On Tue, Jul 9, 2013 at 2:30 PM, Alex Harui <[email protected]> wrote: > It looks like the main difference is that embedded assets like the > BusyCursor and Drag cursors are not as efficiently stored as they used to > be. In Adobe SDKs, the assets came from a SWF. For Apache, we had to > convert the assets to FXG. I thought the FXG compiler would do a better > job of converting FXG back to SWF assets, but it seems to be using a lot > of extra Sprites and Shapes because the FXG has a lot of what seem to be > unnecessary Groups. > > -Alex >
I can corroborate this theory. The FXG spit out by tools like AI, Fireworks, Flash, etc. need to be cleaned up before being used in Flex apps. FWIW, I am currently working on a FXG optimizer tool that will hopefully remove the unnecessary groups that are created. Perhaps this logic can be moved into the compiler at some point as well? Thanks, Om > > On 7/9/13 9:37 AM, "João Fernandes" <[email protected]> > wrote: > > >Ok done! > > > >Thanks! > > > > > >On 9 July 2013 17:28, Alex Harui <[email protected]> wrote: > > > >> João, > >> > >> > >> The mail filters blocked it. Try making the suffix of the link reports > >> something like .xml or .txt and send again. I think it doesn't like > >>.lnk. > >> > >> Thanks, > >> -Alex > >> > >> > >> On 7/9/13 9:25 AM, "João Fernandes" > >><[email protected]> > >> wrote: > >> > >> >Thanks Alex for your time, I've zipped and sent 2 zip files with those > >> >report size files. > >> > > >> >Thank you once again for taking a look at it. > >> > > >> >João Fernandes > >> > > >> > > >> >On 9 July 2013 17:04, Alex Harui <[email protected]> wrote: > >> > > >> >> João, > >> >> > >> >> > >> >> Can you generate a link-report for 4.6 and 4.9.1 and email it to me > >> >> directly? > >> >> > >> >> Are you using Spark DG in the app? I did see additional assets > >>added to > >> >> the default DG styles to support drag/drop. > >> >> > >> >> Thanks, > >> >> -Alex > >> >> > >> >> On 7/9/13 8:57 AM, "João Fernandes" > >> >><[email protected]> > >> >> wrote: > >> >> > >> >> >What do you mean Tom? for testing purpose I've compiled the same > >>app / > >> >> >modules with both 4.5.2 / 4.6.0 / 4.9.1 and for each file there is a > >> >>file > >> >> >size increase with the exact same build mechanism, where isn't that > >>"an > >> >> >issue"? I agree (from the beginning) that it might not be a major > >>issue > >> >> >but > >> >> >it's one nevertheless. > >> >> > > >> >> >What I've reported form the size-report is the difference from MXMLC > >> >>for > >> >> >each SDK Version so clearly, the compiler is adding stuff which > >>wasn't > >> >> >doing previously. > >> >> > > >> >> > > >> >> >On 9 July 2013 16:21, Tom Chiverton <[email protected]> wrote: > >> >> > > >> >> >> Given you don't even get the same file from two compiles of the > >>same > >> >> >>code > >> >> >> at different times, I think 'issue' is stretching it ;-) > >> >> >> > >> >> >> Tom > >> >> >> > >> >> >> > >> >> >> On 09/07/2013 15:18, João Fernandes wrote: > >> >> >> > >> >> >>> 4.5.2 was a prerelease version which we accepted it as good > >>enough > >> >>(no > >> >> >>> signed RSLs) and never upgraded to 4.6.0. We're now upgrading to > >> >>Apache > >> >> >>> 4.9.1 and having this "small" issue. > >> >> >>> > >> >> >>> > >> >> >>> On 9 July 2013 15:08, David Hamiter > >> >> > >>>>><intoxopox@professorfripples.**com<[email protected]>> > >> >> >>> wrote: > >> >> >>> > >> >> >>> Sorry to be a bit off topic, but I'm curious, João... where did > >>you > >> >> >>>find > >> >> >>>> Flex SDK 4.5.2? Was that a nightly build or something? The only > >> >> >>>>official > >> >> >>>> 4.5 I can find is 4.5.1.21328A. If your copy of 4.5.2 appears > >> >>stable, > >> >> >>>>I > >> >> >>>> would love to get a copy of it. > >> >> >>>> > >> >> >>>> David Hamiter > >> >> >>>> > >> >> >>>> -----Original Message----- From: João Fernandes > >> >> >>>> Sent: Tuesday, July 9, 2013 4:27 AM > >> >> >>>> To: [email protected] > >> >> >>>> Subject: Re: bigger swfs in 4.9.1 > >> >> >>>> > >> >> >>>> > >> >> >>>> Ok, I've compared size-report from 4.5.2 and 4.9.1 and I'm > >>noticing > >> >> >>>>that > >> >> >>>> both frames have grown in size, 1.1k in frame 1 and 4k in frame > >>2. > >> >> >>>> I've noticed that there are 132 new sprite entries and 159 new > >> >>shape > >> >> >>>> entries, is this expected? > >> >> >>>> I've compared 4.5.2 with 4.6.0 and 4.6.0 and they are almost > >> >>identical > >> >> >>>> (4.6 > >> >> >>>> produces even better results). > >> >> >>>> > >> >> >>>> João Fernandes > >> >> >>>> > >> >> >>>> > >> >> >>>> On 9 July 2013 08:27, João Fernandes > >><joaopedromartinsfernandes@** > >> >> >>>> gmail.com > >> >> >>>><joaopedromartinsfernandes@**gmail.com > >> >> <joaopedromartinsfernandes@gmail. > >> >> >>>>com> > >> >> >>>> >>wrote: > >> >> >>>> > >> >> >>>> > >> >> >>>> > >> >> >>>> On 9 July 2013 01:51, Alex Harui <[email protected]> wrote: > >> >> >>>>> > >> >> >>>>> > >> >> >>>>> Those shouldn't affect SWF size. But any changes to classes > >>that > >> >> >>>>>go in > >> >> >>>>>> Frame 1 like SystemManager, or in generated code (like the CSS > >> >>data) > >> >> >>>>>> could > >> >> >>>>>> make for larger SWFs. > >> >> >>>>>> > >> >> >>>>>> > >> >> >>>>>> Alex, I don't think it's related to CSS since I see the > >> >>increase > >> >> >>>>>>in > >> >> >>>>>> size > >> >> >>>>>> > >> >> >>>>> just by switching SDKs and comparing generated file sizes, > >>unless > >> >>the > >> >> >>>>> compiler suffered changes, no? Do you think it will be easier > >>to > >> >>find > >> >> >>>>> the > >> >> >>>>> culprint comparing report-size files? > >> >> >>>>> > >> >> >>>>> -- > >> >> >>>>> > >> >> >>>>> João Fernandes > >> >> >>>>> > >> >> >>>>> > >> >> >>>>> > >> >> >>>> -- > >> >> >>>> > >> >> >>>> João Fernandes > >> >> >>>> > >> >> >>>> > >> >> >>> > >> >> >>> > >> >> >> > >> >> > > >> >> > > >> >> >-- > >> >> > > >> >> >João Fernandes > >> >> > >> >> > >> > > >> > > >> >-- > >> > > >> >João Fernandes > >> > >> > > > > > >-- > > > >João Fernandes > >
