evolution strips html anchors from links in outgoing mails

Bug #422719 reported by Niall Gallagher
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
gtkhtml3.14
Won't Fix
Medium
gtkhtml3.14 (Ubuntu)
Triaged
Low
Ubuntu Desktop Bugs

Bug Description

Binary package hint: evolution

Ubuntu 9.04 x86_64
Evolution 2.26.1-0ubuntu2

The bug title says it all.

If I type (or paste) a link like this - http://java.sun.com/javase/6/docs/api/java/lang/Runtime.html#exit(int) - into a mail and I hit space after the link, Evolution will convert the text to a hyperlink and when I send the mail the recipient will receive it correctly. When the recipient clicks the link, their browser will open the Runtime.html page and will jump to the exit(int) section in the page correctly. The "#exit(int)" part of the URL is an html anchor which tells the browser where to jump to in the target page.

However let's say I want the display text of the link in my mail to be different from the associated url. Say I write to a colleague "see this method for details" and I want <this method> to be underlined as a hyperlink to the full url above. I type my message as normal, then I highlight <this method> with the mouse and I click "Insert" menu -> "Link..." and it prompts me for the url. I paste the full url including the anchor and click close. Everything looks fine so I send the mail.

Then my colleague asks me which section in the page I was talking about. Evolution has stripped the "#exit(int)" anchor part from the url. When I view the source of the mail I sent in my Sent folder, I see that Evolution seems to have stripped the anchor from the link before even sending it.

Evolution silently strips html anchors from links in outgoing mails, when the display text of the link differs from the url.

As you might have guessed I'm a Java developer and my colleagues are Java deveopers and this issue comes up all the time. It's really annoying. I think it should be marked urgent. HTML anchors are of course part of the HTML spec since the beginning, so Evolution should support them.

To be honest, there must be some code in Evolution which is responsible for "transforming" links like this. This code should throw an exception on encountering content it can't handle, rather than skipping the content. I would rather see an error message or even have Evolution crash, than have Evolution silently corrupt my outgoing mails.

Changed in evolution (Ubuntu):
importance: Undecided → Low
Revision history for this message
Pedro Villavicencio (pedro) wrote :

that's a gtkhtml issue, reassigning, could you also send it upstream to bugzilla.gnome.org since you're facing the issue there? Thanks in advance.

affects: evolution (Ubuntu) → gtkhtml3.14 (Ubuntu)
Revision history for this message
Niall Gallagher (npgall) wrote :

First- my apologies to the authors of Evolution and GtkHtml for my somewhat snarky tone in my bug report above. You guys don't deserve that and you do great work.

I have filed this bug with GtkHtml upstream: https://bugzilla.gnome.org/show_bug.cgi?id=606351

Revision history for this message
Pedro Villavicencio (pedro) wrote :

thanks for sent it upstream.

Changed in gtkhtml3.14 (Ubuntu):
assignee: nobody → Ubuntu Desktop Bugs (desktop-bugs)
status: New → Triaged
Changed in gtkhtml3.14:
status: Unknown → New
Changed in gtkhtml3.14:
importance: Unknown → Medium
Revision history for this message
Kevin Loughrey (kevinl) wrote :

Unwanted characters at the end of link in Evolution email. Inbuilt parsing of hyperlinks in a manner similar to that which is used by a spell checker.

This is a matter related to the one reported above.

I frequently cut from a browser a URL and place it in my email for the convenience of my recipients. I may then type a space character end the sentence with a full stop or punctuate the sentence with a comma. At this point, Evolution's email text editor includes these characters into the URL. The consequence of this is that when the email recipient clicks on the URL they receive a html 404 error from the web-server and many every-day users are alarmed by this. Many also do not know to try to trim off trailing spurious characters so as to make the link, now displayed in their web-browser, legal.

I would like to suggest there should be a built-in parser for URL links within the Evolution email text editor and when there are trailing characters or other characters that are not legal, they should be removed. If a person wants to put at the end of a URL characters that would typically generate a 404 error, then there is a need for there to be a provision within the editor for the author to indicate that atypical requirement by, for example, putting the link within quotation marks.

I hope this comment is helpful. Thank you for your efforts. They are much appreciated.

Changed in gtkhtml3.14:
status: New → Won't Fix
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.