XML usage question..

Hi all..

So, I've got an XML file that looks like the following :

<?xml version="1.0"?>
<!DOCTYPE requirements [
<!ELEMENT requirements (UseCase+)>
<!ELEMENT UseCase (#PCDATA)>
<!ELEMENT UseCases (UseCase+)>
]>
<requirements>
        <UseCases>
                <UseCase>Use-Case #1</UseCase>
                <UseCase>Use-Case #2</UseCase>
        </UseCases>
</requirements>


I then have the following code to parse & read in the XML :

==============================================================
buildUseCaseCollectionFrom: aList [
    aList detect: [:each | Transcript show: 'Element:' , (each
characterData); cr ]
]

  stream := './requirements.xml' asFile readStream.
  parser := XML.XMLParser new.
  parser validate: false.
  parser parse: stream.

  docroot := parser document root.
  self buildUseCaseCollectionFrom: (docroot elementsNamed: 'UseCases').

==============================================================

When I run it, I get the following output -- seems
like the Collection of Use-Case Elements is not
quite right.. I was expecting two "Element:" lines
each with the respective Use-Case # element but not
both bunched up as a single.. Any ideas on what I've
messed up?

Element:
                 Use-Case #1
                 Use-Case #2

Object: TextCollector new "<-0x53b6698>" error: Invalid argument
Transcript: must be a Boolean
SystemExceptions.MustBeBoolean(Exception)>>signal (ExcHandling.st:254)
SystemExceptions.MustBeBoolean class(SystemExceptions.WrongClass
class)>>signalOn:mustBe: (SysExcept.st:805)
SystemExceptions.MustBeBoolean class>>signalOn: (SysExcept.st:881)
TextCollector(Object)>>mustBeBoolean (Object.st:1399)
[] in Array(Iterable)>>detect:ifNone: (Iterable.st:107)
Array(SequenceableCollection)>>do: (SeqCollect.st:812)
Array(Iterable)>>detect:ifNone: (Iterable.st:108)
Array(Iterable)>>detect: (Iterable.st:172)




_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Il 25/09/2012 18:51, Rick Flower ha scritto:
#detect: is not doing what you want.  It returns the first element for
which the block returns true.  Note that your block does not return a
boolean in the first place.

I think this is returning all the character-data inside the <UseCases>
element, discarding the elements.

Paolo

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Thanks Paolo.. My Smalltalk Collections memory is apparently dead..
I'll
go back and read-up on the collections again.. Been spending way too
much time with C++ lately and my smalltalk memory is fading!!

If I run into another snag I'll reply.. I did notice after posting
my code snippet that the XML was wrong -- the first Elements line
should look like :

<!ELEMENT requirements (UseCases+)>

and not

<!ELEMENT requirements (UseCase+)>


_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Ok..I changed the method shown earlier to the following
and know that it ought to put each element on its own
line :

buildUseCaseCollection2From: aList [
   aList do: [:each | Transcript show: 'Element:' , (each
characterData); cr ]
]

==========================
Element:
                 Use-Case #1
                 Use-Case #2

==========================

Notice the extra line wraps -- one at the beginning and
one at the end.  It seems like my collection is not being
properly setup but I followed the example in the docs
which says that if I do a :

parser document root elementsNamed: 'UseCases'

I ought to get an OrderedCollection with 2 items in it
(in this case).. But I seem to have only 1 item with
both combined..

Here's the updated XML being parsed :

<?xml version="1.0"?>
<!DOCTYPE requirements [
<!ELEMENT requirements (UseCases+)>
<!ELEMENT UseCase (#PCDATA)>
<!ELEMENT UseCases (UseCase+)>
]>
<requirements>
        <UseCases>
                <UseCase>Use-Case #1</UseCase>
                <UseCase>Use-Case #2</UseCase>
        </UseCases>
</requirements>

Any ideas on what I'm not understanding?  Thanks!!

--Rick

_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Got it to parse properly now.. Had two issues :

1) hard ^M (carriage returns) at the end of each XML line -- thanks
    Eclipse! -- didn't cause any problems just annoying

2) Got my array of use-case elements using the following code:

(parser document root elementNamed: 'UseCases') elements reject: [:each
| each isBlankText]

For whatever reason the array for the returned element has
empty entries in it -- and looks like the following in my
case :

An instance of Array
   contents: [
     [1]:

     [2]: <UseCase>Use-Case #1</UseCase>
     [3]:

     [4]: <UseCase>Use-Case #2</UseCase>
     [5]:

   ]

After running the above reject: code it becomes :
An instance of Array
   contents: [
     [1]: <UseCase>Use-Case #1</UseCase>
     [2]: <UseCase>Use-Case #2</UseCase>
   ]

YMMV!


_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

Il 25/09/2012 20:29, Rick Flower ha scritto:
They are not empty, they correspond to the whitespace between <UseCases>
and <UseCase>.

If you want to ignore it, you need to create a subclass of DOM_SAXDriver
and override the #ignorableWhitespace: method.  The overridden method
can simply do nothing.  Then to use your driver call this before
starting the parse:

    parser saxDriver: NoWhitespaceSAXDriver new

Paolo


_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk

On 26.09.2012 00:08, Paolo Bonzini wrote:

Cool! Thanks Paolo!!



_______________________________________________
help-smalltalk mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-smalltalk