Its great to see someone working on this!
And a terrible shame that you cannot reuse the syntax colorer we
already have! (but you may find the source code in
collects/syntax-color/scheme-lexer.rkt to be useful?)
Robby
On Mon, Aug 13, 2012 at 7:46 AM, Greg Hendershott
<greghendersh...@gmail.com> wrote:
> I spent more time on this, including trying to recognzie all the
> variations of numbers like these:
>
> (values
> ;; #b
> #b1.1
> #b-1.1
> #b1e1
> #b1/1
> #b0/1
> #b1e-1
> #b101
> #b2 ;highlight as error
>
> ;; #d
> #d-1.23
> #d1.123
> #d1e3
> #d1e-22
> #d1/2
> #d-1/2
> #d1
> #d-1
> #dZ ;highlight as error
>
> ;; No # reader prefix -- same as #d
> -1.23
> 1.123
> 1e3
> 1e-22
> 1/2
> -1/2
> 1
> -1
>
> ;; #e
> #e-1.23
> #e1.123
> #e1e3
> #e1e-22
> #e1
> #e-1
> #e1/2
> #e-1/2
> #eZ ;highlight as error
>
> ;; #i always float
> #i-1.23
> #i1.123
> #i1e3
> #i1e-22
> #i1/2
> #i-1/2
> #i1
> #i-1
> #iZ ;highlight as error
>
> ;; #o
> #o777.777
> #o-777.777
> #o777e777
> #o777e-777
> #o3/7
> #o-3/7
> #o777
> #o-777
> #08 ;highlight as error
>
> ;; #x
> #x-f.f
> #xf.f
> #x-f
> #xf
> #xG ;highlight as error
>
> )
>
> In this way (Python):
>
> ## numbers: Keep in mind Racket reader hash prefixes,
> ## which can denote the base or the type. These don't map
> ## neatly onto pygment token types; some judgment calls
> ## here. Note that none of these regexps attempt to
> ## exclude identifiers that start with a number, such as a
> ## variable named "100-Continue".
>
> # #b
> (r'#b[-+]?[01]+\.[01]+', Number.Float),
> (r'#b[01]+e[-+]?[01]+', Number.Float),
> (r'#b[-+]?[01]/[01]+', Number),
> (r'#b[-+]?[01]+', Number.Integer),
> (r'#b\S*', Error),
>
> # #d OR no hash prefix
> (r'(#d)?[-+]?\d+\.\d+', Number.Float),
> (r'(#d)?\d+e[-+]?\d+', Number.Float),
> (r'(#d)?[-+]?\d+/\d+', Number),
> (r'(#d)?[-+]?\d+', Number.Integer),
> (r'#d\S*', Error),
>
> # #e
> (r'#e[-+]?\d+\.\d+', Number.Float),
> (r'#e\d+e[-+]?\d+', Number.Float),
> (r'#e[-+]?\d+/\d+', Number),
> (r'#e[-+]?\d+', Number),
> (r'#e\S*', Error),
>
> # #i is always inexact-real, i.e. float
> (r'#i[-+]?\d+\.\d+', Number.Float),
> (r'#i\d+e[-+]?\d+', Number.Float),
> (r'#i[-+]?\d+/\d+', Number.Float),
> (r'#i[-+]?\d+', Number.Float),
> (r'#i\S*', Error),
>
> # #o
> (r'#o[-+]?[0-7]+\.[0-7]+', Number.Oct),
> (r'#o[0-7]+e[-+]?[0-7]+', Number.Oct),
> (r'#o[-+]?[0-7]+/[0-7]+', Number.Oct),
> (r'#o[-+]?[0-7]+', Number.Oct),
> (r'#o\S*', Error),
>
> # #x
> (r'#x[-+]?[0-9a-fA-F]+\.[0-9a-fA-F]+', Number.Hex),
> # the exponent variation (e.g. #x1e1) is N/A
> (r'#x[-+]?[0-9a-fA-F]+/[0-9a-fA-F]+', Number.Hex),
> (r'#x[-+]?[0-9a-fA-F]+', Number.Hex),
> (r'#x\S*', Error),
>
> Talk about brain burn. :)
>
> Also, Jens pointed out I should also handle curly braces. () = [] = {}
>
> https://bitbucket.org/greghendershott/pygments-main
>
> On Sun, Aug 12, 2012 at 6:50 PM, Sam Tobin-Hochstadt <sa...@ccs.neu.edu>
> wrote:
>> Wonderful! This will make the github experience much nicer.
>>
>> On Sun, Aug 12, 2012 at 6:36 PM, Greg Hendershott
>> <greghendersh...@gmail.com> wrote:
>>> So I followed up on this.
>>>
>>> https://bitbucket.org/greghendershott/pygments-main/changeset/240e51e2da13b079482f6b61c215280224f89f06
>>>
>>> ~~~~~
>>> Add RacketLexer.
>>>
>>> Previously Racket files were handled by SchemeLexer. Instead, use a
>>> proper RacketLexer, which handles Racket more appropriately:
>>>
>>> 1. Treat square brackets like parentheses.
>>> 2. Expanded list of keywords.
>>> 3. Different file extensions, MIME types, etc.
>>> 4. Handle #:keyword arguments.
>>> 5. Handle more number literals (e.g. #xFF, #o777, 2e2, #e232, etc.).
>>> 6. Handle #| ... |# multiline comments (although NOT nested).
>>> ~~~~~
>>>
>>> Before I give them a pull request, I wanted to give folks here a
>>> chance to critique it (or even flatly veto it).
>>>
>>> Again, the goal is that eventually this would flow through to GitHub,
>>> and improve the readability of Racket repos and gists.
>>>
>>> On Thu, Jun 2, 2011 at 12:49 PM, John Clements
>>> <cleme...@brinckerhoff.org> wrote:
>>>>
>>>> On Jun 2, 2011, at 6:43 AM, Eli Barzilay wrote:
>>>>
>>>>> An hour ago, Vincent St-Amour wrote:
>>>>>> At Thu, 2 Jun 2011 06:30:00 -0400,
>>>>>> Greg Hendershott wrote:
>>>>>>> aliases = ['scheme', 'scm', 'ss', 'racket', 'rkt']
>>>>>>
>>>>>> rktl would probably fit in there too.
>>>>>
>>>>> (This is just in case someone takes it on more seriously: IIRC,
>>>>> another issue with pygments was either ignoring square brackets, or
>>>>> highlighting them as errors.)
>>>>
>>>> Gosh, sounds complicated. I don't think any of us could handle that!
>>>>
>>>>
>>>> John (goad, goad) Clements
>>>>
>>>>
>>>> _________________________________________________
>>>> For list-related administrative tasks:
>>>> http://lists.racket-lang.org/listinfo/users
>>> ____________________
>>> Racket Users list:
>>> http://lists.racket-lang.org/users
>>
>>
>>
>> --
>> sam th
>> sa...@ccs.neu.edu
> ____________________
> Racket Users list:
> http://lists.racket-lang.org/users
____________________
Racket Users list:
http://lists.racket-lang.org/users
