Question: How to escape a cr/lf character for browser type custom actions

[DoGauss]

Hello Henri,

normally one would escape an carriage return or line feed charater in URLs with %0d (ascii(13)) or %0a (ascii(10)) respectively.
Unfortunately that doesn't seem to work on browser type custom actions - at least on Windows with the current version 0.9.10.

The following example action does create a new email - to and subject fields are also correct, but the mail body is just the literal string "Bitte ein Ticket erstellen und bearbeiten.%0d%0aGruß Dominik"

Is this a bug or do I have to excape a new line differently? Sorry for asking stupid questions, but Google was not a friend of mine on this one yet...

Thanks
Dom

[action_Email to servicedesk]
filter_target_service = True
leave_popwin_open = False
name = Email an Helpdesk
re_status_information_enabled = False
os =
description = Sendet eine Email an den Helpdesk
enabled = True
filter_target_host = True
re_status_information_pattern =
version = 1
re_host_reverse = False
re_status_information_reverse = False
re_service_reverse = False
re_service_enabled = False
re_host_pattern =
re_host_enabled = False
re_service_pattern =
type = browser
close_popwin = True
string = mailto:servicedesk@my.org&subject=Nagios Meldung: $HOST$ - $SERVICE$ - $STATUS-INFO$&body=Bitte ein Ticket erstellen und bearbeiten.%0d%0aGruß Dominik

[HenriWahl]

Does it work without that escapes? Is the result the same with different browsers?

[DoGauss]

I've tested it with every browser at my disposal: IE10, FF22, Chrome 22, opera 12
they all work as I'd expect it: %0d%0a causes a new line in the mail body.

It seem as if nagstamon would prevent the correct parsing of the escape sequences.
If I had to make an educated guess, I'd say the complete action string get's URL encoded before execution and thus breaking the ability to use escape sequences - that would mean: no chance of having new lines in emails.

of course: creating emails this way without new lines works just fine

[HenriWahl]

I played with your example and found it not working with &subject but ?subject.

Line 651-656 in Actions.py:

def _URLify(self, string):
    """
    return a string that fulfills requirements for URL
    exclude several chars
    """
    return urllib.quote(string, ":/=?&@")

So yes URLs get cleaned.

I tested Browser URLs without that _URLify() and it works as you expected. Obviously the browsers could URLify the URLS themselves and this method should not be needed here.

[DoGauss]

Thanks for clarifying.

nonetheless thanks for pointing out a syntactical error. the mailto protocol expects a question mark as separator between the first argument (most often the to field) and other arguments. additional parameters are linked with an ampersand.

[HenriWahl]

I think your way to send emails should work in future version - next github commit to be exact.