OpenEuphoria

1. foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 21, 2006
• 742 views

Hi...

Where are we at on this issue? I searched/read the forum archives on this issue
last night. Any conclusions / plans? I ask because I can't believe that it
wasn't included (and some others) right from the onset. In the "Euphoria
Reference Manual" Part I-Section 2, it states:

"Euphoria data structures are almost infinitely flexible." Further down, "Not
only can a Euphoria program easily represent all conventional data structures
but you can create very useful, flexible structures that would be extremely
hard to declare in a conventional language." That may be true, but what good
is all that if you can't get at / manipulate them easily? IMHO, sequences are
*so huge* in Euphoria that we should scrap some of the functions that are
supported but perhaps rarely used, in favor of beefing up "sequence support".

In the meantime, anybody ever come across a "foreach" C function that could
be called from Euphoria? Later...
--
duke

3. Re: foreach routine

• Posted by Derek Parnell <ddparnell at bigpond.com> Dec 21, 2006
• 691 views
• Last edited Dec 22, 2006

duke normandin wrote:
> 
> Hi...
> 
> Where are we at on this issue? 

Here is a first cut at an interim solution...

-- foreach.e

-- Extended version. Covers the full functionality.

--/func every_item(pTheList, pRID, pData, pStart, pEnd, pSpan)
--/desc Process each specified item in the list
--/ret SEQUENCE: If no errors, a 4-element sequence;
--               [1] = 0 (zero)
--               [2] = 0 (if pRID sees each item)
--                   = integer (whatever pRID returned to end the processing)
--               [3] = The index of the last item processed.
--               [4] = The (possibly updated) list.
--               If errors, a 2-element sequence;
--               [1] = Error code:
--                      1 => The supplied low index is < 1
--                      2 => The supplied low index is > list length
--                      3 => The supplied high index is < 1
--                      4 => The supplied high index is > list length
--                      5 => The supplied routine id is invalid
--               [2] = The parameter in error.
--
-- This function calls the routine id /i(pRID) for each of the items
-- requested in the list /i(pTheList). The range and direction is
-- specified by /i(pStart) and /i(pEnd). /i(pStart) is the index
-- of the first item to be processed and /i(pEnd) is the index
-- of the last item to be processed. If /i(pStart) is higher than
-- /i(pEnd) then the items are processed in right-to-left order
-- otherwise they are processed in left-to-right order. /n
--
-- The /i(pSpan) parameter specifies the 'gap' between successive
-- items to be processed. To process each item use 1 as the span, to
-- process every second item use 2, etc... /n
--
-- The /i(pRID) parameter is a routine id of a caller defined function
-- which is called for each item processed. It is passed three parameters:
--     [1] => The index of the item being processed.
--     [2] => The value of the item being processed.
--     [3] => The /i(pData) parameter. This can be any object.
-- The routine must return either a single integer or a one-element
-- sequence. If it returns a 0 (zero) then next item is processed. If
-- it returns a non-zero integer, the processing of items is stopped.
-- If it returns a sequence, the first element is used to replace the
-- item just processed in the list.
--
-- Example:
--/code
-- sequence res
-- function showit(integer pIdx, object pItem, object pExtra)
--     printf(1, "%d: %s\n", {pIdx, pItem})
--     if pItem = pExtra then return 1 end if
--     return {pItem - ('A' - 'a')}
-- end function
--
-- res = every_item("ABCDEF", routine_id("showit"), 'D', 2, 5, 1)
-- printf(1, "%d %d %d %s\n", res)
--/endcode

----------------------------------------------------------
global function every_item(sequence pTheList, integer pRID, object pData,
                           integer pStart, integer pEnd, integer pSpan)
----------------------------------------------------------

    integer lDir
    object lRes
    integer lLastI
    integer lLow
    integer lHigh

    if length(pTheList) = 0 then
        return {0,0,0, pTheList}
    end if

    if pStart > pEnd then
        lDir  = -1
        lLow  = pEnd
        lHigh = pStart
    else
        lDir  = 1
        lHigh = pEnd
        lLow  = pStart
    end if

    if lLow < 1 then
        return {1, lLow}
    end if

    if lLow > length(pTheList) then
        return {2, lLow}
    end if

    if lHigh < 1 then
        return {3, lHigh}
    end if

    if lHigh > length(pTheList) then
        return {4, lHigh}
    end if

    if pRID < 0 then
        return {5, pRID}
    end if

    lRes = 0
    for i = pStart to pEnd by lDir * pSpan do
        lLastI = i
        lRes = call_func(pRID, {i, pTheList[i], pData})
        if integer(lRes) then
            if lRes != 0 then
                exit
            end if
        else
            pTheList[i] = lRes[1]
        end if
    end for

    return {0, lRes, lLastI, pTheList}

end function

--/func foreach(pTheList, pRID)
--/desc Process every item in the list, from left to right.
--/ret SEQUENCE: If no errors, a 4-element sequence;
--               [1] = 0 (zero)
--               [2] = 0 (if pRID sees each item)
--                   = integer (whatever pRID returned to end the processing)
--               [3] = The index of the last item processed.
--               [4] = The (possibly updated) list.
--               If errors, a 2-element sequence;
--               [1] = 5 (The supplied routine id is invalid)
--               [2] = The pRID value.
--
-- This function calls the routine id /i(pRID) for each of the items
-- in the list /i(pTheList). They are processed in left-to-right order. /n
--
-- The /i(pRID) parameter is a routine id of a caller defined function
-- which is called for each item processed. It is passed three parameters:
--     [1] => The index of the item being processed.
--     [2] => The value of the item being processed.
--     [3] => 0
-- The routine must return either a single integer or a one-element
-- sequence. If it returns a 0 (zero) then next item is processed. If
-- it returns a non-zero integer, the processing of items is stopped.
-- If it returns a sequence, the first element is used to replace the
-- item just processed in the list.
--
-- Example:
--/code
-- sequence res
-- function showit(integer pIdx, object pItem, object pExtra)
--     printf(1, "%d: %s\n", {pIdx, pItem})
--     if pItem = 'D' then return 1 end if
--     return {pItem - ('A' - 'a')}
-- end function
--
-- res = foreach("ABCDEF", routine_id("showit"))
-- printf(1, "%d %d %d %s\n", res)
--/endcode

----------------------------------------------------------
global function foreach(sequence pTheList, integer pRID)
----------------------------------------------------------

    return every_item(pTheList, pRID, 0, 1, length(pTheList), 1)

end function

--/func foreach_reverse(pTheList, pRID)
--/desc Process every item in the list, from right to left.
--/ret SEQUENCE: If no errors, a 4-element sequence;
--               [1] = 0 (zero)
--               [2] = 0 (if pRID sees each item)
--                   = integer (whatever pRID returned to end the processing)
--               [3] = The index of the last item processed.
--               [4] = The (possibly updated) list.
--               If errors, a 2-element sequence;
--               [1] = 5 (The supplied routine id is invalid)
--               [2] = The pRID value.
--
-- This function calls the routine id /i(pRID) for each of the items
-- in the list /i(pTheList). They are processed in right-to-left order. /n
--
-- The /i(pRID) parameter is a routine id of a caller defined function
-- which is called for each item processed. It is passed three parameters:
--     [1] => The index of the item being processed.
--     [2] => The value of the item being processed.
--     [3] => 0
-- The routine must return either a single integer or a one-element
-- sequence. If it returns a 0 (zero) then next item is processed. If
-- it returns a non-zero integer, the processing of items is stopped.
-- If it returns a sequence, the first element is used to replace the
-- item just processed in the list.
--
-- Example:
--/code
-- sequence res
-- function showit(integer pIdx, object pItem, object pExtra)
--     printf(1, "%d: %s\n", {pIdx, pItem})
--     if pItem = 'D' then return 1 end if
--     return {pItem - ('A' - 'a')}
-- end function
--
-- res = foreach_reverse("ABCDEF", routine_id("showit"))
-- printf(1, "%d %d %d %s\n", res)
--/endcode

----------------------------------------------------------
global function foreach_reverse(sequence pTheList, integer pRID)
----------------------------------------------------------

    return every_item(pTheList, pRID, 0, length(pTheList), 1, 1)

end function

-- 
Derek Parnell
Melbourne, Australia
Skype name: derek.j.parnell

5. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 22, 2006
• 681 views

Derek Parnell wrote:
> 
> duke normandin wrote:
> > 
> > Hi...
> > 
> > Where are we at on this issue? 
> 
> Here is a first cut at an interim solution...
> }}}
<eucode>
> -- foreach.e

[snip and saved] 

Nice Code! and well documented! The Euphoria community may just put you in
the Hall-of-Fame you know ;)

Like I said in another post, I basically trying to emulate Perl's `foreach'
routine, as in:
     foreach $pair (@pairs){
        do something
     }

where @pairs is a simple array (no sweat! => a "U4ia" sequence)
      $pair is a variable used in "list" context so it gets a string

So in Perl, what happens here is that the foreach routine is fed an array
(sequence), it walks through it and returns each element one at a time.
This element gets shoved into $pair (in this example). $pair is then
massaged whichever way you need.

In Lua, they have `foreach' to walk through a hash, and `foreachi' to walk
through an "indexed" array. Slick too!
--
duke

7. Re: foreach routine

• Posted by ags <eu at 531pi.co.nz> Dec 22, 2006
• 688 views

duke normandin wrote:
> 
> Coming from Perl, I'm trying to emulate the following:
> 
> read(STDIN, $buffer, $ENV{CONTENT_LENGTH});
> @pairs = split(/&/, $buffer);
> foreach $pair (@pairs){
>     ($name, $value) = split(/=/, $pair);
>      blah
>      blah
> }

BTW you will probably miss tightly integrated regular expression too.  I've been
meaning to wrap the perlre.dll for a while now but it's a bit of a headache
everytime I look at the API :)

That and the fact that you can't just drop in a RE in things like split, unpack,
etc.

But I use Perl and Euphoria equally as much... Perl for CGI and general
scripting for server stuff and Euphoria for Windows programming and general
scripting on Windows.

Gary

9. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 22, 2006
• 673 views

ags wrote:
> 
> duke normandin wrote:
> > 
> > Coming from Perl, I'm trying to emulate the following:
> > 
> > read(STDIN, $buffer, $ENV{CONTENT_LENGTH});
> > @pairs = split(/&/, $buffer);
> > foreach $pair (@pairs){
> >     ($name, $value) = split(/=/, $pair);
> >      blah
> >      blah
> > }
> 
> BTW you will probably miss tightly integrated regular expression too.  I've
> been meaning to wrap the perlre.dll for a while now but it's a bit of a
> headache
> everytime I look at the API :)
> 
> That and the fact that you can't just drop in a RE in things like split,
> unpack,
> etc.
> 
> But I use Perl and Euphoria equally as much... Perl for CGI and general
> scripting
> for server stuff and Euphoria for Windows programming and general scripting
> on Windows.

My gut feeling tells me that U4ia can do CGI and various backend stuff OK.
The issue for me is how efficiently? So far, lots of hoops to jump through
to get simple things done. However I'm serious about U4ia. I'm still reading the
reference manual. In section 2.3 it states:

"Programming in Euphoria is based entirely on creating and manipulating
flexible, dynamic sequences. Sequences are IT - there are no other
structures to learn"

Nice --- but hardly no tools to do the above "manipulating" of sequences.
Its sad!
--
duke

11. Re: foreach routine

• Posted by Derek Parnell <ddparnell at bigpond.com> Dec 22, 2006
• 697 views

duke normandin wrote:
> 
> Coming from Perl, I'm trying to emulate the following:
> 
> read(STDIN, $buffer, $ENV{CONTENT_LENGTH});
> @pairs = split(/&/, $buffer);
> foreach $pair (@pairs){
>     ($name, $value) = split(/=/, $pair);
>      blah
>      blah
> }

These might help ...

-- split.e

--/func split_char(pDelim, pText)
--/desc This will create subsequences of pText which is delimited by pDelim
--/ret SEQUENCE: A set of subsequences.
--This looks for the delimiter in pText and breaks it up into subsequences
--at the delimiter boundaries.
--
-- Example:
--/code
--  res = split_char('=', "age=59")
--  -- Returns {"age", "59"}
--  res = split_char(';', "abc;defg;hi;jkl;;qwerty")
--  -- Returns {"abc","defg","hi","jkl","","qwerty"}
--/endcode

global function split_char(integer pDelim, sequence pText)
    sequence lResult
    integer lPos
    integer lNewPos

    lResult = {}
    lNewPos = 1
    lPos = find(pDelim, pText)
    while lPos > 0 do
        lResult = append(lResult, pText[lNewPos..lPos-1])
        lNewPos = find(pDelim, pText[lPos+1..$])
        if lNewPos > 0 then
            lPos += lNewPos
            lNewPos = lPos - lNewPos + 1
        else
            lNewPos = lPos+1
            lPos = 0
        end if
    end while

    lResult = append(lResult, pText[lNewPos..$])
    return lResult
end function

--/func split(pDelim, pText)
--/desc This will create nested subsequences of pText which is delimited by
pDelim
--/ret SEQUENCE: A set of nested subsequences.
--/i(pDelim) is a set of one-character delimiters. /i(pText) is recursively
-- split based on successive delimiters in /i(pDelim).
--
-- Example:
--/code
--  res = split("&=", "dude=duke&age=59&style=kewl")
--  -- Returns {{"dude","duke"}, {"age", "59"}, {"style", "kewl"}}
--/endcode
--
--However, if /i(pDelim) is either an integer or a single character sequence
-- the result is equivalent to the split_char() function.
--
--Example:
--/code
--  res = split('=', "age=59")
--  -- Returns {"age", "59"}
--  res = split(";", "abc;defg;hi;jkl;;qwerty")
--  -- Returns {"abc","defg","hi","jkl","","qwerty"}
--/endcode

global function split(object pDelim, sequence pText)
    sequence lResult

    if integer(pDelim) then
        return split_char(pDelim, pText)
    end if

    lResult = split_char(pDelim[1], pText)
    if length(pDelim) > 1 then
        for i = 1 to length(lResult) do
            lResult[i] = split(pDelim[2..$], lResult[i])
        end for
    end if

    return lResult
end function

--/func toKeyValue(pPairs)
--/desc Converts the set of paired subsequences into two sequences of keys and
values.
--/ret SEQUENCE: A 2-element sequence. The first contains a list of keys, the
2nd a list of corresponding values.
--The input is assumed to be a list of 2-element (pairs) sequences. The
--output is just the pairs rearranged such that the first output subsequence
-- is a list of all the keys, and the second output subsequences is
-- a list of the corresponding values for those keys.
--
--Example:
--/code
--  res = toKeyValue( {{"dude","duke"}, {"age", "59"}, {"style", "kewl"}} )
--  -- Returns ...
--  --   { {"dude", "age", "style"}, {"duke", "59", "kewl"} }
--/endcode

global function toKeyValue(sequence pPairs)
    sequence lResult

    lResult = repeat( repeat({}, length(pPairs)), 2)

    for i = 1 to length(pPairs) do
        lResult[1][i] = pPairs[i][1]
        lResult[2][i] = pPairs[i][2]
    end for

    return lResult
end function

--/func findKeyValue(pKey, pKeyValues, pDefault)
--/desc Returns the value associated with the /i(pKey) parameter.
--/ret OBJECT: The associated value of the key.
--The /i(pKeyValues) parameter must be a 2-element sequence, the
--first containing keys and the second containing values for those keys.
--
--This looks for the /i(pKey) value in the keys subsequence and when
--found returns the corresponding value. If not found, it returns the
--supplied /i(pDefault) value.
--
--Example:
--/code
--  sequence x
--  x = toKeyValue( split("&=", "dude=duke&age=59&style=kewl") )
--  ? findKeyValue("age", x, -1)    -- returns "59"
--  ? findKeyValue("dude", x, -1)   -- returns "duke"
--  ? findKeyValue("rabbit", x, -1) -- returns -1
--  ? findKeyValue("style", x, -1)  -- returns "kewl"
--/endcode

global function findKeyValue(object pKey, sequence pKeyValues, object pDefault)
    integer lPos

    lPos = find(pKey, pKeyValues[1])
    if lPos > 0 then
        return pKeyValues[2][lPos]
    else
        return pDefault
    end if

end function

-- 
Derek Parnell
Melbourne, Australia
Skype name: derek.j.parnell

13. Re: foreach routine

• Posted by Jules Davy <jdavy at dsl.pipex.com> Dec 22, 2006
• 688 views

> Nice --- but hardly no tools to do the above "manipulating" of sequences.
> Its sad!
> --
> duke

take a look at Ricardo Forno's general functions library, I find it very useful
for sequence splicing and dicing.

15. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 22, 2006
• 690 views

Derek Parnell wrote:
> 
> duke normandin wrote:
> > 
> > Coming from Perl, I'm trying to emulate the following:
> > 
> > read(STDIN, $buffer, $ENV{CONTENT_LENGTH});
> > @pairs = split(/&/, $buffer);
> > foreach $pair (@pairs){
> >     ($name, $value) = split(/=/, $pair);
> >      blah
> >      blah
> > }
> 
> These might help ...
> }}}
<eucode>
> -- split.e

[code snipped]

What can I say but *awesome* !! -- and -- thanks a bunch. I've got enough
code to study now so that I'll either "Get it!" or I should imigrate
"down-under" and chase "roos" for a living. ;)  Later...
--
duke

17. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 22, 2006
• 679 views

Jules Davy wrote:
> 
> > Nice --- but hardly no tools to do the above "manipulating" of sequences.
> > Its sad!
> > --
> > duke
> 
> take a look at Ricardo Forno's general functions library, I find it very
> useful
> for sequence splicing and dicing.

Thanks! I'll have a look! 
--
duke

19. Re: foreach routine

• Posted by Chris Bensler <bensler at nt.net> Dec 22, 2006
• 699 views

duke normandin wrote:
> 
> What I was going to say is:
> 
> CONTENT_LENGTH will either be empty or contain the number of chars/bytes
> sent. So 
> if atom(tmp) then
> 
> checks to see if getenv() returned a number, right? If true, we have that
> many chars/bytes. If 0 CONTENT_LENGTH was empty.
> 
> Do I need another cup of coffee? or something? It seems that the logic
> is backwards. TIA...  (now hit the send button, duke)
> --
> duke

The getenv() routine will return -1 if the environment variable CONTENT_LENGTH
doesn't exist. If it does exist, a string will be returned.
Since CONTENT_LENGTH is supposed to be a number, we use value() to convert the
string to a number.

value() also returns a sequence of length = 2, where the 1st element is an error
code and the 2nd is the actual number value.

Hence the 2 error checks.

HTH

Chris Bensler
~ The difference between ordinary and extraordinary is that little extra ~
http://empire.iwireweb.com - Empire for Euphoria

21. Re: foreach routine

• Posted by cklester <cklester at yahoo.com> Dec 22, 2006
• 684 views

duke normandin wrote:
> My gut feeling tells me that U4ia can do CGI and various backend stuff OK.
> The issue for me is how efficiently? So far, lots of hoops to jump through
> to get simple things done.

I've not really had to jump through any "hoops." Using some libraries freely
available in the archive, plus building a little bit of my own, I've got
several web sites being served by Euphoria.

I've built a content management system called Building Blocks CMS. It has
made development of dynamic web sites pretty easy. I'm even incorporating
Ajax functionality now. Cool stuff!

-=ck
"Programming in a state of Euphoria."
http://www.cklester.com/euphoria/

23. Re: foreach routine

• Posted by Derek Parnell <ddparnell at bigpond.com> Dec 22, 2006
• 699 views
• Last edited Dec 23, 2006

duke normandin wrote:
> 
> Chris Bensler wrote:
> > The getenv() routine will return -1 if the environment variable
> > CONTENT_LENGTH
> > doesn't exist. If it does exist, a string will be returned.
> 
> Are you sure? I've never proven this to myself, but it has always been my
> understanding that with respect the method=POST, the ENV. VAR. CONTENT_LENGTH
> would be set to the amount of bytes sent out. THEREFORE, this is why we "read"
> that many bytes from STDIN, and only that many. So my point is that
> CONTENT_LENGTH
> will in one case be a positive number or a -1 -- both atoms. No?

Yes and no. The Euphoria function getenv() always returns a string if the
environment variable exists, and -1 if it doesn't exist. In your case,
CONTENT_LENGTH will most likely always be there due to the CGI environment, but
when getenv() returns its content, it is a string representation of a number and
not the number itself. So if the number of characters to read in from STDIN is
59, getenv("CONTENT_LENGTH") will return "59" and not 59. Thus you have to
convert the string "59" to its number form.

  res = value( "59" )

The value() function always returns a 2-element sequence. The first element is
the success/failure code and the second is the converted number. So in this case,
value() returns {0,59} where the leading 0 indicates successful conversion
occured. So in code ...

object result
result = getenv("CONTENT_LENGTH")
if sequence(result) then 
   -- a string was returned.
   result = value(result)
   if result[1] = 0 then
       --- a good conversion.
       readsize = result[2]
   else
       --- bad data in string!
   end if
else
   --- CONTENT_LENGTH doesn't exist
end if

 
> The above of course does not apply to method=GET, because in the latter, all
> the data is stuffed in the ENV. VAR. QUERY_STRING. Here we can check to see
> if getenv("QUERY_STRING") returns an atom i.e -1, because QUERY_STRING will
> either contain a string, or will be undefined. 
> --
> duke


-- 
Derek Parnell
Melbourne, Australia
Skype name: derek.j.parnell

25. Re: foreach routine

• Posted by Chris Bensler <bensler at nt.net> Dec 23, 2006
• 691 views

duke normandin wrote:
> 
> Derek Parnell wrote:
> > 
> > duke normandin wrote:
> > > 
> > > Chris Bensler wrote:
> > > > The getenv() routine will return -1 if the environment variable
> > > > CONTENT_LENGTH
> > > > doesn't exist. If it does exist, a string will be returned.
> > > 
> > > Are you sure? I've never proven this to myself, but it has always been my
> > > understanding that with respect the method=POST, the ENV. VAR.
> > > CONTENT_LENGTH
> > > would be set to the amount of bytes sent out. THEREFORE, this is why we
> > > "read"
> > > that many bytes from STDIN, and only that many. So my point is that
> > > CONTENT_LENGTH
> > > will in one case be a positive number or a -1 -- both atoms. No?
> > 
> > Yes and no. The Euphoria function getenv() always returns a string if the
> > environment
> > variable exists, and -1 if it doesn't exist. In your case, CONTENT_LENGTH
> > will
> > most likely always be there due to the CGI environment, but when getenv()
> > returns
> > its content, it is a string representation of a number and not the number
> > itself.
> > So if the number of characters to read in from STDIN is 59,
> > getenv("CONTENT_LENGTH")
> > will return "59" and not 59. Thus you have to convert the string "59" to its
> > number form. 
> 
> I got it! Damn! I've got to watch the data types more carefully. Not used to
> it. ;)
> To recap, getenv returns the atom -1 if the ENV.VAR. is unset OR a string
> containing
> whatever. In the case of CONTENT_LENGTH I was expecting an integer. I now know
> to
> convert CONTENT_LENGTH *to* an integer.
> 
> >   res = value( "59" )
> 
> [snip]
> 
> I really don't need to do all that testing.
> 
> What the Perl/PHP etc world has done all this time is:
> 
> 1. getenv("REQUEST_METHOD")
> 2. if == "POST"
>    2.a value_array = value(getenv("CONTENT_LENGTH")
>    2.b stash get_bytes(0, value_array[2])
> 3. if == "GET"
>    3.a stash getenv("QUERY_STRING")
> 4. if == "some_bogus_method"
>    4.a cgi_die("gracefully")
> 5. amen
> 
> That is what I want to reproduce in Euphoria come "hell or high water" ;)
> 
> Once I have the above "stash", then we use your "foreach" and "split" code to
> massage the FORM[key]=values
> I love it !! Thanks you all for the mini EU tutorials!!
> --
> duke

Good programming practice is to always trap error conditions whenever you can.

I get the impression though, that you meant that you don't want to have to do
the error testing all the time. Make a routine that takes care of it for you
instead.

I use this routine..
This avoids the error by making it default to an empty string.
This routine will always return a valid string or "".

global function cgi_getenv(string name)
 object tmp
  tmp = getenv(name)
  if atom(tmp) then return "" end if
  return tmp
end function

Similarly for value(), I have..
This avoids the error by making it default to 0.
This always returns the converted number, or 0.

global function get_value(object x)
 sequence ret
  ret = value(x)
  if ret[1] = GET_SUCCESS then
    return ret[2]
  else
    return 0
  end if
end function

The above routines are not generally a good way to deal with error cases, but
for specific purposes such as this situation, it should be fine.

Using those two routines, I could make a new routine specifically to return the
content_length..
This routine returns the number value of the content length.

global function cgi_content_length()
 integer len
  len = cgi_getenv("CONTENT_LENGTH")
  return get_value(len)
end function

And a routine to handle the request method...

global function cgi_request_method()
  return cgi_getenv("REQUEST_METHOD")
end function

For further processing of the POST or QUERY, you will need some code like the
following, to convert url-encoded strings to plain text.

function replace_hex(sequence s)
 sequence f,hex
  -- find and store the index of all instances of '%' in s
  f = {}
  for i = 1 to length(s) do
    if s[i] = '%' then
      f &= i
    end if
  end for
  if length(f) then
    for i = length(f) to 1 by -1 do
hex = get_value("#"&s[f[i]+1..f[i]+2]) -- convert the string code to a
      character code
      s = s[1..f[i]-1] & hex & s[f[i]+3..length(s)]
    end for
  end if
  return s
end function

------------------------------------------------
-- This is the function that you are mostly after
-- cgi_parse_request() will convert a url-encoded string to key/value pairs
-- in the form of {{key1,key2,...},{val1,val2,...}}
-- EG. "firstname=Duke&lastname=Normandin"
--  becomes {{"firstname","lastname"},{"Duke","Normandin"}}

global function cgi_parse_request(sequence s)  
  s ={ s , {} }
  s[1] = join(split(s[1],'+'),' ') -- convert '+' to ' '
  s[1] = join(split(s[1],'&'),';') -- convert '&' to ';'
  s[1] = split(s[1],';')

  if length(s[1]) then s[2] = repeat("",length(s[1])) end if
  for i = 1 to length(s[1]) do
    s[1][i] = split(s[1][i],'=')
    s[1][i][1] = replace_hex(s[1][i][1])
    if length(s[1][i]) > 1 then
      s[2][i] = replace_hex(s[1][i][2])
    end if
    s[1][i] = s[1][i][1]
  end for
  return s
end function

Here is a routine to handle fetching the post data.
It will return a sequence of key/values as described for cgi_parse_request()

global function cgi_get_post()
 sequence buf
 integer len
  len = cgi_content_length()
  buf = get_bytes(STDIN,len)
  return cgi_parse_request(buf)
end function

Here is the routine to handle QUERY data.

global function cgi_get_query()
 sequence buf
  buf = cgi_getenv("QUERY_STRING")
  return cgi_parse_request(buf)
end function

Now you can make a routine to actually do the whole thing for you..

global function cgi_get_request()
 sequence req
  req = cgi_request_method()
  if equal(req,"POST") then
    return cgi_get_post()
  elsif equal(req,"GET") then
    return cgi_get_query()
  else
    -- error, invalid request method
  end if
end function

Buene?


You may be wondering why the pairs are stored as {key_list,data_list} instead of
{{key1,data1},{key2,data2}}. The reason is because it's easier to search for a
key that way.

EG.
sequence pairs
integer f
pairs = cgi_get_request()
f = find("firstname",pairs[1])
if f then
 puts(STDOUT,"Your firstname is: " & pairs[2][f] & "\n")
end if


Chris Bensler
~ The difference between ordinary and extraordinary is that little extra ~
http://empire.iwireweb.com - Empire for Euphoria

27. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 23, 2006
• 668 views

Chris Bensler wrote:

> > [snip]
> > 
> > I really don't need to do all that testing.
> > 
> > What the Perl/PHP etc world has done all this time is:
> > 
> > 1. getenv("REQUEST_METHOD")
> > 2. if == "POST"
> >    2.a value_array = value(getenv("CONTENT_LENGTH")
> >    2.b stash get_bytes(0, value_array[2])
> > 3. if == "GET"
> >    3.a stash getenv("QUERY_STRING")
> > 4. if == "some_bogus_method"
> >    4.a cgi_die("gracefully")
> > 5. amen
> > 
> > That is what I want to reproduce in Euphoria come "hell or high water" ;)
> > 
> > Once I have the above "stash", then we use your "foreach" and "split" code
> > to
> > massage the FORM[key]=values
> > I love it !! Thanks you all for the mini EU tutorials!!
> > --
> > duke
> 
> Good programming practice is to always trap error conditions whenever you can.

I agree! But how often you need to trap errors depends on the language you are
using and the number of lines of code you need to generate to get something
done.
 
> I get the impression though, that you meant that you don't want to have to do
> the error testing all the time. Make a routine that takes care of it for you
> instead.

That's what I meant. As a EU novice, I'm quickly seeing that my coding style
and practices *need* to be different than those I was using with Perl and
PHP.

That said, I've been doing CGI / Web / SQL stuff since 1996 using Perl mostly. I
know how the CGI and back-end stuff works!! I just need to learn to do it
correctly "the EU way". ;)) Nice collection of functions BTW!! I'm sure they
get the job done. Can't wait to put it all together and try it out. Thanks
for the input!
--
duke

29. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 23, 2006
• 675 views

cklester wrote:
> 
> duke normandin wrote:
> > 
> > 
> > Once I have the above "stash", then we use your "foreach" and "split" code
> > to
> > massage the FORM[key]=values
> 
> Have you checked out the CGI code available in the archive? I just do this:
> 
> include cgi.e
> process_cgi()
> -- then
> x = get_key_value("whatever")
> 
> ...or something close to that. Easy! :)

I have searched the archives, but I don't think that I ran across cgi.e. I will
check it out. Thanks!
--
duke

31. Re: foreach routine

• Posted by duke normandin <dnormandin at bsdrocksperlrolls.com> Dec 23, 2006
• 688 views

Chris Bensler wrote:
> 

> > That said, I've been doing CGI / Web / SQL stuff since 1996 using Perl
> > mostly.
> > I
> > know how the CGI and back-end stuff works!! I just need to learn to do it
> > correctly "the EU way". ;)) Nice collection of functions BTW!! I'm sure they
> > get the job done. Can't wait to put it all together and try it out. Thanks
> > for the input!
> 
> Sorry, I didn't mean for it to sound preachy. I can tell that you know web
> programming
> and I look forward to the contributions I'm sure you will be making in no time
> :)

No worries! Really!!
 
> You are right in thinking that there appears to be alot of 'hoops' to jump
> through
> when using Euphoria. Don't let it turn you away though. Eu is an excellent
> language.
> Eu's greatest hurdle is to get more people to use it, so we have more people
> contributing quality code and banging heads.
>
> This thread is precisely why I started the thread about standardized euphoria
> and this issue is why I have been working on a design and implementation of
> an improved set of standard includes for quite a long time. Something like a
> CGI library should be a part of the standard includes (guess where I ripped
> the cgi functions I pasted for you? ;). People shouldn't have to construct a
> hammer before they can build a house and I hope I can help to resolve that for
> us. Anyways I'm getting off topic.

The above paragraphs are reason enough for me to adopt EU as my "native"
language. I like your attitude and style, and I have a sense that, for the
most part, you mirror the prevailing attitude(s) of this fine community. I'm
not a C programmer, so I can't be of much use with the guts of EU, but I
sure can't wait to get up to speed with the language itself in order to
start making a contribution. You know, I may be familiar with Perl and its
CPAN etc, but I left that community and more-and-more the language itself
for a reason or two....! But methinks I've found a home!

Anyway, thank you for your candor and insights. You and Derek Parnell and
others have been just great. Later....
--
duke