Of course you realize that this can be broken by perfectly valid XML:

<?xml version="1.0"?>
<doc>
<!-- <?xml version="1.0"?>-->        <!-- an XML comment -->
<![CDATA[<?xml version="1.0"?> ]]>   <!-- a CDATA section -->
</doc>
[download]

I can't really see any "good" XML way to do this though. Why on Earth doesn't your customer use tar or zip or whatever instead of going a definitely non-XML (and risky) way?

Note that this particular example will work with the program, because the regex is anchored at the start of the line. However, in general it is possible that a "fake" XML declaration will appear at the beginning of a line, so it is a good point. For example:

<![CDATA[
<?xml version="1.0"?> ]]>
[download]

Also, while playing with the little program I wrote, I used the cat command to concatenate a bunch of XML files I have, and discovered that some of them did not have an end-of-line at the end of the last line, so the result would be like this:

</ESP-Component><?xml version="1.0"?>
[download]

Where the start declaration of one file would end up on the same line as the closing declaration of the previous file. In this case, this simple mechanism for recognizing when a file starts would not work, obviously.

Of course, I agree that some other packing method would be much better. But people do weird things... :-)

--ZZamboni

The poor man's "tar":

binmode(STDOUT);
for my $file ( @ARGV ) {
    open(FILE,"<$file") or die "Can't read $file: $!\n";
    binmode(FILE);
    print "$file: ",-s $file,$/;
    while(<FILE>) {
        print;
    }
}
[download]

binmode(STDIN);
my $file;
while(  defined( $file= <STDIN> )  ) {
    chomp $file;
    $file =~ s/: (\d+)$//  or  die "Invalid header: $file\n";
    my $size= $1;
    open(FILE,">$file")  or  die "Can't create $file: $!\n";
    warn "Extracting $size bytes to $file...\n";
    my $len= $size < 4096 ? $size : 4096;
    my $read;
    while(  $read= sysread(STDIN,$rec,$len)  ) {
        $size -= $read;
        $len= $size   if  $size < $len;
        print FILE $rec;
    }
}
[download]

use strict;
my $n="0000";
while (<>) {
  if (/^<\?xml version/) {
    open OF, ">outfile_$n.xml" or die "open: $!\n";
    $n++;
  }
  print OF $_;
}
[download]

--ZZamboni

I think you have a better chance of a useful answer if you specify the problem you want to solve first.

-- Abigail

Abigail-sama, I am righteously admonished. Please allow me to correct myself...

(BTW, good to see you posting here again.)

I suppose I was more concentrating on code, rather than phrasing my 'question'. I was looking more for a conversation about coding style & technique so that I could learn a better way to do something. I'm sure someone of much brain such as yourself could do this in a few lines. Perhaps I should rephrase it in a more generic, theoretical way, rather than a real world example?

I'm attempting to simply split a file into smaller files based on a search criteria. I've explained the real world application. But my approach is simply one of finding the opening XML Processing Instruction which identifies the version of the XML file. One subtopic of discussion might be why I shouldn't do that, perhaps.. For now I'm splitting up a file into smaller pieces based on a regular expression. I'm asking is there is a simpler way to do this, or if my coding style is perhaps obtuse? Comments of all temperatures are welcome. (I'm wearing my asbestos weave robes today. (; )

Here some sample data:

<?xml version = "1.0"?>
<!DOCTYPE AdvanceShippingNotice SYSTEM "AdvanceShippingNotice.dtd">
<AdvanceShippingNotice>
   <AdvanceShipping>
      <AdvanceShippingDetail>
      </AdvanceShippingDetail>
   </AdvanceShipping>
</AdvanceShippingNotice>
<?xml version = "1.0"?>
<!DOCTYPE AdvanceShippingNotice SYSTEM "AdvanceShippingNotice.dtd">
<AdvanceShippingNotice>
   <AdvanceShipping>
      <AdvanceShippingDetail>
      </AdvanceShippingDetail>
      <AdvanceShippingDetail>
      </AdvanceShippingDetail>
      <AdvanceShippingDetail>
      </AdvanceShippingDetail>
   </AdvanceShipping>
   <GlobalDocumentFunctionCode>ASN</GlobalDocumentFunctionCode>
</AdvanceShippingNotice>
<?xml version = "1.0"?>
<!DOCTYPE AdvanceShippingNotice SYSTEM "AdvanceShippingNotice.dtd">
<AdvanceShippingNotice>
   <AdvanceShipping>
      <AdvanceShippingDetail>
      </AdvanceShippingDetail>
   </AdvanceShipping>
   <fromRole/>
   <toRole/>
   <thisDocumentGenerationDateTime>
      <DateTimeStamp>20010619120949</DateTimeStamp>
   </thisDocumentGenerationDateTime>
   <thisDocumentIdentifier>
      <ProprietaryDocumentIdentifier/>
   </thisDocumentIdentifier>
   <GlobalDocumentFunctionCode>ASN</GlobalDocumentFunctionCode>
</AdvanceShippingNotice>
[download]

I have left the details of the data out for the sake of brevity, and since they don't matter.

The results of this should be three new files written to the filesystem which are themselves valid XML files.

Actually, let me correct that... Known bugs:

No zero padding of numbers in filenames to support alphabetical ordering

No actual XML validation



Wait! This isn't a Parachute, this is a Backpack!