And here it is with PetitParser using the islands support. I think this is the nicest to read: codeParser := ( '<code>' asParser, '</code>' asParser negate star flatten , '</code>' asParser) ==> [ :t | t second asUppercase ]. parser := codeParser island star ==> [: t | '' join: t flatten ]. originalString := 'this is a string with <code>some funny lines</code> and here is another <code>haha</code>'. parser parse: originalString â--> this is a string with SOME FUNNY LINES and here is another HAHA" Cheers, Doru
On May 12, 2016, at 10:35 PM, Denis Kudriashov <dionisiydk@gmail.com> wrote:
Hi.
I always solve such kind of problems with streams. It is super easy and much easy then regex (I hate regex). For your case it would be something like:
in := source readStream. result := String streamContents: [:out | [in atEnd] whileFalse: [ out nextPutAll: (in upToAll: '<code>'). code := in upToAll: '</code>'. out nextPutAll: code asUppercase; nextPutAll: '</code>']. ]
And it could be much nicer with Xtreams but I not remember it API (maybe tomorrow I will remember).
2016-05-12 21:50 GMT+02:00 Mariano Martinez Peck <marianopeck@gmail.com>: Hi guys,
I wonder which is the easier way to do the following. I have a string which inside could have something like 'this is a string with <code>some funny lines</code> and here is another <code>haha</code>'. I need to parse that string, get all places where I have things surrounded with <code>SOMETHING</code>, get the "SOMETHING" (in previous example, that would be 'some funny lines'), execute that (this is something internal) , and from that I get the real string (imagine in this case the answer is 'SOME FUNNY LINES'). Finally, I need to replace the orignal string... So ... given the input:
'this is a string with <code>some funny lines</code> and here is another <code>haha</code>'
And given my specific domain logic transformation (in this example I assume a simple #asUppercase), I would like to get:
'this is a string with SOME FUNNY LINES and here is another HAHA'
I got it working with below lines. But it is a hack and terrible slow (I imagine). So...anyone has an idea how can I do this simpler/faster? Maybe some RB re-write rule?
Thanks in advance
| dom string originalString stringToBeAbleToParse xmlDocument replacements finalString | replacements := Dictionary new. originalString := 'this is a string with <code>some funny lines</code> and here is another <code>haha</code>'. stringToBeAbleToParse := '<hack>', originalString, '</hack>'. dom := XMLDOMParser on: stringToBeAbleToParse. dom configuration isValidating: false. xmlDocument := dom parseDocument. (xmlDocument allElementsNamed: 'code') do: [ :aXMLElement |
"Let's simulate my domain transformation logic as a simple #asUppercase" replacements at: aXMLElement asString put: (([:code | code asUppercase ]) value: aXMLElement nodes first asString). ]. finalString := originalString. replacements keysAndValuesDo: [ :originalText :new | finalString := finalString copyReplaceAll: originalText with: new. ]. finalString
-- Mariano http://marianopeck.wordpress.com
-- www.tudorgirba.com www.feenk.com "From an abstract enough point of view, any two things are similar."