[Pdx-pm] hash from array question

Eric Wilhelm scratchcomputing at gmail.com
Fri May 25 16:39:00 PDT 2007

# from Thomas J Keller
# on Friday 25 May 2007 02:23 pm:

>line 32 print "genomes string: $genomes\n";
>line 33 print "split gives: ", split(/,\s*/,$genomes),"\n";
>line 34 my %genomes = split(/,\s*/,$genomes) ;
>line 35 print map { "Key: $_  has value: $genomes{$_}\n" } sort keys %

Yep.  What Andy said.

While $genomes and %genomes are distinct, it is usually helpful to avoid 
using the same word with different sigils within a single scope.  Had 
there not been a $genomes, the use of %$genomes would have triggered the 
error [Global symbol "$genomes" requires ...] rather than the slightly 
less obvious error [Can't use string ... as a hash reference].


  my %genomes = split(/,\s*/, $config{"$project.genomes"});

Anyway, now for the comic relief.  I'll admit to not having any clue what
I'm doing here, but I find the scary avoidance of exception handling rather
ironic when coupled with the fact that string comparison is done so elegantly
via "==" operator overloading.  Also note the "lets all invent our own
string libraries" fun coming from both ends.  And of course the 2-line 
grep/map statement that would be required to do this in a 10-line method in
Perl.  Whee!

bool wxMozillaBrowser::ScrollToElementByID(wxString id)
  //fprintf(stderr, ((wxT("id: ")+ id + wxT("\n")).mb_str()));
  nsCOMPtr<nsIDOMWindow> domWindow;
  nsresult rv;
  rv = m_Mozilla->mWebBrowser->GetContentDOMWindow(getter_AddRefs(domWindow));
  if (!domWindow)
    return FALSE;
  if (NS_FAILED(rv))
    return FALSE;

  rv = domWindow->GetDocument(getter_AddRefs(doc));
  if (NS_FAILED(rv))
    return FALSE;

  nsString element_id = wxString_to_nsString(id, wxConvISO8859_1);
  nsCOMPtr<nsIDOMElement> domElement;
  rv = doc->GetElementById(element_id, getter_AddRefs(domElement));

  if(domElement) {
    // GRR all our doc is a wiki, yay.  So how do I use this domElement?
    // We must make it be a member of the class that has the method.
    nsCOMPtr<nsIDOMNSHTMLElement> hElement(do_QueryInterface(domElement));
    // TODO maybe should check that this succeeds?  
    return TRUE;

  // ok, by-id got us nothing, so try to find the first named anchor
  // fprintf(stderr, "seaching for a name=\n");
  nsCOMPtr<nsIDOMNodeList> a_tags;
    NS_LITERAL_STRING("a"), getter_AddRefs(a_tags)

  if(!a_tags) return FALSE;

  PRUint32 count;

  if(!count) return FALSE;

  for(PRUint32 i = 0; i < count; i++) {
    // fprintf(stderr, "check tag %i\n", i);

    nsCOMPtr<nsIDOMNode> node;
    rv = a_tags->Item(i, getter_AddRefs(node));
    if (NS_FAILED(rv) || !node) continue;

    nsCOMPtr<nsIDOMHTMLAnchorElement> anc;
    anc = do_QueryInterface(node);
    if(!anc) continue;

    // make thing, pass in to get return value, lather rinse, repeat...
    nsAutoString name;
    rv = anc->GetName(name);
    if (NS_FAILED(rv)) continue;

    fprintf(stderr, ((wxT("now check ") +
      nsString_to_wxString(name, wxConvISO8859_1) + wxT("\n")).mb_str()));
    if(name == element_id) {
      nsCOMPtr<nsIDOMNSHTMLElement> hElement(do_QueryInterface(node));
      return TRUE;

  return FALSE;

"It works better if you plug it in!"
--Sattinger's Law

More information about the Pdx-pm-list mailing list