Need Inifile-Parser with duplicated section handling

Here is snippet of the upi-file. The problems are the duplicate sections which causes the TIniFIle and nearly all other parsers to give up or return only the first section :-/

[DELETE FOLDER]
Path=%UserProfile%\.ARIS62
[CREATE FOLDER]
Path=%UserProfile%\.ARIS62
[COPY]
SOURCE=C:\_integra\upm\copy\user.cfg
DESTINATION=%UserProfile%\.ARIS62
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\ABC
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\Analyse
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\Applications
 
[CREATE REGISTRY KEY]
CLASS=
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\Applications2
 
[CREATE REGISTRY VALUE]
CLASS=REG_SZ
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\Applications2
DATA=ARIS.ANIApp.6.2
NAME=AniApp
 
[CREATE REGISTRY VALUE]
CLASS=REG_SZ
PATH=HKEY_CURRENT_USER\Software\IDS\Aris\6.2\Applications2
DATA=ARIS.CHARTApp.6.2
NAME=ChartApp
                                              

1:
2:
3:
4:
5:
6:
7:
8:
9:
10:
11:
12:
13:
14:
15:
16:
17:
18:
19:
20:
21:
22:
23:
24:
25:
26:
27:
28:
29:
30:
31:
32:
33:
34:
35:
36:
37:
38:
39:
40:
41:
42:
43:
44:
45:
46:
47:

This looks like you may have a golden opportunity to learn how to do data parsing. ;-)  

In other words, as far as I know, there isn't anything avaialbe that isn't goind to choke on that; however, it shouldn't be too hard to create your own parser that would extract the data.  Kindof depends on what you want to do after it is extracted.

It looks like this is a script of some sort for an installation process.  In that case, you should be able to parse it and execute the steps as you do so.  Although, I would probably build up a batch file from the data so that you could parse it all and check for problems (if you have sufficient information to do that ;-).

@TDS

The reason you are having a problem using the TInifile class is that this file is not a 'well formed' ini file.  I've used one of the Windows APIs to enumerate the sections, duplicated as you've already noted, and can only get the contents of the first section with a duplicate name.

You might want to explore regular expression parsing of this file.  The actions are located within the square brackets and are followed by one or more lines of name=value pairs.

Of course, you could make a copy of the file, manually eedit it so that the sections are no longer duplicated (for instance, just a a sequencing number in each of the squeare bracketted names), and then work with it.

It kind of depends on what you plan to do with it.  If you just need to be able to read it in order to use the data to drive your processing of the file, then I would look at manually editing it.  If, on the other hand, you are needing to use it as it was originall intended, then you have a problem.

I've found a unit called FastIniFiles (v3.2). It handles sections with linked lists so I can walk through it.
Will post here if i'll have luck :-)

I think the easiest solution would be the following:
1. read the entire contents into a TStringlist variable
2. iterate through the stringlist items (lines)
3. at each item, check for an equals sign (=) character
if found
 parse the item into another stringlist variable using the '=' as the delimiter
else
 use the second-through-next-to-last character, eliminating the square brackets from the action name

================================
If you are trying to mimic the Symantec actions, you need to perform these actions in the order in which they appear.  Linked list processing might not preserve that order.

If the name/value lines contain a space character, the method I described above will not work.  The DelimitedText always includes spaces in addition to the Delimiter character you assign. bummer :-(

Here is an example of parsing the file and putting the results into a string grid.  It correctly parses name/value lines containing a space.

procedure TSDIAppForm.Button1Click(Sender: TObject);
  var
    ts:tstringlist;
    itemloop:integer;
    delimposn:integer;
    actioncmd:string;
 
begin
    ts:=tstringlist.Create;
    ts.loadfromfile('C:\Documents and Settings\aikimark\My Documents\malformedini.txt');
    for itemloop := 0 to ts.Count - 1 do
      begin
        delimposn:=pos('=',ts[itemloop]);
        if delimposn =0 then
          actioncmd:=copy(ts[itemloop],2,length(ts[itemloop])-2)
        else
            begin
              stringgrid1.Cells[0,stringgrid1.RowCount-1]:=actioncmd;
              stringgrid1.Cells[1,stringgrid1.RowCount-1]:=copy(ts[itemloop],1,delimposn-1);
              stringgrid1.Cells[2,stringgrid1.RowCount-1]:=copy(ts[itemloop],delimposn+1,length(ts[itemloop]));
              stringgrid1.RowCount:=stringgrid1.RowCount+1;
            end;
 
      end;
    stringgrid1.RowCount:=stringgrid1.RowCount-1;
end;
                                              

1:
2:
3:
4:
5:
6:
7:
8:
9:
10:
11:
12:
13:
14:
15:
16:
17:
18:
19:
20:
21:
22:
23:
24:
25:
26:

1. I wasn't suggesting you put this in a string grid.  I used a string grid on my PC to show me the parsed data on a user form.  You only need to pay attention to the acutal line parsing code, not the displaying code.

2. I didn't know your data was unicode.  That wasn't in the example you posted.

3. The lastest version of Delphi handles unicode.

I know that the latest version handles unicode but it don't handle duplicate sections. To avoid this and not having wasted time i switched back to BDS 2006 and the working fast ini files unit.

@TDS

What does BDS 2006 provide that D2009 doesn't provide?

Did you solve this with FastIniFiles?

BDS handles strings as normal ASCII. But D2009 uses Unicode (UTF-8) so one char are two bytes. In the FastIniFiles code is some assembler to rewrite. And that thing i couldn't do. Maybe the author of this unit will do it later.