258802 - nsExternalProtocolService: need default app for scheme

Status: RESOLVED FIXED

(Core :: Networking, enhancement)

Description

[Daniel Veditz [:dveditz]]

We support external protocols by handing them off to the OS, but there is no
current way to know what application will be handling them. Would be nice if we
could let the user know (as we do with helper apps) when prompting whether it's
OK to load an externally handled link.

Would need a new method on nsIExternalProtocolService, not just
isExposedProtocol, but the actual app. All platform implementations of
nsOSHelperAppService would need to be updated.

Comment 1

[Darin Fisher]

biesi: you interested in taking this?

Comment 2

[Christian :Biesinger (don't email me, ping me on IRC)]

sure, why not :)

Comment 3

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: patch v0

something along this lines probably. This patch shows the complete
shell\command\@ string, i.e. path to the application + arguments, which is
probably not optimal.

Comment 4

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: patch v0.1

show <Unknown> as application desc if we can't find a real one

Comment 5

[Christian :Biesinger (don't email me, ping me on IRC)]

bug 252189 adds a way to get a nice human-readable description from the
shell\command filename

Comment 6

[Darin Fisher]

biesi: the patch looks pretty good to me.  is there anything we can do on OSX or
GNOME?

Comment 7

[Christian :Biesinger (don't email me, ping me on IRC)]

Comment on attachment 158867 [details] [diff] [review]
patch v0.1

I have no idea about OSX stuff.

For Gnome, and /etc/mailcap (also on OS/2), we could probably use the filename,
like this patch does for windows.


ok, let's get this patch in while waiting for bug 252189 to get fixed. the
path+filename is good enough for now.

Comment 8

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: unix patch

to be applied in addition to the above patch

Comment 9

[Christian :Biesinger (don't email me, ping me on IRC)]

Comment on attachment 158883 [details] [diff] [review]
unix patch

oops, this breaks gtk1 builds

Comment 10

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: unix patch v2

Comment 11

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 158892 [details] [diff] [review]
unix patch v2

So... can't we factor out the code you're copying from HandlerExists?

Comment 12

[Christian :Biesinger (don't email me, ping me on IRC)]

Comment on attachment 158892 [details] [diff] [review]
unix patch v2

yeah, I shouldn't be so lazy ;)

Comment 13

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: unix patch, v3

factored out. also makes this code use a non-released GConfClient to check the
enabled-state of the app.

Comment 14

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 159089 [details] [diff] [review]
unix patch, v3

r=bzbarsky.  Nice catch on the client thing!

Comment 15

[Ere Maijala (slow)]

Comment on attachment 158867 [details] [diff] [review]
patch v0.1

The patch looks ok to me.

Comment 16

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 158867 [details] [diff] [review]
patch v0.1

sr=bzbarsky, but have a UI owner check out the localized string changes?

Comment 17

[[e-mail]]

Comment on attachment 158867 [details] [diff] [review]
patch v0.1

Nit: as discussed on IRC it seems to me you can simplify the description code
by declaring desc as an nsXPIDLString.

Comment 18

[Christian :Biesinger (don't email me, ping me on IRC)]

Attachment: final patch

Comment 19

[Christian :Biesinger (don't email me, ping me on IRC)]

checked in. bug 264647, bug 264648, and bug 264649 filed for implementations for
other platforms. marking fixed.