Re: [PATCH 2/2] Add parenthesis to charset value in HTTP header for NOTIFY messages
- From: Ross Burton <ross burton intel com>
- To: Sven Neumann <s neumann raumfeld com>
- Cc: gupnp-list gnome org
- Subject: Re: [PATCH 2/2] Add parenthesis to charset value in HTTP header for NOTIFY messages
- Date: Tue, 19 Apr 2011 15:30:45 +0100
On 19 April 2011 15:13, Sven Neumann <s neumann raumfeld com> wrote:
> On Tue, 2011-04-19 at 16:08 +0300, Zeeshan Ali (Khattak) wrote:
>> On Tue, Apr 19, 2011 at 1:30 PM, Ross Burton <ross burton intel com> wrote:
>> > On 19 April 2011 11:14, Sven Neumann <s neumann raumfeld com> wrote:
>> >> - "text/xml; charset=utf-8",
>> >> + "text/xml; charset=\"utf-8\"",
>> >
>> > Whilst I expect that this is actually mandated by the specs, I wonder
>> > if there is anything that breaks with them there...
>>
>> Although in theory I'm all for this change, we really should do some
>> IOP testing before applying this patch as it could severely break
>> things. If we can't do that, we might want to only apply it if it
>> fixes IOP with some known device out there.
>
> Well, fact is that libgupnp uses quotes in some places and omits them in
> others. I know at least one device (a Philips Blue-Ray Player) that
> didn't detect GUPnP-based devices because the device description
> document was not delivered with the correct HTTP header. This was fixed
> with commit 2745f17e but the fix wasn't applied to all places.
>
> Honstely, I just stumpled across this today when comparing the HTTP
> headers created by our iPhone app with the headers created by libgupnp.
> I have not made any IOP tests with this change.
Let's for now be consistent and correct, and do quotes everywhere. If
we discover a device that stops working when the quotes are present,
we can detect that and react accordingly. But until we do, let's do
what the spec says.
Ross
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]