2013/5/4 Stuart Henderson <[email protected]> > On 2013/05/04 12:30, Vadim Zhukov wrote: > > This patch shuts up update-mime-database(1) when having KDE4 and/or > > chemical packages installed. Looking at the code, there are already > > some non-standard items (x-* ones), so I don't thing we're going > > to make things worse than they already are. > > Given that RFC6648 (June 2012) recommends *not* using the x- prefix for > "unofficial" protocol extensions, I think things like chemical/* have a > similar status to e.g. x-scheme-handler. At present, considering various > x-* types are already there, and chemical/* is widely used in its field, > I'd be fairly happy to add that as a top-level type. >
To be clear: you'd okay this change... > There was a previous attempt at registering chemical/, which didn't have > sufficient support; I would encourage a suitable person to try again > under the new RFC6838 rules (January 2013) which seem considerably > easier. > > I'm less sure about adding all and uri ... all/* seems to me as if > it's an internal-use hack, and uri/* is annoying as it appears that > x-scheme-handler is meant to be for the same use so it's splitting the > database up, it seems wrong and confusing that some programs should look > in uri/ when determining how to handle some protocol, and other programs > should look in x-scheme-handler/ for the same thing... Though OTOH if we > spam the console with warning messages for things which are a "normal" > situation we risk people ignoring messages when they *are* important... > Don't know.. but not this, right? Does anyone else want to vote, especially on wonky "uri/" and "all/"? I don't see any good we'll have by not adding those except saving a few dozens of bytes in CVS, but I could be wrong easily. :) And you'll have about eight lines of spam on each call of update-mime-database if you have kdelibs-4.x installed, even when installing/removing/updating non-KDE4 things. -- WBR, Vadim Zhukov
