Get form data from webdialogs?

Chris Fullmer

Ah, that did not cross my mind. All my "options" values are scene names. So I'll have to check if scene names can contain commas - if so, I'll have to re-think my delimiter. Thanks!

EDIT - yes, scene names can contain commas. What delimiter do I use then? I'm stumped. Do I forcibly remove commas from the user's scene names before processing? That seems like a bad idea.

thomthom

@chris fullmer said:

Do I forcibly remove commas from the user's scene names before processing? That seems like a bad idea.
No, that's not good. That's making the user work for the computer instead of the computer working for the user.

How about || ?
Much more unlikely to appear in a tab.
Though, there is still a risk of it.

You could make a escape character scheme. Escape | with | - then off course you also need to escape \ with \ .

This example is all Ruby, but you can easily port it to JS.

Escape a string:

string = page.name string.gsub!('\\', '\\\\') string.gsub!('|', '\\|')

Restore it:
string.gsub!('\\\\', '\\') string.gsub!('\\|', '|')

If you do it for all input - output then you will be safe that there will never be any conflict.

TIG

Could you use 'tab' as the delimiter - these are never going to appear in a Scene name ?

      $("select option;selected").each(function () {
        str += $(this).text() + "\t";
      });

Then on the Ruby side use chosen=str.split("\t") to make an array of the items ???

thomthom

@tig said:

Could you use 'tab' as the delimiter - these are never going to appear in a Scene name ?

Actually they can. Paste a string containing Tab and it'll be there. Or it can be added via some Ruby script... pages[0].name = "Foo\tBar"

TIG

OK how about 'newline' ?

          $("select option;selected").each(function () {
            str += $(this).text() + "\n";
          });

then in Ruby chosen=str.split("\n") ?
Or even \r or \f ?

thomthom

That would be even more unlikely to appear, but it's still possible - for instance the Ruby API will allow \n. Who knows what an importer for instance might put in there when importing from some random data. Or if it takes input from somewhere without sanitising.

I mean, it's unlikely, but still possible. Where as using an escape character scheme would make it 100% safe. And it's a simple thing as well.

<span class="syntaxdefault">def escape_pipe</span><span class="syntaxkeyword">(</span><span class="syntaxdefault"> string </span><span class="syntaxkeyword">)<br /></span><span class="syntaxdefault">  string</span><span class="syntaxkeyword">.</span><span class="syntaxdefault">gsub</span><span class="syntaxkeyword">(</span><span class="syntaxstring">'\\'</span><span class="syntaxkeyword">,</span><span class="syntaxdefault"> </span><span class="syntaxstring">'\\\\'</span><span class="syntaxkeyword">).</span><span class="syntaxdefault">gsub</span><span class="syntaxkeyword">(</span><span class="syntaxstring">'|'</span><span class="syntaxkeyword">,</span><span class="syntaxdefault"> </span><span class="syntaxstring">'\\|'</span><span class="syntaxkeyword">)<br /></span><span class="syntaxdefault">end<br /><br />def restore_pipe</span><span class="syntaxkeyword">(</span><span class="syntaxdefault"> string </span><span class="syntaxkeyword">)<br /></span><span class="syntaxdefault">  string</span><span class="syntaxkeyword">.</span><span class="syntaxdefault">gsub</span><span class="syntaxkeyword">(</span><span class="syntaxstring">'\\\\'</span><span class="syntaxkeyword">,</span><span class="syntaxdefault"> </span><span class="syntaxstring">'\\'</span><span class="syntaxkeyword">).</span><span class="syntaxdefault">string</span><span class="syntaxkeyword">.</span><span class="syntaxdefault">gsub</span><span class="syntaxkeyword">(</span><span class="syntaxstring">'\\|'</span><span class="syntaxkeyword">,</span><span class="syntaxdefault"> </span><span class="syntaxstring">'|'</span><span class="syntaxkeyword">)<br /></span><span class="syntaxdefault">end<br /></span>

bigcatln

so far there is no elegant solution for it!
Javascript may be useful but have problem when handling mutibyte characters

thomthom

@bigcatln said:

so far there is no elegant solution for it!
Javascript may be useful but have problem when handling mutibyte characters

Eh? Javascript is unicode compatible... Ruby on the other hand, the 1.8 branch we're stuck with in SU, only deal with strings as ASCII characters.

bigcatln

The problem is JSON doen't support Multibyte characters in fact. when you restore Json str from javascript in ruby,it can't be handled correctly

thomthom

@bigcatln said:

The problem is JSON doen't support Multibyte characters in fact. when you restore Json str from javascript in ruby,it can't be handled correctly

But that's a problem with Ruby 1.8 which only handle ASCII strings - and not a problem with JavaScript...

Pout

I always do as chris does:
collect in one string, send to SU, split there if needed