W3C home > Mailing lists > Public > public-device-apis@w3.org > April 2010

ReSpec: potentially breaking change

From: Robin Berjon <robin@robineko.com>
Date: Tue, 13 Apr 2010 13:36:31 +0200
Message-Id: <68F6894F-A041-4066-8C51-7A3AA48E8A9D@robineko.com>
To: public-device-apis@w3.org, chairs <chairs@w3.org>, Spec Prod <spec-prod@w3.org>
Hi all,

Dom spotted a bug in the way that ReSpec handled the fields in exception definitions both in parsing and generation. I have now fixed the bug in the live version of ReSpec.

If you do not have any exception definitions, or if those you have don't introduce their fields with "readonly attribute" or "attribute" you're safe. If you did, however, your document is now broken. The fix is to remove those indications since they aren't actually allowed in WebIDL.

If this causes too much carnage I know how to make it backwards compatible but I would rather avoid that as it would essentially mean that this bug in the input format is to be maintained forever.

This brings me to a more general question. Quite a few groups have started using ReSpec (I know of at least DAP, WebApps, MAWG, RDFa, XML Security, and I've seen commits to the biblio DB that seem to indicate more) but when this sort of problem arises I don't know how to get in touch with everyone. I also get a fair amount of feature requests and discussions about how to implement them that could probably best be discussed as a group. Should we have a ReSpec mailing list? Or should we officially highjack spec-prod at the risk of annoying its current denizens? Should we introduce library versioning to avoid this sort of issue? Should we work on v2? Etc. etc.


Robin Berjon
  robineko  hired gun, higher standards
Received on Tuesday, 13 April 2010 11:37:01 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:32:19 UTC