1
Vote

Truncated subject line; better special characters handling needed.

description

Using these example mailto links with Hotmail I have found the following:
 
mailto:anyone@at.all?subject=#$&|~`!@%^*()_-+={[}]:;<,>.?/'"\   ~`!@%^*()_-={[}]:;<,>.?/'\ are characters which are handled.
 

$&| are characters which truncate everything following.

 
  • is replaced with a space.
     
    " is simply ignored (disappears).

comments

jcao219 wrote Oct 17, 2010 at 6:33 AM

Thanks for this report!
I have been meaning to get around to fixing this, but recently I've been busy.

I think it should be trivial to update the parser to fix it.

As for the Protected area thing, I'm not sure how that happens.
I'm not getting it on a Windows 7 test machine.

crenshaw wrote Oct 17, 2010 at 3:55 PM

Great to hear. The "protected area" warning is a one-time UAC prompt one receives the first time GmailDefaultMaker is activated from within IE. Now that I've chosen "Don't warn me for this program" it doesn't do it anymore. As long as there's no way for the mailto to get your handler to do anything then it's fine however.

Thanks and keep up the great work!

crenshaw wrote Nov 8, 2010 at 12:00 AM

Crap sorry. In IE9 beta it didn't look like it posted the first 2 times.. please remove my extra comments :/

jcao219 wrote Nov 8, 2010 at 1:54 AM

I think this is an easy fix, stay tuned.

jcao219 wrote Nov 8, 2010 at 2:14 AM

I think it's fixed now,
(2.0.1.2)

crenshaw wrote Nov 8, 2010 at 2:51 AM

Unforuntately I'm still seeing same truncation as before using that example mailto: in the description of the issue. The quotation seems to be causing the truncation.

jcao219 wrote Nov 9, 2010 at 2:38 AM

Ah I'm sure I've fixed it now.
I thought I had fixed it earlier but I realize now that my browser was automatically encoding the URL. Today testing on Internet Explorer revealed that the problem still existed and the URL wasn't completely encoded.

crenshaw wrote Nov 9, 2010 at 5:50 PM

Using 2.0.1.3 it still does not work. It's still not completely encoded and it needs to be done before it hits your parser because that's most likely the root of the problems with "" as I believe your code is taking this literally to mean the end of the string (truncating everything after).

I'm updating the Description with my results using all of the usual keyboard characters.

jcao219 wrote Nov 9, 2010 at 10:35 PM

The #$&| characters should really always encoded in the link itself because they can cause confusion in any URL.

What the changes earlier fixed was it would decode html special entities (such as ").
The problem with the comic site was because '"' was inside the mailto link instead of %22 (which is standard).

Thanks your helpful descriptions,
I've fixed part of the quoting problem (the truncation part).
The part with the quotation mark disappearing with Internet Explorer is not fixable; I realized that Internet Explorer seemed to be removing unencoded quotation marks before the link even got to the parser.

(Use this release: http://gmaildefault.codeplex.com/releases/view/55413)

crenshaw wrote Nov 10, 2010 at 1:04 AM

Truncation due to quotation marks does seem to be fixed in 2.0.1.4! Thanks :)

Same behavior for #$&| and + however as previously reported. Do these get sent to your parser? If so, could it not url encode them itself? Or is it as with IE removing the " signs? Thanks again so much for the ongoing work on a really fine product.