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

Re: [jasspa] java indent and highligting

Expand Messages
  • Jon Green
    ... A few more to add to that maybe? ; fix throws/implements/extends indent at start of line, typically a ; statement continuation. indent .indent.java s
    Message 1 of 29 , Oct 4, 2006
    • 0 Attachment
      Jon Green wrote:
      > Bryan Schofield wrote:
      >> crap... I pasted in a line accidentally...
      >> delete the following:
      >>
      >> 0 indent .indent.java-enum 1 20
      >>
      >
      > Would you want to add this as well?
      > Jon.
      >
      > ; fix throws indent at start of line, typically a statement
      > ; continuation.
      > indent .indent.java s "^\\s*throws\\s" t
      >

      A few more to add to that maybe?

      ; fix throws/implements/extends indent at start of line, typically a
      ; statement continuation.
      indent .indent.java s "^\\s*throws\\s" t
      indent .indent.java s "^\\s*implements\\s" t
      indent .indent.java s "^\\s*extends\\s" t
    • Bryan Schofield
      I took a look at the java hilighting. It was missing some java keywords and it had a bunch of c keywords that don t really apply. Below is a suggestion for
      Message 2 of 29 , Oct 5, 2006
      • 0 Attachment
        I took a look at the java hilighting. It was missing some java
        keywords and it had a bunch of c keywords that don't really apply.
        Below is a suggestion for hilighting. In hkjava.emf if you replace
        from around line 60
        hilight .hilight.java 0x12 "import" .scheme.prepro
        to around line 131
        hilight .hilight.java 1 "virtual" .scheme.type



        ; annotations
        hilight .hilight.java 4 "@" "\\}\\W" "" .scheme.prepro
        ; constants: A_CONSTANT_VARAIBLE
        hilight .hilight.java 1 "[A-Z_]+" .scheme.constant
        ; class names: AJavaClassName
        hilight .hilight.java 1 "[A-Z]+[a-z]+\\w*" .scheme.type

        ; Java keywords
        ; The following conventions are used:
        ; .scheme.type primitive types
        ; .scheme.operator flow control keywords
        ; .scheme.prepro package declaration and imports
        ; .scheme.error reserved keywords that have not been implemented
        ; .scheme.keyword other java keywords that are not covered above
        hilight .hilight.java 1 "abstract" .scheme.keyword
        hilight .hilight.java 1 "assert" .scheme.keyword
        hilight .hilight.java 1 "boolean" .scheme.type
        hilight .hilight.java 1 "break" .scheme.operator
        hilight .hilight.java 1 "byte" .scheme.type
        hilight .hilight.java 1 "case" .scheme.operator
        hilight .hilight.java 1 "catch" .scheme.operator
        hilight .hilight.java 1 "char" .scheme.type
        hilight .hilight.java 1 "class" .scheme.keyword
        hilight .hilight.java 1 "const" .scheme.error ; reserved but
        not implemented in java
        hilight .hilight.java 1 "continue" .scheme.operator
        hilight .hilight.java 1 "default" .scheme.operator
        hilight .hilight.java 1 "do" .scheme.operator
        hilight .hilight.java 1 "double" .scheme.type
        hilight .hilight.java 1 "else" .scheme.operator
        hilight .hilight.java 1 "enum" .scheme.keyword
        hilight .hilight.java 1 "extends" .scheme.keyword
        hilight .hilight.java 1 "false" .scheme.constant
        hilight .hilight.java 1 "final" .scheme.keyword
        hilight .hilight.java 1 "finally" .scheme.operator
        hilight .hilight.java 1 "float" .scheme.type
        hilight .hilight.java 1 "for" .scheme.operator
        hilight .hilight.java 1 "goto" .scheme.error ; reserved but
        not implemented
        hilight .hilight.java 1 "if" .scheme.operator
        hilight .hilight.java 1 "implements" .scheme.keyword
        hilight .hilight.java 0x12 "import" .scheme.prepro
        hilight .hilight.java 1 "instanceof" .scheme.keyword
        hilight .hilight.java 1 "int" .scheme.type
        hilight .hilight.java 1 "interface" .scheme.keyword
        hilight .hilight.java 1 "long" .scheme.type
        hilight .hilight.java 1 "native" .scheme.keyword
        hilight .hilight.java 1 "new" .scheme.keyword
        hilight .hilight.java 1 "null" .scheme.constant
        hilight .hilight.java 0x12 "package" .scheme.prepro
        hilight .hilight.java 1 "private" .scheme.keyword
        hilight .hilight.java 1 "protected" .scheme.keyword
        hilight .hilight.java 1 "public" .scheme.keyword
        hilight .hilight.java 1 "return" .scheme.keyword
        hilight .hilight.java 1 "short" .scheme.type
        hilight .hilight.java 1 "static" .scheme.keyword
        hilight .hilight.java 1 "strictfp" .scheme.keyword
        hilight .hilight.java 1 "super" .scheme.keyword
        hilight .hilight.java 1 "switch" .scheme.operator
        hilight .hilight.java 1 "synchronized" .scheme.keyword
        hilight .hilight.java 1 "this" .scheme.keyword
        hilight .hilight.java 1 "throw" .scheme.keyword
        hilight .hilight.java 1 "throws" .scheme.keyword
        hilight .hilight.java 1 "transient" .scheme.keyword
        hilight .hilight.java 1 "true" .scheme.constant
        hilight .hilight.java 1 "try" .scheme.operator
        hilight .hilight.java 1 "void" .scheme.keyword
        hilight .hilight.java 1 "volatile" .scheme.keyword
        hilight .hilight.java 1 "while " .scheme.operator
      • Jon Green
        ... Thanks Bryan, I have re-worked hkjava.emf with all of the suggestions received to date. If you install this then delete the myjava.emf file (or rename to
        Message 3 of 29 , Oct 5, 2006
        • 0 Attachment
          Bryan Schofield wrote:
          > I took a look at the java hilighting. It was missing some java
          > keywords and it had a bunch of c keywords that don't really apply.
          > Below is a suggestion for hilighting. In hkjava.emf if you replace

          Thanks Bryan,

          I have re-worked hkjava.emf with all of the suggestions received to
          date. If you install this then delete the myjava.emf file (or rename to
          myjava-1.emf so that it is not picked up) and then execute the file (or
          restart the editor).

          This includes the token highlighting as suggested and the indent as
          suggested. The language template includes the new indent method and the
          existing 'C' method. The indent mode may be changed with
          esc-x major-mode-setup.

          I am proposing that this gets shipped with the next release. Is this OK
          with everybody? Does this solve the Java problems that people are
          experiencing?

          Thanks
          Jon.
        • Thomas Hundt
          Hold it. This release is unusable. Basic functionality has been broken: private static final Logger LOG = Logger.getLogger(PeakHourReportRunner.class); now
          Message 4 of 29 , Oct 5, 2006
          • 0 Attachment
            Hold it. This release is unusable.

            Basic functionality has been broken:

            private static final Logger LOG =
            Logger.getLogger(PeakHourReportRunner.class);

            now comes out as

            private static final Logger LOG =
            Logger.getLogger(PeakHourReportRunner.class);

            and if-else statements come out as

            if (something)
            do_something();
            else
            do_something_else();

            (Bryan feels the whole world should have to put in braces, but --
            idealism aside -- I know I'm going to be editing plenty of peoples' code
            without all the braces where he would like them. It IS a legitimate
            Java feature not to have to put in extra braces where they aren't
            necessary, and having the editor properly indent them will allow one to
            find out where one screwed it up, right? I mean that is why we are
            doing this, yes?)

            I'm rather have it work the way it used to, minus any of the new
            "improvements", if basic functionality like this goes out the window.

            Sorry, Jon. It was looking great there for a while, but this is too much.

            Regards,

            -Th


            Jon Green wrote:
            > Bryan Schofield wrote:
            >> I took a look at the java hilighting. It was missing some java
            >> keywords and it had a bunch of c keywords that don't really apply.
            >> Below is a suggestion for hilighting. In hkjava.emf if you replace
            >
            > Thanks Bryan,
            >
            > I have re-worked hkjava.emf with all of the suggestions received to
            > date. If you install this then delete the myjava.emf file (or rename to
            > myjava-1.emf so that it is not picked up) and then execute the file (or
            > restart the editor).
            >
            > This includes the token highlighting as suggested and the indent as
            > suggested. The language template includes the new indent method and the
            > existing 'C' method. The indent mode may be changed with
            > esc-x major-mode-setup.
            >
            > I am proposing that this gets shipped with the next release. Is this OK
            > with everybody? Does this solve the Java problems that people are
            > experiencing?
            >
            > Thanks
            > Jon.


            --
            Thomas Hundt <tom@...> +1-415-867-6698
          • Jon Green
            ... Hi Thomas, Thanks for the info and knee jerk reaction, just what I was looking for. M-x major-mode-setup will allow you to turn on C indentation again
            Message 5 of 29 , Oct 5, 2006
            • 0 Attachment
              Thomas Hundt wrote:
              > Hold it. This release is unusable.
              >
              > Basic functionality has been broken:
              >
              > private static final Logger LOG =
              > Logger.getLogger(PeakHourReportRunner.class);
              >
              > now comes out as
              >
              > private static final Logger LOG =
              > Logger.getLogger(PeakHourReportRunner.class);
              >
              > and if-else statements come out as
              >
              > if (something)
              > do_something();
              > else
              > do_something_else();
              >
              > (Bryan feels the whole world should have to put in braces, but --
              > idealism aside -- I know I'm going to be editing plenty of peoples' code
              > without all the braces where he would like them. It IS a legitimate
              > Java feature not to have to put in extra braces where they aren't
              > necessary, and having the editor properly indent them will allow one to
              > find out where one screwed it up, right? I mean that is why we are
              > doing this, yes?)
              >
              > I'm rather have it work the way it used to, minus any of the new
              > "improvements", if basic functionality like this goes out the window.
              >
              > Sorry, Jon. It was looking great there for a while, but this is too much.
              >
              > Regards,
              >
              > -Th
              >

              Hi Thomas,

              Thanks for the info and knee jerk reaction, just what I was looking for.

              M-x major-mode-setup

              will allow you to turn on 'C' indentation again this should restore what
              you had before. I'll make C mode the default and then leave the other
              method in for the 'Java pure' people.

              Regards
              Jon.
            • Jon Green
              ... OK, here is the next one! This time the default mode of operation is the built in indentation with the bracket alignment disabled. Note that your history
              Message 6 of 29 , Oct 5, 2006
              • 0 Attachment
                Thomas Hundt wrote:
                > Hold it. This release is unusable.
                >
                > Basic functionality has been broken:
                >
                > private static final Logger LOG =
                > Logger.getLogger(PeakHourReportRunner.class);
                >
                > now comes out as
                >
                > private static final Logger LOG =
                > Logger.getLogger(PeakHourReportRunner.class);
                >
                > and if-else statements come out as
                >
                > if (something)
                > do_something();
                > else
                > do_something_else();
                >
                > (Bryan feels the whole world should have to put in braces, but --
                > idealism aside -- I know I'm going to be editing plenty of peoples' code
                > without all the braces where he would like them. It IS a legitimate
                > Java feature not to have to put in extra braces where they aren't
                > necessary, and having the editor properly indent them will allow one to
                > find out where one screwed it up, right? I mean that is why we are
                > doing this, yes?)
                >
                > I'm rather have it work the way it used to, minus any of the new
                > "improvements", if basic functionality like this goes out the window.
                >
                > Sorry, Jon. It was looking great there for a while, but this is too much.
                >
                > Regards,
                >
                > -Th
                >

                OK, here is the next one!

                This time the default mode of operation is the built in indentation with
                the bracket alignment disabled.

                Note that your history will likely be retained so you will have to
                reconfigure the major mode to return to some default using

                M-x major-mode-setup

                The defaults are:

                Brace Hilighting:OFF
                Enable Bracket Alignment:OFF (Enables old brace alignment)
                Alternative Indent Mode:OFF (Pure Java mode)

                I understand that neither method is ideal. With the 'C' type indent then
                there are issues with "@" statements. These are not present in the
                Alternative indent - but this then needs to be fully bracketed.

                Regards
                Jon.
              • Bryan Schofield
                unfortunately, the indent syntax isn t powerful enough to describe the situtations needed. For example... if (true) continuedStatement(); if (true) {
                Message 7 of 29 , Oct 6, 2006
                • 0 Attachment
                  unfortunately, the indent syntax isn't powerful enough to describe the
                  situtations needed.

                  For example...

                  if (true)
                  continuedStatement();

                  if (true) {
                  indentedStatement();
                  }

                  if (true)
                  {
                  indentedStatement();
                  }

                  I do not know how to tell indent that the next line is a continuation,
                  unless it is a "{". I can write indent rules to handle the first 2 or
                  the last 2 but not all 3 at the same time.


                  indent .indent.java c "=\\s*$" t

                  handles:

                  private static final Logger LOG =
                  Logger.getLogger(PeakHourReportRunner.class);
                • jon@jasspa.com
                  ... Hi Bryan, We are aware of the short comings of the indentation. We have yet to come up with a better way of describing the indentation in terms of its
                  Message 8 of 29 , Oct 6, 2006
                  • 0 Attachment
                    > unfortunately, the indent syntax isn't powerful enough to describe the
                    > situtations needed.
                    >
                    > For example...
                    >
                    > if (true)
                    > continuedStatement();
                    >
                    > if (true) {
                    > indentedStatement();
                    > }
                    >
                    > if (true)
                    > {
                    > indentedStatement();
                    > }
                    >
                    > I do not know how to tell indent that the next line is a continuation,
                    > unless it is a "{". I can write indent rules to handle the first 2 or
                    > the last 2 but not all 3 at the same time.
                    >
                    >
                    > indent .indent.java c "=\\s*$" t
                    >
                    > handles:
                    >
                    > private static final Logger LOG =
                    > Logger.getLogger(PeakHourReportRunner.class);
                    >

                    Hi Bryan,

                    We are aware of the short comings of the indentation. We have yet to come
                    up with a better way of describing the indentation in terms of its
                    grammer. At the moment we seem to do reasonably well with very little
                    knowledge of the language in the template. All of the items that you have
                    raised above require a little more information about the syntax in order
                    to make them work properly. This is actually why C is burnt in because it
                    is easier to handle it like this than coming up with all of the correct
                    rules that are programmable.

                    We have considered how we might express this, but so far have not come up
                    with a sufficiently good way of minimally and generically describing the
                    syntax of the target language for indentation.

                    If anybody has any thoughts on how we can do this I would like to hear
                    them because it is not trivial.

                    Anybody looking for a Comp. Sci. project at University????

                    Regards
                    Jon.
                  • Bryan Schofield
                    I see I missed the official release... but the java indentation needs a little something to fix handling of Strings indent .indent.java e indent
                    Message 9 of 29 , Oct 9, 2006
                    • 0 Attachment
                      I see I missed the official release... but the java indentation needs
                      a little something to fix handling of Strings

                      indent .indent.java e "\"" "\"" "\\"
                      indent .indent.java e "\'" "\'" "\\"



                      On 10/6/06, jon@... <jon@...> wrote:
                      > > unfortunately, the indent syntax isn't powerful enough to describe the
                      > > situtations needed.
                      > >
                      > > For example...
                      > >
                      > > if (true)
                      > > continuedStatement();
                      > >
                      > > if (true) {
                      > > indentedStatement();
                      > > }
                      > >
                      > > if (true)
                      > > {
                      > > indentedStatement();
                      > > }
                      > >
                      > > I do not know how to tell indent that the next line is a continuation,
                      > > unless it is a "{". I can write indent rules to handle the first 2 or
                      > > the last 2 but not all 3 at the same time.
                      > >
                      > >
                      > > indent .indent.java c "=\\s*$" t
                      > >
                      > > handles:
                      > >
                      > > private static final Logger LOG =
                      > > Logger.getLogger(PeakHourReportRunner.class);
                      > >
                      >
                      > Hi Bryan,
                      >
                      > We are aware of the short comings of the indentation. We have yet to come
                      > up with a better way of describing the indentation in terms of its
                      > grammer. At the moment we seem to do reasonably well with very little
                      > knowledge of the language in the template. All of the items that you have
                      > raised above require a little more information about the syntax in order
                      > to make them work properly. This is actually why C is burnt in because it
                      > is easier to handle it like this than coming up with all of the correct
                      > rules that are programmable.
                      >
                      > We have considered how we might express this, but so far have not come up
                      > with a sufficiently good way of minimally and generically describing the
                      > syntax of the target language for indentation.
                      >
                      > If anybody has any thoughts on how we can do this I would like to hear
                      > them because it is not trivial.
                      >
                      > Anybody looking for a Comp. Sci. project at University????
                      >
                      > Regards
                      > Jon.
                      >
                      >
                      >
                      > __________________________________________________________________________
                      >
                      > This is an unmoderated list. JASSPA is not responsible for the content of
                      > any material posted to this list.
                      >
                      > To unsubscribe, send a mail message to
                      >
                      > mailto:jasspa-unsubscribe@yahoogroups.com
                      >
                      > or visit http://groups.yahoo.com/group/jasspa and
                      > modify your account settings manually.
                      >
                      >
                      >
                      > Yahoo! Groups Links
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                    • Bryan Schofield
                      Perhaps you add in a hook to allow a function to specified as the last argument in an indent rule. This function would return an indicator on whether the
                      Message 10 of 29 , Oct 9, 2006
                      • 0 Attachment
                        Perhaps you add in a hook to allow a function to specified as the last
                        argument in an indent rule. This function would return an indicator on
                        whether the indent rule should actually be applied. The function could
                        be passed the line number and do whatever language-specific analysis
                        is required.

                        indent .indent.java c "if.*)\\s*$" t java-check-if

                        define-macro java-check-if
                        ; @1 = line number
                        ; ...
                        set-variable $result true
                        !emacro
                      • jon@jasspa.com
                        ... OK Thanks, As it turns out it looks like we need a patch to jst.etf so I can slip this one in as well as they are both macro problems. Fortunately there
                        Message 11 of 29 , Oct 9, 2006
                        • 0 Attachment
                          > I see I missed the official release... but the java indentation needs
                          > a little something to fix handling of Strings
                          >
                          > indent .indent.java e "\"" "\"" "\\"
                          > indent .indent.java e "\'" "\'" "\\"
                          >

                          OK Thanks,

                          As it turns out it looks like we need a patch to jst.etf so I can slip
                          this one in as well as they are both macro problems. Fortunately there are
                          no binary problems at the moment, but does mean that some of the packages
                          have to be re-built.

                          This is why one release per year is more than enough for us! As usual it
                          does look like there are some more sweeping changes in the pipeline that
                          will come fairly soon after the release on the development side however we
                          cannot afford to delay this release any longer because of the amount of
                          time it takes us to prove it is stable.

                          As an aside: In building the 2006 release I was forced back to 20050505 on
                          some platforms where I had not installed 2006 yet and was surprised by
                          some of the features that were missing (e.g. being able to delete
                          directories in the file browser, infact all of the file browser controls).
                          Obviously most users will not be aware of this but we have been using
                          these internally for the last year and miss them when they are not there -
                          hence we really cannot delay this release any more.

                          Regards
                          Jon.
                        • jon@jasspa.com
                          ... I think you are right. I did try and prototype this type of schema about a year ago and got into a real mess because of problems with the macro context
                          Message 12 of 29 , Oct 9, 2006
                          • 0 Attachment
                            > Perhaps you add in a hook to allow a function to specified as the last
                            > argument in an indent rule. This function would return an indicator on
                            > whether the indent rule should actually be applied. The function could
                            > be passed the line number and do whatever language-specific analysis
                            > is required.
                            >
                            > indent .indent.java c "if.*)\\s*$" t java-check-if
                            >
                            > define-macro java-check-if
                            > ; @1 = line number
                            > ; ...
                            > set-variable $result true
                            > !emacro

                            I think you are right.

                            I did try and prototype this type of schema about a year ago and got into
                            a real mess because of problems with the macro context (i.e. you are in a
                            macro that then calls back into native and then want to come out and run a
                            macro again) - anyway I ran out of time doing some other stuff. I do
                            believe that this is the way to go but have not found enough time to
                            follow it through to a conclusion and get it working.

                            The only problem with this type of scheme is if you mess up in your macro
                            the editor will die a horrible death - error recovery is likely to be
                            difficult.

                            Jon.
                          Your message has been successfully submitted and would be delivered to recipients shortly.