| drobilla | falktx: You using -Wcast-align too? | 00:02 |
|---|---|---|
| falktx | not sure now | 00:02 |
| drobilla | As I write down how atoms are by definition 64-bit aligned, knowing that my code is totally not -Wcast-align clean by any measure, I wonder how actually true that is :) | 00:02 |
| drobilla | The NASA JPL coding guidelines are a good source for insane compiler warnings and code quality measures, by the way. | 00:03 |
| falktx | nope, no cast-align | 00:03 |
| falktx | but I'll add it now | 00:03 |
| drobilla | (LV2 iiiiiiinnnnnn spaaaaaaaaaaaacccccccceeeeeeeee) | 00:03 |
| falktx | I want to use '-Wsign-conversion' but Qt clashes with this | 00:04 |
| drobilla | Gtk2 violates a few I use too, the strict prototypes ones IIRC. | 00:06 |
| drobilla | Those only apply to C though. | 00:06 |
| drobilla | (Great for ensuring that every function that can be static is static in C though) | 00:07 |
| falktx | what about "-ansi -pedantic" ? | 00:07 |
| falktx | that's usually fun... | 00:07 |
| drobilla | Most of my C++ is C++0x these days, but I do use pedantic | 00:07 |
| drobilla | http://pastebin.com/vFTePKRe | 00:08 |
| drobilla | I keep two levels, ultra-strict is for "stuff I look and and would like to be clean for, but aren't" | 00:09 |
| drobilla | I skip through errors in emacs, so any warning really gets in the way, strict is day-to-day dev flags. | 00:09 |
| *** falktx has quit IRC | 00:31 | |
| *** falktx has joined #lv2 | 00:53 | |
| *** HarryHaaren has quit IRC | 01:02 | |
| drobilla | Output ports with a variably sized type MUST be initialised by the host before every run() to an atom:Chunk with size set to the available space. The plugin reads this size to know how much space is available for writing. In all cases, the plugin MUST write a complete atom (including header) to outputs. However, to be robust, hosts SHOULD initialise output ports to a safe sentinel (e.g. the null Atom) before | 01:03 |
| drobilla | calling run(). | 01:03 |
| drobilla | Oops. This is a contradiction. | 01:03 |
| drobilla | Not sure what the best rule here is. I'd be tempted to say hosts must set the type to 0 and initialize the size, but this may conflict with future use of the reserved type 0 (with non-zero size) for references. | 01:08 |
| drobilla | <characteristicShortNameGripe>InputPort and OutputPort were kind of silly when Inlet and Outlet are there for the taking | 01:19 |
| *** HarryHaaren has joined #lv2 | 01:22 | |
| drobilla | Maybe it's best left undefined for now, then. i.e. the only thing plugins can depend on is that the size indicates the amount of space available for writing | 01:24 |
| *** Gethiox4 has quit IRC | 01:32 | |
| drobilla | falktx: Speaking of warning flags, LV2 itself is inherently not const-correct thanks to our friend connect_port() | 01:33 |
| drobilla | (Also known as The Thing We Did Totally Wrong) | 01:33 |
| falktx | heh, forgot about that one | 01:34 |
| drobilla | We really should have left room for extensibility in LV2_Descriptor | 01:35 |
| drobilla | (extension_data can do it, of course) | 01:35 |
| drobilla | Anchakor: My solution to RDF wankery knee-jerk was to talk about the mechanics completely without referring to specific type at all, so it's all bits and bytes until section 3 | 02:14 |
| drobilla | Then type hierarchy, which is mostly numbers anyway. No Turtle until at least page 3 | 02:15 |
| drobilla | 1. Introduction 2. Mechanics 3. Semantics (4-* ?). I think this works well. Saying "Semantics" might be a red flag, but it's just such a tidy split :) | 02:16 |
| drobilla | rgareus: Deadline for paper-submissions: January 27th, 2014 (23:59 HAST) | 02:45 |
| rgareus | drobilla: sounds about right. | 02:46 |
| drobilla | rgareus: ... Hawaiian-Aleutian Standard Time? Am I missing something? :) | 02:46 |
| rgareus | drobilla: that's correct. we're on Hawaiian time | 02:46 |
| drobilla | Okay then. | 02:46 |
| rgareus | drobilla: it gives pretty much everyone a good extra day, and it allows Frank to lazily close submission after breakfast in europe. | 02:47 |
| drobilla | heh | 02:47 |
| rgareus | (though it's a pretty safe bet that there'll be a deadline extenstion.) | 02:48 |
| rgareus | extension, even. | 02:48 |
| drobilla | I didn't hear that. | 02:48 |
| rgareus | but you can submit and update (until submission really closes) | 02:48 |
| drobilla | 4:59 a.m. EST, so basically, "before sleep time" for me | 02:48 |
| falktx | rgareus: who is going to read the papers btw? | 02:51 |
| falktx | I really didn't want to "waste" too much time on it | 02:51 |
| falktx | I mean, right now I see the paper more like a formality | 02:52 |
| rgareus | falktx: a peer review comittee will read it. at least 2 independent persons per paper | 02:52 |
| falktx | not something that really needs to be perfect | 02:52 |
| falktx | rgareus: can submissions get rejected? if so, why? and how many usually do? | 02:55 |
| falktx | (sorry for all the questions, I'm a very curious person) | 02:55 |
| rgareus | falktx: yes, there's a fixed number of presentation slots. | 02:56 |
| rgareus | falktx: and the rejection rate depends on the number of submissions. the highest so far was about 50% when LAC was at stanford. | 02:56 |
| rgareus | falktx: last year I think rejection rate was 10-20% | 02:56 |
| rgareus | falktx: Frank Neumman (AudioFranky on IRC) is chair of the review committe | 02:57 |
| rgareus | falktx: he knows this stuff. | 02:57 |
| rgareus | falktx: I only learn most of this over beers at the conference. -- scientific ethics demand from me that I don't peek :) | 02:58 |
| rgareus | (which I could do as root) | 02:58 |
| falktx | hmm, I need to speak to him then | 02:58 |
| falktx | see if I can persuade him... | 02:59 |
| rgareus | falktx: well, the review is done by independent peers, but most of which are just the usual LAU/LAD crowd. | 03:00 |
| rgareus | falktx: eg. http://lac.linuxaudio.org/2012/ has a "Review committee" list for said year | 03:00 |
| rgareus | falktx: unless you write complete BS, I dare say there's a high chance of being accepted one way or another. | 03:02 |
| drobilla | falktx: I wouldn't extrapolate too much from my worries, in a conference like this 'show and tell' papers about projects don't really need to adhere to some lofty standard | 03:02 |
| falktx | I have the topics for the paper in place, just need to write the content now | 03:04 |
| drobilla | I'm a Computer Scientist describing some lofty ideas that are a bit to left field for this crowd, and am basically trained to try really hard to make them bulletproof because it's my job, and in the sort of conferences usual in my field, aggressive rejection definitely happens. | 03:04 |
| drobilla | e.g. the conference I submitted my last failure to had a 15% acceptance rate | 03:04 |
| drobilla | I'm only really woried about people like Fons who I'm pretty sure personally hate me :) | 03:05 |
| rgareus | The two main idea ideas behind LAC papers: (A) written documentation is always good and (B) provide creditablity | 03:06 |
| drobilla | Yeah, this process is really making me realize I've been fucking up by not doing it every year, because then I'd at least have publications to show for this absurdly massive amount of work I've done. | 03:06 |
| rgareus | drobilla: you are free to submit the paper blinded for review. | 03:06 |
| drobilla | rgareus: Yeah, big mystery there :) | 03:07 |
| rgareus | drobilla: I was just going to say that | 03:07 |
| rgareus | drobilla: that's why we stopped bothering. | 03:07 |
| drobilla | I'm sure it's fine. I'm sort of wired to assume everyone hates me. | 03:08 |
| rgareus | there's a very strong correlation author <> topic in linuxaudio | 03:08 |
| drobilla | (and this stuff in particular) | 03:09 |
| drobilla | rgareus: speaking of creditability, shame your lightning talk on GL UIs isn't in the proceeding. How should I cite that? | 03:14 |
| drobilla | Fancy UIs in general are a good use case for needing a real protocol, should be mentioned. | 03:14 |
| rgareus | drobilla: mmh. I don't know. I was drafted for a speaker that could not come and jumped in for this lightning talk with 4 hours preparation! | 03:14 |
| drobilla | heh | 03:15 |
| rgareus | drobilla: there's no real publication. | 03:15 |
| drobilla | Just asking in case you have a .bib entry hanging around | 03:15 |
| *** NickSB2 has joined #lv2 | 03:16 | |
| rgareus | drobilla: I suppose http://lac.linuxaudio.org/2013/program/83 will work as reference | 03:16 |
| drobilla | Even the W3C stuff doesn't really publish, my bibliography is ultra questionable from a scholarly standpoint, but such is tech in the trenches | 03:16 |
| rgareus | title: openGL UIs for LV2 plugins ; year: 2013 ; <URL> | 03:17 |
| rgareus | 'tech in the trenches' hell yeah! | 03:19 |
| drobilla | Yeah, I like "in the trenches". Usually contrasted against either completely out of touch theory types, or Enterprisey™ wankers :) | 03:20 |
| rgareus | the vast majority of RFCs do not have any bib. or just a few references | 03:20 |
| drobilla | Yeah. Unfortunately bibtex styles are all over the place, I usually end up just ramming stuff in there wherever to make the output look sensible. | 03:22 |
| drobilla | (Not a huge fan of ACL style) | 03:22 |
| *** falktx_ has joined #lv2 | 03:30 | |
| *** falktx has quit IRC | 03:33 | |
| *** HarryHaaren has quit IRC | 03:35 | |
| * drobilla decides that citing LISP is a stretch | 03:35 | |
| drobilla | http://www.cs.uni.edu/~wallingf/blog-images/people/john-mccarthy.jpg | 03:36 |
| * drobilla wants a poster of this so bad | 03:37 | |
| rgareus | LOL | 03:39 |
| *** falktx_ has quit IRC | 07:21 | |
| *** edogawa has joined #lv2 | 07:32 | |
| *** Gethiox has joined #lv2 | 09:54 | |
| *** timbyr has quit IRC | 10:22 | |
| *** timbyr has joined #lv2 | 10:24 | |
| *** shanipribadi has joined #lv2 | 13:12 | |
| *** falktx has joined #lv2 | 13:13 | |
| *** NickSB2 has quit IRC | 14:23 | |
| *** Gethiox2 has joined #lv2 | 15:15 | |
| *** Gethiox has quit IRC | 15:15 | |
| falktx | and now I realized I had even more bad lv2 time types | 15:29 |
| falktx | time:beat is a double, I had it as long | 15:29 |
| falktx | http://lv2plug.in/ns/ext/time/#beat | 15:29 |
| falktx | errrrr :( | 15:29 |
| *** falktx has quit IRC | 15:57 | |
| *** falktx has joined #lv2 | 16:05 | |
| *** rncbc|AFK has quit IRC | 16:08 | |
| *** rncbc has joined #lv2 | 16:18 | |
| * drobilla wonders if a C atom validator is possible | 17:04 | |
| drobilla | falktx: i.e. your host was sending that to plugins? (when receiving you really should check anyway) | 17:04 |
| falktx | yes, sorry | 17:05 |
| falktx | not sure how I made that, I copied some from jalv. is it possible that jalv has it wrong too? | 17:05 |
| drobilla | I need to flesh out those docs. | 17:06 |
| drobilla | I don't think jalv even sends beat. | 17:06 |
| drobilla | (beat != barBeat) | 17:06 |
| falktx | yep | 17:07 |
| drobilla | Mostly derived from jack transport, but beat stuff became float to make ticks go away | 17:08 |
| falktx | I kinda want ticks though | 17:10 |
| drobilla | Why? | 17:15 |
| falktx | BBT = bar beat ticks, what's wrong with ticks? | 17:16 |
| drobilla | Ticks means choosing some ppqn. It seems objectively better to send a massively higher resolution float you can easily convert to a tick if you want | 17:16 |
| drobilla | You can consider it ticks with PPQN somewhere around 3 billion :) | 17:19 |
| wrl | just multiply the fractional part by some arbitrary ppqn to get ticks | 17:20 |
| falktx | I'm just following the jack design | 17:21 |
| falktx | it's easier when porting software | 17:21 |
| drobilla | Perhaps. The mishmash of things that start from both 1 and 0 in jack transport is silly IMO | 17:26 |
| drobilla | Oh well. "Easier" by one division. You'll live :P | 17:26 |
| *** rncbc has quit IRC | 18:12 | |
| *** mlpug has joined #lv2 | 18:17 | |
| *** mlpug has quit IRC | 19:48 | |
| *** mlpug has joined #lv2 | 19:54 | |
| *** edogawa is now known as CarlosSantana | 20:02 | |
| *** CarlosSantana is now known as edogawa | 20:03 | |
| *** mlpug has quit IRC | 20:18 | |
| *** NickSB2 has joined #lv2 | 21:33 | |
| *** kaspar_e has joined #lv2 | 21:34 | |
| * drobilla desperately tried to cram code into these 40 character columns | 21:57 | |
| drobilla | tries* | 21:57 |
| *** ssj72 has quit IRC | 22:04 | |
| *** ssj72 has joined #lv2 | 23:14 | |
| *** edogawa has quit IRC | 23:35 | |
Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!
