- From: cowwoc <cowwoc@bbs.darktech.org>
- Date: Wed, 08 Jan 2014 17:55:06 -0500
- To: public-media-capture@w3.org
- Message-ID: <52CDD74A.5010707@bbs.darktech.org>
On 08/01/2014 5:33 PM, Jan-Ivar Bruaroey wrote: > On 1/8/14 4:57 PM, Harald Alvestrand wrote: >> On 01/08/2014 10:31 PM, Jan-Ivar Bruaroey wrote: >>> On 1/8/14 1:27 PM, Harald Alvestrand wrote: >>>> The constraintName property (and the sdpLineNumber property in the >>>> WebRTC draft) were added for reasons that seemed good at the time; >>>> you can't indicate a line number with an error code. >>> >>> But you can in a message-field, which I understand DOMError will >>> grow. Why is that not sufficient? >> Because that means you can only get it out via parsing the text of >> the field, which means that the format of the text message is part of >> your API, cannot be allowed to vary between implementations, and has >> to be specified in your specifications. > > You're missing my question. Who needs SDP line numbers other than for > debugging? What software needs this information? > > Is someone working on a live SDP editor I don't know about? If so, > then parsing a message-field is the least objectionable thing that > person is doing. I don't think we can predict ahead of time what components users will or will not want to operate on. They might want to transform the error message to improve readability and log it for later review. In any case, I'd err on the side of declaring more programmatic access rather than less. As Harald said, we don't want to force people to starting parsing error messages. Gili
Received on Wednesday, 8 January 2014 22:55:36 UTC