<yoav> rniwa: If you're still around, I'd love to discuss srcN implementation concerns
--> jdaggett ([email protected]) has joined #webkit
<rniwa> yoav: I've expressed all my concerns in the mail I sent ot
<rniwa> out*
--> msanchez ([email protected]) has joined #webkit
<yoav> rniwa: Yeah, I saw that, but I'd love to have some more details on that optimization and its impact on srcN's parsing flow
 Since srcN attributes are rarely expected to be addressed explicitly, I wonder if there's isn't a way to avoid the upper bound restriction
<rniwa> yoav: mike just replied on webkit-dev thread
 yoav: saying that people rarely need more than 2-3 src-* specified
 yoav: why do you think lifting the restriction is important?
 yoav: I seriously can't think of a use case where you'd end up specifying 10-20 src-N's
<yoav> rniwa: I don't see Mike's response yet for some reason...
 I completely agree that 2-3 srcs are typical and I wouldn't expect more than 6-7 breakpoints in any normal Web site
<rniwa> yoav: I also share the concern others have expressed
 yoav: which is that we have a list of lists
 yoav: and having N attributes to express this information is distasteful
<-- tasak has quit (Remote host closed the connection)
--> tasak (tasak@nat/google/x-neazqxnonyhyugaw) has joined #webkit
<rniwa> yoav: don't get me wrong. i'm trying to block your work. i think the feature addresses a very important use case
 yoav: i'm just disagreeing with the specific API being proposed to be implemented in webkit
<-- svillar_ has quit (Quit: Ex-Chat)
<yoav> rniwa: What worries me about that restriction is that it may make authoring more difficult, by giving authors less freedom to "space out" the attribute order, and add more attributes in the future
--> svillar ([email protected]) has joined #webkit
<-- Company has quit (Read error: Connection reset by peer)
<yoav> rniwa: With that said, if this is a requirement from WebKit, I don't think it's a compromise I can live with, it's just something I wish we can avoid
<noamr> elecro: could you comment on https://bugs.webkit.org/show_bug.cgi?id=123246? do we need PlatformCertificateInfo there?
<yoav> rniwa: s/don't//
 rniwa: To be clear, it is a compromise I can live with, but wish we can avoid it
<-- noraesae has quit (Quit: noraesae)
 miChou has quit (Quit: Leaving.)
 tasak has quit (Ping timeout: 252 seconds)
<rniwa> yoav: I actually think we should revise the spec not to have N attributes
 yoav: and just have single attribute with some internal format
<rniwa> that'll be simpler than the current solution
--> mZugn (Marton@nat/u-szeged/x-fvxglviypmmskjsg) has joined #webkit
<rniwa> that invokes N attributes and M values in each one of them :/
<-- mihnea has quit (Remote host closed the connection)
<yoav> rniwa: I'm not against that, but not a "syntax person" myself, so I don't have a particular opinion on what would be easier for authors
<rniwa> yoav: being easy for author is one thing
<yoav> It's a tradeoff between one single huge attribute and multiple, smaller ones
<rniwa> yoav: but being consistent with the rest of platform is another
--> tasak (~tasak@2401:fa00:4:1012:b8fb:97d5:d684:3818) has joined #webkit
<-- greenspa has quit (Remote host closed the connection)
 narkeeso has quit (Quit: Computer has gone to sleep.)
 simathur has quit (Quit: ChatZilla 0.9.90.1 [Firefox 24.0/20130910160258])
<yoav> rniwa: Very well. I think I understand your point of view. Thanks! :)
<rniwa> yoav: np<yoav> rniwa: If you're still around, I'd love to discuss srcN implementation concerns
--> jdaggett ([email protected]) has joined #webkit
<rniwa> yoav: I've expressed all my concerns in the mail I sent ot
<rniwa> out*
--> msanchez ([email protected]) has joined #webkit
<yoav> rniwa: Yeah, I saw that, but I'd love to have some more details on that optimization and its impact on srcN's parsing flow
 Since srcN attributes are rarely expected to be addressed explicitly, I wonder if there's isn't a way to avoid the upper bound restriction
<rniwa> yoav: mike just replied on webkit-dev thread
 yoav: saying that people rarely need more than 2-3 src-* specified
 yoav: why do you think lifting the restriction is important?
 yoav: I seriously can't think of a use case where you'd end up specifying 10-20 src-N's
<yoav> rniwa: I don't see Mike's response yet for some reason...
 I completely agree that 2-3 srcs are typical and I wouldn't expect more than 6-7 breakpoints in any normal Web site
<rniwa> yoav: I also share the concern others have expressed
 yoav: which is that we have a list of lists
 yoav: and having N attributes to express this information is distasteful
<-- tasak has quit (Remote host closed the connection)
--> tasak (tasak@nat/google/x-neazqxnonyhyugaw) has joined #webkit
<rniwa> yoav: don't get me wrong. i'm trying to block your work. i think the feature addresses a very important use case
 yoav: i'm just disagreeing with the specific API being proposed to be implemented in webkit
<-- svillar_ has quit (Quit: Ex-Chat)
<yoav> rniwa: What worries me about that restriction is that it may make authoring more difficult, by giving authors less freedom to "space out" the attribute order, and add more attributes in the future
--> svillar ([email protected]) has joined #webkit
<-- Company has quit (Read error: Connection reset by peer)
<yoav> rniwa: With that said, if this is a requirement from WebKit, I don't think it's a compromise I can live with, it's just something I wish we can avoid
<noamr> elecro: could you comment on https://bugs.webkit.org/show_bug.cgi?id=123246? do we need PlatformCertificateInfo there?
<yoav> rniwa: s/don't//
 rniwa: To be clear, it is a compromise I can live with, but wish we can avoid it
<-- noraesae has quit (Quit: noraesae)
 miChou has quit (Quit: Leaving.)
 tasak has quit (Ping timeout: 252 seconds)
<rniwa> yoav: I actually think we should revise the spec not to have N attributes
 yoav: and just have single attribute with some internal format
<rniwa> that'll be simpler than the current solution
--> mZugn (Marton@nat/u-szeged/x-fvxglviypmmskjsg) has joined #webkit
<rniwa> that invokes N attributes and M values in each one of them :/
<-- mihnea has quit (Remote host closed the connection)
<yoav> rniwa: I'm not against that, but not a "syntax person" myself, so I don't have a particular opinion on what would be easier for authors
<rniwa> yoav: being easy for author is one thing
<yoav> It's a tradeoff between one single huge attribute and multiple, smaller ones
<rniwa> yoav: but being consistent with the rest of platform is another
--> tasak (~tasak@2401:fa00:4:1012:b8fb:97d5:d684:3818) has joined #webkit
<-- greenspa has quit (Remote host closed the connection)
 narkeeso has quit (Quit: Computer has gone to sleep.)
 simathur has quit (Quit: ChatZilla 0.9.90.1 [Firefox 24.0/20130910160258])
<yoav> rniwa: Very well. I think I understand your point of view. Thanks! :)
<rniwa> yoav: np