justin = { main feed , music , code , askjf , pubkey };
Ask Justin Frankel
No reasonable question unanswered since 2009!

Suggested topics: programming, music, sleep, coffee, etc.

Note: please do not ask questions about REAPER features, bugs or scheduling, use the forums instead.


Name: Ask: Human (enter yes):
[newer questions][unreplied] | [replied] | [recent comments] | [all]
[older questions]

Question: channel9.msdn.com/Shows/Going+Deep/Expert-to-Expert-Erik-Meijer-... 51:55 Lars Bak (author of v8) discusses your concern of multithreading in js.
Asked by Will (68.104.121.x) on March 29 2014, 5:58am
Reply on March 29 2014, 2:50pm:
    Yeah, they're saying you could have completely separate VMs that pass messages back and forth, which can be very useful but is also no substitute for real threads. My point is that for simple apps the web stack is getting there, but for intensive media applications, it has a long and unclear road ahead.
Comment...
Question: I can't let VC6 go too, it's got that something special. but for the argument sake: have you tried in VC13 /LTCG:PGO and /favor:xxx? I don't like the idea of cpu dispatching in ICC but to be fair i'm using an AMD cpu and never had performance related problems w/ R. I remember clearly the days you made the switch and everything turned lightning fast immediate
Asked by gio (94.66.76.x) on March 26 2014, 9:46pm
Reply on March 26 2014, 9:55pm:
    I haven't tried profile-guided optimizations, it'd be tough to decide what to run for testing, would want to come up with a bunch of test projects perhaps...

    Re: ICC performance: actually, from my initial tests, VC2013 vs VC6+ICC actually was pretty comparable on a Ivy Bridge i7. On a c2d though VC6+ICC was significantly faster. Unsure of how it would look on an AMD chip, but I'd guess that the VC6+ICC build would at least be comparable to the VC2013 build.
Comment...
Question: VC2013 barfs but VC6 does not? Could you list 2 or 3 things VC6 supports properly which VC2013 does not?
Asked by Will (24.234.85.x) on March 26 2014, 4:56pm
Reply on March 26 2014, 7:05pm:
    The latest ICC on VC2013 was barfing on some code. Normal VC2013 is fine (but VC6+ICC produces faster code in general than VC2013 w/o ICC).
Comment...
Question: Is this you? www.theverge.com/2014/3/25/5547628/oculus-investor-says-facebook...
Asked by Justin (76.90.239.x) on March 26 2014, 1:05am
Reply on March 26 2014, 1:12am:
    I was thinking about ordering a rift 2 devkit, but now not so much.
Comment...
Question: Facebook recently bought Oculus, the VR company that employs (among other people) John Carmack. Is this real life? Or some weird tech-bubble false-reality?
Asked by Eric (70.191.121.x) on March 26 2014, 12:41am
Reply on March 26 2014, 1:12am:
    Heh you couldn't make that shit up.
Comment...
Question: Does that mean you're using VC2013 as your primary C IDE now?
Asked by Will (24.234.85.x) on March 25 2014, 8:55pm
Reply on March 25 2014, 11:27pm:
    I use it often, but I don't really have a "primary" anything. It isn't used to compile our release builds yet, either (I did test VC13 with the latest ICC, but it barfed a lot... it'll get mature soon enough though).
Comment...
Question: The modern web stack offers so much! Are you gonna die clinging to win32 and visual c++ 6.0? :)
Asked by Will (24.234.85.x) on March 24 2014, 6:44pm
Reply on March 24 2014, 7:03pm (edited at March 24 2014, 7:04pm):
    I've been enjoying VC2013 and Xcode5/LLVM (Win64 and SWELL rounding out the API targeting). The modern web stack does offer a lot, but it also is missing even more. There is a long, LONG way to go for a lot of interesting things. Random off the top of my head pick: multithreading support?
Comment...
Question: Would you use this address instead? nsis.neotext.org
Asked by Nick (173.31.164.x) on March 22 2014, 3:20am
Reply on March 24 2014, 2:36am:
    Nope
Comment...
Question: Global auto destroy ptrlist:now you're telling me :D The small problem is that whole SWS is filled with that idiom using SWSProjConfig
Asked by Martin (93.136.3.x) on March 18 2014, 2:31pm
Reply on March 18 2014, 3:41pm:
    email me :)
Comment...
Question: Problem is I can't be picky when things get deleted since they reside in WDL_PtrListDeleteOnDestroy within WDL_PtrList which manages stuff in all active projects. So sometimes the object may get deleted because it's not needed, and sometimes when Reaper quits :) However, this seems to do the trick for now: if (EnumProjects(0, NULL, 0)) DestroyAudioAccessor(m
Asked by Martin (93.139.111.x) on March 18 2014, 1:17am
Reply on March 18 2014, 1:29am (edited at March 18 2014, 1:51am):
    Just don't delete anything that you don't need to in the unload, then (and don't make your global ptrlists DeleteOnDestroy)... On OSX you will not get your DLL unloaded, too.

    Anyway fixing it on our end too, mostly.
Comment...
Question: I wrote some small patches for NSIS back in the day, and you were kind enough to include me in the credits. As a nerdy teen i found that cool.. thanks man! Anyway you mentioned liking East London. If you ever need a coffee buddy here give me a shout!
Asked by PeterWindridge (81.157.72.x) on March 17 2014, 10:47pm
Reply on March 17 2014, 11:36pm:
    Ah nice! Thanks for your contributions!
Comment...
Question: I have DestroyAudioAccessor call in destructor of an object that gets deleted in REAPER_PLUGIN_ENTRYPOINT() on !rec If accessor is not NULL I get Access violation reading location (location is not NULL). Is this Reaper's fault or I shouldn't do that?
Asked by Martin (93.139.111.x) on March 17 2014, 10:11pm
Reply on March 17 2014, 11:35pm (edited at March 17 2014, 11:39pm):
Comment...
Question: Re LE: I would expect exactly what you described for REAPER_FOURCC('a','b','c','d') >> "dcba" but I get "abcd". e.g. Y800 should hex to 0x30303859 but it isn't, I get 0x59383030. Unless things has gone terribly wrong here could you have a closer look please? Using 2009 Mini-SDK, thanks.
Asked by gio (94.66.14.x) on March 13 2014, 3:25pm
Reply on March 13 2014, 4:27pm:
    On x86 (LE), REAPER_FOURCC(1,2,3,4) is 0x01020304, and "4321" in memory... so REAPER_FOURCC is backwards from what you expect, but consistent across platforms and has always been that.
Comment...
Question: Clarification plz: Is the byte order of REAPER_FOURCC @reaper_plugin.h correct for little-endian?
Asked by gio (94.66.14.x) on March 13 2014, 6:51am
Reply on March 13 2014, 2:32pm:
    Not sure what "Correct" means, but REAPER_FOURCC('a','b','c','d') will result in "dcba" being stored in memory on all platforms.
Comment...
Question: Taking back my question...items were being resampled! Don't you hate when things like that happen during testing? :)
Asked by Martin (93.139.50.x) on March 13 2014, 5:17am
Reply on March 13 2014, 2:32pm:
    Phew!
Comment...
Question: Is there a reason that GetAudioAccessorSamples is so much slower than PCM_source->GetSamples(). The difference is staggering when getting 4 minutes of wav. We're talking around 3 seconds as opposed to 25 seconds.
Asked by Martin (78.0.233.x) on March 13 2014, 4:47am
Reply on March 13 2014, 2:32pm:
    Some additional work/mutex locking perhaps but not enough to cause that. hm.
Comment...
Question: Is there a contact email for a ninjam tech guru? I'm composing a book about online jamming solutions and hope to include Ninjam in it.
Asked by Steve (124.125.48.x) on March 9 2014, 2:47pm
Reply on March 11 2014, 11:02pm:
    You can send me an email (my first name at my company name dot com).
Comment...
Question: I haven't been succeeding at the latter lately, but I got another question! Can I hide install directory choose when it's open over already install in nsis?
Asked by Nick Forystek (173.31.164.x) on March 6 2014, 12:16pm
Reply on March 11 2014, 11:02pm:
    Yeah, pretty sure you can, but I forget at the moment.
Comment...
Question: mp3DirectCut can split mp3 files without reencoding, but interface is horrible... could you code something like that? It would be great.
Asked by Rodrigo (189.26.143.x) on March 3 2014, 2:46pm
Reply on March 4 2014, 2:55am:
    Hmm it mostly would be pretty straightforward, but UI is pain, and I forget how the bit reservoir works/would affect such a thing.
Comment...
Question: Re: "Using software you make." I get the distinct feeling that some workstation developers actually don't, at least not on the level of their most loyal users. Big problem?
Asked by PunkyBrewster (76.226.54.x) on March 3 2014, 2:40pm
Reply on March 4 2014, 2:54am:
    "Using every day" != "Using like 0.01% of the users" IMO. There's a balance to be found, and a lot of the massive feedback you get needs to be sorted and filtered inside, where it can be worked into a consistent vision... or something.
Comment...
[newer questions][unreplied] | [replied] | [recent comments] | [all]
[older questions]
Copyright 2024 Justin Frankel. | RSS