance of the assertion...


                    RRS
                  • crlender
                    ... expected is an error ;) ... in code that doesn t run as you want it to. Dav, I may be misreading you, but I don t see how the example I posted could be
                    Message 9 of 15 , Nov 2, 2008
                    • 0 Attachment
                      --- In ydn-javascript@yahoogroups.com, dav.glass@... wrote:
                      > I used the wrong term (syntax error), to me code that doesn't run as
                      expected is an error ;)
                      >
                      > So not putting the braces in and compressing the line could result
                      in code that doesn't run as you want it to.

                      Dav, I may be misreading you, but I don't see how the example I posted
                      could be interpreted in any other way than the example with braces. The
                      ECMAScript standard specifies how these structures are to be parsed and
                      evaluated; any difference would indicate a (severe) bug in the script
                      engine.

                      Just in case, let me mention that I personally do use braces even if
                      they're not necessary, but a compressor doesn't get any bonus points
                      for readable output.


                      - Conrad
                    • D. Wayne Fincher
                      Code conventions aren t just about making code more readable . Anyone that attacks my post on that one single issue is taking part in a straw man. That is,
                      Message 10 of 15 , Nov 2, 2008
                      • 0 Attachment
                        Code conventions aren't just about "making code more readable".

                        Anyone that attacks my post on that one single issue is taking part in
                        a straw man. That is, instead of presenting a rebuttal based on the
                        argument at hand, you take another argument which is easier to rebut
                        and refute that one.

                        In this case, I never mentioned anything about the readability of the
                        code. And yet, a few posts only rebutted me on that point.

                        Classic straw man approach. ;)


                        --- In ydn-javascript@yahoogroups.com, "crlender" <crlender@...> wrote:
                        >
                        > --- In ydn-javascript@yahoogroups.com, dav.glass@ wrote:
                        > > I used the wrong term (syntax error), to me code that doesn't run as
                        > expected is an error ;)
                        > >
                        > > So not putting the braces in and compressing the line could result
                        > in code that doesn't run as you want it to.
                        >
                        > Dav, I may be misreading you, but I don't see how the example I posted
                        > could be interpreted in any other way than the example with braces. The
                        > ECMAScript standard specifies how these structures are to be parsed and
                        > evaluated; any difference would indicate a (severe) bug in the script
                        > engine.
                        >
                        > Just in case, let me mention that I personally do use braces even if
                        > they're not necessary, but a compressor doesn't get any bonus points
                        > for readable output.
                        >
                        >
                        > - Conrad
                        >
                      Your message has been successfully submitted and would be delivered to recipients shortly.