Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2017-12-11

00:01:14jhMikeSThis (old function selection code) actually makes it work again but doesn't seem like what you intend: https://pastebin.com/p3QzWJEQ
00:04:47Bilgusare you sure? I'm pretty sure it still wouldn't be null??
00:08:09jhMikeSwhat wouldn't be NULL? the function?
00:08:37Bilgusget_context_map
00:08:53jhMikeSnot following...
00:09:28Bilgusok so its looking for context & CONTEXT_PLUGIN && get_context_map != NULL
00:09:30jhMikeSthe internal ones don't set CONTEXT_PLUGIN
00:09:40jhMikeSyeah, two things
00:10:00Bilgusget context_map still isn't NULL
00:11:13Bilgusthe same thing happens when initializing cur_action.. at #862
00:11:18jhMikeSwhy would that matter? the flag is 0
00:13:17Bilguspictureflow passes back -> LAST_ITEM_IN_LIST__NEXTLIST(CONTEXT_PLUGIN|1)
00:15:38jhMikeSwhich sends it to pf_context_buttons[], right?
00:15:43[Saint]lol - I always get soo tense when I see you guys messing about with list handling.
00:15:53[Saint]you're upsetting my delicate ecosystem of bandaids and prayer!
00:16:19jhMikeS[Saint]: gotta mess up all the lists!
00:17:04Bilgusyes which has no end of list marker
00:17:05jhMikeSBilgus: then after it exhausts pf_context_buttons, it specifies CONTEXT_TREE, unless it's M3 of course
00:17:38Bilgusand thats just an index that isn't in pf_context
00:17:42[Saint]The icon display choices for some(most?) of the lists are just so arbitrary sometimes I ended up just going "fuck it".
00:17:49jhMikeSBilgus: why should it? as I read it that tells it to increment and try the next plugin context
00:18:07[Saint]In my icon set *everything* that's either a standalone setting or a setting with sublists just gets the same icon.
00:18:22[Saint]I couldn;t handle the inconsistency so I just nuked it from orbit with a coat of paint.
00:18:34Bilgusit needs to be in action.c ~#604 context = get_next_context(cur->items, i);
00:18:34Bilgus (+) cur->items = get_context_mapping(context);
00:19:10Bilgusor something along those lines instructing it to go back to the internal mapping
00:19:58Bilgusonly reason the that context tree one is working is because its never falling through to it
00:20:33Bilgusyou only noticed the issue on your device because it has def USE_CORE_PREVNEXT
00:21:20jhMikeSit's not defined on my device
00:21:37jhMikeSit's a gigabeat S so has no scrollwheel
00:21:42jhMikeSsame for the clip
00:24:37Bilgushmm
00:25:16jhMikeSthe last item just links the structs together: "try plugin context index 1".
00:26:22BilgusI just added it and compiled and it still crashes
00:26:42jhMikeSadded what?
00:27:03Bilgusthe diff you posted
00:27:30jhMikeSmine wasn't crashing
00:28:45BilgusO_o
00:29:31jhMikeSI'll try the clip now
00:31:09jhMikeSthat's fine with the diff too
00:32:08jhMikeSpictureflow doesn't regulate its speed very well
00:32:41BilgusI think I have a better option worked out brb
00:32:58 Quit pamaury (Remote host closed the connection)
00:38:34Bilgusah ok if you were to pass back CONTEXT_PLUGIN like with USE_CORE_PREVNEXT it would be back to crashing the only reason what you posted works is because its switching to the internal functions again
00:39:35Bilgusoh duh nevermind −−
00:39:39*Bilgus stupids
00:42:22 Quit petur (Remote host closed the connection)
00:47:51 Quit ender` (Quit: In a perfect world… spammers would get caught, go to jail, and share a cell with many men who have enlarged their penisses, taken Viagra and are looking for a new relationship.)
01:00
01:01:07 Quit krabador (Quit: Leaving)
01:09:16Bilgusok jhMikeS I got down to why I was so confused and what you changed does fix the issue you pointed out but not the issue I exposed that made the same issue happen but slightly different series of events ok so the reason your diff fixed the issue is that pictureflow passes in CTX_PLUGIN which allows it's button mapping to take the first table it encounters and use it and then switches to the internal mapping for any subsequent
01:10:43Bilgusbut if there were more than 1 table for it to look up like with USE_CORE_PREVNEXT then it still fails which is what I exposed
01:17:03 Quit michaelni (Ping timeout: 248 seconds)
01:17:34Bilgussee: https://github.com/Rockbox/rockbox/blob/03dd4b92be7dcd5c8ab06da3810887060e06abd5/apps/plugins/lib/pluginlib_actions.c
01:18:02***Saving seen data "./dancer.seen"
01:27:22jhMikeSthere were two tables to look at though (hence CONTEXT_PLUGIN|1 )
01:30:41 Join michaelni [0] (~michael@213-47-41-20.cable.dynamic.surfer.at)
01:31:15jhMikeSpictureflow doesn't use anything there
02:00
02:23:43 Quit Guest56968 (Ping timeout: 248 seconds)
02:24:02BilgusJhMikeS you were right I'm not sure why I was having such a hard time wrapping my head around it
02:28:25 Quit dys (Ping timeout: 240 seconds)
02:28:38Bilgusg#1762
02:28:41fs-bluebot_3Gerrit review #1762 at http://gerrit.rockbox.org/r/1762 : 3Fix error with action subsystem and custom context mapping by William Wilgus
03:00
03:04:13 Quit Moarc (Ping timeout: 248 seconds)
03:06:32 Join Moarc [0] (~chujko@a105.net128.okay.pl)
03:10:41jhMikeSBilgus: I was thinking you could just embed a pointer to the next context map or something in the element
03:13:12jhMikeSthen again that has its own nasty repercussions...
03:18:03***Saving seen data "./dancer.seen"
03:19:59 Quit this_is_a_nick (Remote host closed the connection)
03:29:00 Join Bilgus_ph [0] (4cf32773@gateway/web/freenode/ip.76.243.39.115)
03:29:57 Quit snw (Quit: User was destroyed by a weapon of mass destruction.)
03:30:37 Join snw [0] (~snow_bcks@ganon.dot-server.net)
03:34:32Bilgus_phjhMikeS: I think the only safe option besides checking the context + map like it was before is what is in pluginlib_actions, pictureflow is the only thing using get_custom_action besides it though
03:41:20 Quit Bilgus_ph (Quit: Page closed)
03:53:44jhMikeSBilgus: the callback is such a wart though
03:55:20jhMikeSOne question I have is if hwcodec is supposed to forked, expunged or whatever, why is it still in the build rounds so that you still have to tiptoe around it?
03:56:31 Join this_is_a_nick [0] (~amofiuhr_@ip60-237-64-186.ct.co.cr)
04:00
04:07:23BilgusNo one has moved towards removing it..
04:10:09jhMikeSit's a lot of moving if mr. someone has to do all the work at once
04:11:35Bilgusi think the better option is to fork it where it sits and then remove the targets
04:11:56Bilgusand as we go through we can remove the remaining conditionals
04:13:08jhMikeSI did that latter by hand once. Didn't actually take very long. After that, I got distracted away from what I was doing there.
04:16:20Bilguswell, no I suspect not its pretty nicely delineated..
04:33:52[Saint]in some places.
04:34:25[Saint]jhMikeS: sounds like my push for main menu tidying
04:34:48[Saint]I got about 90% of the way there and went "what the fuck am I doing with my life?" and rage quit the git repo.
04:36:24jhMikeSmain menu tidying?
04:37:07[Saint]Apparently I'm the only one that notices it or is bothered by it but the menus are a dumpster fire.
04:37:34[Saint]Things just seem haphazardly scattered all over the place.
04:37:51jhMikeSyou mean the code or the UI?
04:38:23[Saint]A little A, a little B.
04:39:06[Saint]Mostly the end result. The UI. How it's organised in the source doesn't matter /too/ much.
04:39:34jhMikeSwell....it does if you have to edit it :)
04:40:55jhMikeSI find it a bit more strewn about in the source. Most of the menu I never even look at. I think I opened the FF/RW playback menu for the first time about 2 weeks ago to see what it was.
04:42:50jhMikeSAnd "Skip Length" right after that. I just opened a few more I've never looked in. :)
04:47:18[Saint]Main menu:
04:47:18[Saint]...
04:47:18[Saint]Settings - System
04:47:18DBUGEnqueued KICK [Saint]
04:47:18[Saint]...
04:47:18[Saint]System
04:47:31[Saint]That bothers me to an unreasonable extent.
04:49:58Bilguswould renaming it to Debug help? :p
04:50:27BilgusI hate the menu system internally but its kinda the nature of the beast
04:51:43[Saint]I honestly think it would be more at home as something like "Settings - Info"
04:52:03jhMikeSThe first three aren't really debug stuff
04:52:08[Saint]At the very least there's no logical reason why it needs to be a toplevel menu entry.
04:52:37[Saint]no one has ever needed toplevel menu access to this menu structure ever.
04:53:07jhMikeSI don't mind that it is though. It is frequently visited. But the more useful screens are in shortcuts anyway.
04:53:39[Saint]frequently visited by whom?
04:53:44jhMikeSme
04:53:52[Saint]the dozen developers or the thousands of users?
04:54:01[Saint]rhetorical question BTW
04:54:14jhMikeSDebug (You'll Be Here A Lot)
04:54:42[Saint]For the overwhelming majority that's literally never oging to be true though.
04:54:46[Saint]Kinda my point.
04:55:11[Saint]I don't even think it should exist in the menu structure without a magic token.
04:55:21jhMikeSdeveloper mode?
04:55:43[Saint]We could easily make it a toplevel menu but hide it by default. Then jam the other three entires from System into Settings - Info
04:55:46jhMikeSjust add a config entry for it by hand
04:55:57jhMikeSno menu item
04:56:00[Saint]We already have the mechanism to hide toplevel menu entires and reposition them through config.cfg
04:56:19jhMikeSnever really looked into that
04:56:42[Saint]from memory it's the menu_order config param.
04:56:57[Saint]perhaps main_menu_order, we seem to like descriptive identifiers.
04:58:45[Saint]developers who _really_ wanted it there 24/7 could just add the change to the config in as fixed.cfg
04:59:06[Saint]I am still right that fixed.cfg is parsed last and overrides config.cfg am I not?
04:59:14[Saint]I haven't actually tested that in a while.
04:59:17jhMikeSGeneral Settings should be largely flatted right into Settings
04:59:23jhMikeS*flattened
04:59:27[Saint]Yes!
04:59:35[Saint]A man after my own heart.
05:00
05:00:02[Saint]So far that's two changes I've proposed that you've come up with independently.
05:00:13[Saint]Makes me think I'm not entirely insane, or we both are.
05:00:16[Saint]Either works for me.
05:01:40jhMikeSit's too many clicks deep to get into some items.
05:02:25jhMikeSsome clearly playback-related functionality is General Settings for some reason
05:02:36[Saint]Just changing repeat or shuffle is quite ridiculously complex if you're not aware of the Quickscreen.
05:02:46[Saint]Which in my experience seems to be about ~90% of users
05:02:57jhMikeSThat's only two deep to see it
05:03:39jhMikeSwouldn't one think hmmm..."sounds like a playback thing?"
05:04:18jhMikeSthough it has a certain urgency, almost as immediate as a volume control
05:05:54[Saint]It would be difficult to achieve and I'm aware it wouldn't be a popular opinion but I'm inclined to think Rockbox needs a "simple" menu setting.
05:06:30jhMikeSmaybe unlock achievements? :)
05:06:40[Saint]The Sound Settings menu is....well, it's something.
05:07:51CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
05:07:51*[Saint] eye twitches
05:08:10[Saint]Some of the child menus have capital letters and others don't.
05:08:16[Saint]I can't unsee it.
05:09:12[Saint]Settings - Theme Settings - Browse Themes, and
05:09:12[Saint]Settings - Sound Settings - Equalizer are ones I've found just now poking around.
05:09:29[Saint]Once you enter them their list title is lower cased.
05:09:57 Join _jhMikeS_ [0] (~jethead71@d192-24-173-177.try.wideopenwest.com)
05:09:57 Quit jhMikeS (Disconnected by services)
05:09:58 Nick _jhMikeS_ is now known as jhMikeS (~jethead71@d192-24-173-177.try.wideopenwest.com)
05:14:35jhMikeSat least it just lists independent effects in a straight menu for the most part
05:15:59jhMikeSone thing that I didn't put together right away was the connection between Channel Configuration and Stereo Width
05:17:16[Saint]There aren't many "self cancelling" menu combinations, but, yes - that's one of them.
05:17:53[Saint]stero width, 200%
05:17:53[Saint]channel configuration, mono
05:17:59[Saint]checkmate athiests.
05:18:07***Saving seen data "./dancer.seen"
05:19:35[Saint]I mean...how do you even handle cases like that with the way we handle menus?
05:19:40jhMikeSCrossfeed is similar now
05:20:03[Saint]There's no way to effictively tell the user something is happening if we "hand hold" and do it for them.
05:20:32[Saint]Setting a stereo width other than 100% should probably set the channel configuration to custom.
05:20:44[Saint]But we can't tell the user we did it or that it happened in a meaningful way.
05:20:57[Saint]So we just let them apply settings that might do literally nothing.
05:22:46[Saint]Hmmmm, there's a thought. Do we break EU law by not enabling a safe listening volume target by default?
05:24:00jhMikeSyou mean maximum volume setting achievable by default?
05:24:09[Saint]Yes.
05:24:48jhMikeSwell, heck, I don't know the letter of that law
05:26:39[Saint]Hmmmm, like all things EU it seems complex.
05:26:43[Saint]"None of the standards currently prescribe any maximum pressure limit nor require any specific labelling in respect of noise emissions but require that a statement be put in the instruction manual to warn against adverse effects of exposure to excessive sound pressure.:
05:26:49jhMikeSask the kind of Sweden if it's too loud
05:27:50[Saint]One thing I'm really not sure of but I /think/ I'm right about...
05:28:02[Saint]Our dB measure for volume is all just viscious lies, no?
05:28:55[Saint]It realy looks to me like it would be a shitload simpler to just scale everything to a nice clean positive percentile.
05:29:05jhMikeSno, it's dBFS, relative to the unity volume limit
05:32:26jhMikeSit's just that the 100% gain level corresponds to very different physical power levels depending on many things like the player and headphones
05:34:25[Saint]Maybe I get tripped up on the wording of it. Knowing it's dBFS makes a little more sense to me, but I don't really understand how we can then go into positive decibel ranges.
05:34:47[Saint]in dBFS (correct me if I'm wrong, please), isn't 0dB always going to be peak amplitude?
05:35:50jhMikeS+6 = amplify the signal by a factor of 2 (roughly). 20*log(2) ~= 6.0205999...
05:35:51[Saint]My understanding was that dBFS always had 0 or sub-zero values with 0 being peak amplitute.
05:37:19 Quit CR0W (Ping timeout: 248 seconds)
05:37:24jhMikeS-6 = amplify the signal by a factor of about 1/2. 20*log(0.5) ~= -6.0205999...
05:38:22jhMikeSso 0 means the amp puts out a full-scale signal at the limits of the amp. anything gain above that will cause clipping.
05:38:46[Saint]I need to redo my signal theorum modules again I guess.
05:38:47[Saint]I just seem to be categorically at odd with understanding this.
05:39:25[Saint]I don't even really understand the context of 'peak amplitude' here. Doesn't a properly dithered digitial signal havy infinite amplitude?
05:39:56[Saint]Errr, sorry for derailing - in University I just never had anyone who could put this to me in a way that I "got".
05:40:10[Saint]signal theory wasn't ever my bag.
05:42:41 Join CR0W [0] (narf@2001:0:53aa:64c:2cec:18dd:b306:4fe9)
05:42:41 Quit CR0W (Changing host)
05:42:41 Join CR0W [0] (narf@unaffiliated/em64t)
05:46:17jhMikeSummm....no if it's dithered you can lower the quantization floor arbitrarily, but it's like say you can put a tiny white dot on an arbitarily large black canvas to get aribrary dark gray levels
05:49:54 Join saratoga [0] (12650869@gateway/web/freenode/ip.18.101.8.105)
05:50:03saratoga0dBFS -> full scale wav file will play without clipping
05:50:46saratoga6dBFS - > half amplitude (-0.5 to 0.5) wav file will just barely play without clipping
05:50:55[Saint]in a model system only, presumably?
05:51:07saratogaassuming we calibrated correctly, in rockbox
05:51:35saratogai bet a few of the drivers are off by a dB or so though
05:51:52[Saint]in a log system that's a pretty wide berth.
05:52:47saratogathe definition of full scale often depends on the exact voltages you feed the amp anyway, so hardware tends to be conservative
05:53:02saratoganot a big deal if you waste a db of volume control, but it sounds bad if you clip
05:53:38jhMikeSsaratoga: I think it was the chip designers that had to calibrate
05:54:00saratogayeah if there is a datasheet
05:54:16saratogaAMS was literally just dfkt increasing the gain 1 dB at a time and measuring THD
05:54:24[Saint]I'm likely diving off the deep end of my understanding here but 0dBFS /can/ still clip, can it not?
05:54:30saratogaturns out we were off by quite a bit for a while
05:54:53saratogathe hardware won't clip it, but of course the audio file could have clipping already in it from bad mastering
05:55:15jhMikeS[Saint]: you can start to get distortion a bit before that even as the amp gets closer to its limts
05:56:47saratogaregardless of what the spec sheet says, I usually define 0dBFS as where ever the distortion starts to increase faster than amplitude
05:57:14saratogasince the exact value often depends on supply voltages, which in a DAP aren't always what the spec sheet says they should be anyway
05:58:13[Saint]right, that makes more sense to me.
05:58:25[Saint]the breakpoint between amplitude and distortion increases.
05:59:38 Join saratoga_ [0] (12650869@gateway/web/freenode/ip.18.101.8.105)
05:59:52[Saint]I think I've said it before but I really wish I had someone like you two who could actually use plain English to describe these principles instead of just waving at the math and hand-waving 'something something signal theory'.
06:00
06:00:49saratoga_(reading the logs) dither in audio is the same idea as dithering in images (add random noise to cause the value to switch between levels), but in terms of what it actually accomplishes, dither in images is more like oversampling, while dithering in audio doesn't really have an analog
06:00:52jhMikeSin the end though, the math is where it's at
06:01:31saratoga_what it really does is decorrelate the quantization error from the input, but what that actually means is just that noise gets both stronger and also much harder to hear (since it no longer tracks the input signal)
06:01:39[Saint]Its no use understanding the math, which I do, if you don't have some tangible concepts to align them with.
06:01:53[Saint]visual descriptors are pretty bloody important in education.
06:01:59 Quit saratoga (Ping timeout: 260 seconds)
06:02:19saratoga_for signals you pretty much have to start with the math, and then do it enough that it starts to make intuitive sense
06:02:48[Saint]And presumably not have a professor with a giant stick up his arse.
06:03:06saratoga_you'd have to be a genius to really grasp something like that intuitively without working it out a hundred times on pen and paper
06:03:48[Saint]Tenure is a bitch. Some educators just seem to check out.
06:04:16[Saint]I guess I was just unlucky and uninterested. Mostly the latter I suppose because no one bothered to make it interesting.
06:06:08saratoga_i was fortunate enough that mp3 players were just coming out when i was in college, so i was pretty motivated to figure out how to program one
06:06:09 Quit TheSeven (Ping timeout: 246 seconds)
06:07:14jhMikeSto make something interesting, have an interest in it. the rest just sort of does itself.
06:14:33 Quit CR0W (Ping timeout: 246 seconds)
06:16:47[Saint]I think I fixate too much on semantics.
06:18:02[Saint]0dB being "the point at which we won't clip" seems at odds with interpolation.
06:19:03[Saint]IIUC, it very well could clip even without a singlular sample at 0dB.
06:19:13jhMikeSindeed
06:19:32jhMikeSsince it can overshoot the known sample values
06:19:47saratoga_yeah, the intersample over problem
06:19:56 Join CR0W [0] (~narf@adsl-76-249-176-22.dsl.milwwi.sbcglobal.net)
06:19:56 Quit CR0W (Changing host)
06:19:56 Join CR0W [0] (~narf@unaffiliated/em64t)
06:20:04saratoga_but again, if you define 0dB as the point where you don't have distortion, you automatically handle those correctly
06:20:13[Saint]I feel like I made a little breakthrough there.
06:20:13jhMikeSbut the real value shouldn't be a massive outlier if there's band limiting
06:21:58jhMikeSneeds moar jpeg
06:29:50 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:38:19 Quit saratoga_ (Quit: Page closed)
06:39:26 Quit TheSeven (Ping timeout: 255 seconds)
07:00
07:00:20 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
07:04:01 Quit CR0W (Ping timeout: 240 seconds)
07:10:07 Join CR0W [0] (~narf@adsl-76-249-176-22.dsl.milwwi.sbcglobal.net)
07:10:07 Quit CR0W (Changing host)
07:10:07 Join CR0W [0] (~narf@unaffiliated/em64t)
07:16:43 Quit PurlingNayuki (Ping timeout: 250 seconds)
07:18:08***Saving seen data "./dancer.seen"
07:23:53 Join PurlingNayuki [0] (~Thunderbi@2001:da8:215:6905:192:ac4a:f78a:121e)
07:47:08 Join advcomp2019_ [0] (~advcomp20@63-229-188-192.sxct.qwest.net)
07:47:08 Quit advcomp2019_ (Changing host)
07:47:08 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
07:49:50 Quit advcomp2019 (Ping timeout: 268 seconds)
07:56:22 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:a44b:4e64:1e1d:5416)
08:00
08:10:59 Join deevious [0] (~Thunderbi@193.226.142.214)
08:12:06 Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
08:16:31 Quit jhMikeS (Ping timeout: 240 seconds)
08:17:29 Join ender` [0] (krneki@foo.eternallybored.org)
08:27:03 Quit PurlingNayuki (Ping timeout: 258 seconds)
08:47:51 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
09:00
09:08:07 Join ernestask [0] (~ernestask@78-56-62-157.static.zebra.lt)
09:10:18 Join dys [0] (~dys@tmo-123-54.customers.d1-online.com)
09:10:54 Join JannF [0] (~jann@HSI-KBW-046-005-019-085.hsi8.kabel-badenwuerttemberg.de)
09:13:56 Join petur [0] (~petur@91.183.48.77)
09:13:56 Quit petur (Changing host)
09:13:56 Join petur [0] (~petur@rockbox/developer/petur)
09:18:11***Saving seen data "./dancer.seen"
09:21:26 Quit JannF (Ping timeout: 255 seconds)
09:57:50 Quit dys (Ping timeout: 246 seconds)
10:00
10:09:38 Join dys [0] (~dys@2003:5b:203b:100:6af7:28ff:fe06:801)
10:13:03 Quit _meg (Ping timeout: 248 seconds)
10:14:41 Join _meg [0] (~notsure@211.25.203.45)
10:14:55 Join JannF [0] (~jann@HSI-KBW-109-192-015-103.hsi6.kabel-badenwuerttemberg.de)
10:23:23 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:35:06 Join JannF1 [0] (~jann@HSI-KBW-109-192-015-103.hsi6.kabel-badenwuerttemberg.de)
10:39:04 Quit JannF (Ping timeout: 240 seconds)
10:40:42 Join jhMikeS [0] (~jethead71@d192-24-173-177.try.wideopenwest.com)
10:45:02 Quit jhMikeS (Ping timeout: 248 seconds)
10:51:49copper[Saint]: needs moar NwAvGuy http://nwavguy.blogspot.com/2011/09/noise-dynamic-range.html
10:51:52copperremember him?
10:51:58copperhttps://spectrum.ieee.org/tech-history/silicon-revolution/nwavguy-the-audio-genius-who-vanished
10:53:04 Quit pamaury (Ping timeout: 248 seconds)
10:53:38copperI learned the basics from reading his blog
11:00
11:07:00 Join Guest56968 [0] (~ac@nat-eduroam-76-gw-01-lne.lille.inria.fr)
11:18:15***Saving seen data "./dancer.seen"
11:42:43 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:54:26 Quit Guest56968 (Quit: WeeChat 1.9.1)
12:00
12:06:42wodzpamaury: Great discovery! MIPS compatibility Ingenic way :P
12:11:58 Quit paulk-gagarine (Quit: Leaving)
12:13:08 Join paulk-gagarine [0] (~paulk-gag@gagarine.paulk.fr)
12:43:40 Quit CommunistWitchDr (Ping timeout: 240 seconds)
12:59:02pamaurywodz: yeah, it's seriously fucked up but now it works
13:00
13:01:15pamauryby the way, when doing all this experimentation, I ended up building a mip "library" that uses as macros and tries to hides some MIPS details, I'll push it to gerrit tonight, feel free to have a look
13:01:34pamauryit would be good if we can avoid duplicating a lot of code in crt0 by having macros there
13:01:43wodzpamaury: I am stack with atj as codec crashes early. It looks like something is overwriting codec buffer but I can't find out what. I am thinking of employing MMU to protect beginning of codec buffer just after codec load. This should allow to catch the offender.
13:01:53wodzpamaury: This is however quite a bit of work
13:02:23wodzpamaury: sure will look (but not today most probably)
13:04:13 Quit atsampson (Ping timeout: 276 seconds)
13:11:26pamaurywodz: yeah if you just want to protect a small part of the buffer, it's a bit tricky because the wired entries of the TLB might not be enough
13:13:18wodzpamaury: Well I need to protect only .text and .rodata part. I wonder if dma overwriting could be catched as well
13:18:18***Saving seen data "./dancer.seen"
13:21:31pamaurydma will be more difficult, you could always have some wrapper around dma that checks the destination
13:29:56pamauryat some point I think thinking about implementing proper MMU protection for .text, .data and buflib (ie invalidating previous virtual addresses when a buffer moves), but it's a lot of work
13:30:02pamaury* was thinking
13:37:26 Quit petur (Quit: Connection reset by beer)
13:47:31 Quit MrBismuth (Read error: Connection reset by peer)
13:56:00 Quit robertd11 (Ping timeout: 248 seconds)
14:00
14:01:40 Join robertd1 [0] (~root@186-88-142-98.genericrev.cantv.net)
14:04:09 Quit robertd1 (Client Quit)
14:04:18 Join robertd11 [0] (~root@186-88-142-98.genericrev.cantv.net)
14:07:06 Quit robertd11 (Client Quit)
14:08:29 Quit this_is_a_nick (*.net *.split)
14:08:29 Quit ChanServ (*.net *.split)
14:11:10 Join this_is_a_nick [0] (~amofiuhr_@ip60-237-64-186.ct.co.cr)
14:16:10 Join ChanServ [0] (ChanServ@services.)
14:16:10Mode"#rockbox +o ChanServ " by orwell.freenode.net
14:21:05 Join robertd1 [0] (~root@186-88-142-98.genericrev.cantv.net)
14:23:41 Quit robertd1 (Client Quit)
14:24:28 Join robertd11 [0] (~root@186-88-142-98.genericrev.cantv.net)
14:38:10 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:a44b:4e64:1e1d:5416)
14:46:32 Join amayer [0] (~amayer@107-1-97-172-ip-static.hfc.comcastbusiness.net)
15:00
15:10:48 Join LinusN [0] (~linus@giant.haxx.se)
15:18:21***Saving seen data "./dancer.seen"
15:23:04 Quit robertd11 (Ping timeout: 268 seconds)
15:26:38 Quit wodz (Ping timeout: 248 seconds)
15:38:30 Join storm_dragon [0] (~Storm@137.118.191.109)
15:38:56 Part LinusN
15:39:03 Part storm_dragon ("With a flash of lightning and a blast of air from massive wings, the storm dragon takes flight and vanishes into the fading twilight.")
15:55:52 Join Natter [0] (d15d1d67@gateway/web/freenode/ip.209.93.29.103)
15:56:53 Quit Natter (Client Quit)
16:00
16:17:25 Join jhMikeS [0] (~jethead71@d192-24-173-177.try.wideopenwest.com)
16:34:19 Quit _meg (Ping timeout: 255 seconds)
16:35:22 Join _meg [0] (~notsure@211.25.203.45)
16:38:38 Quit this_is_a_nick (Read error: Connection reset by peer)
16:44:52 Join this_is_a_nick [0] (~amofiuhr_@ip60-237-64-186.ct.co.cr)
17:00
17:02:34 Join SovietShaman [0] (quasselcor@97-87-177-85.dhcp.stls.mo.charter.com)
17:02:43 Quit this_is_a_nick (Read error: Connection reset by peer)
17:17:00 Quit pamaury (Ping timeout: 240 seconds)
17:18:10 Join krabador [0] (~krabador@unaffiliated/krabador)
17:18:23***Saving seen data "./dancer.seen"
17:47:54 Join atsampson [0] (~ats@cartman.offog.org)
18:00
18:34:32 Join pratch [0] (d15d1d67@gateway/web/freenode/ip.209.93.29.103)
18:35:02pratchHi, I am unable to charge my ipod classic via a wall usb.
18:35:26pratchAny advise to this?
18:48:18 Quit dys (Ping timeout: 250 seconds)
18:49:38 Quit pratch (Ping timeout: 260 seconds)
18:55:10 Quit maraz (Ping timeout: 248 seconds)
18:57:49 Join maraz [0] (maraz@kapsi.fi)
19:00
19:02:06 Quit JannF1 (Ping timeout: 248 seconds)
19:10:43 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
19:17:46 Quit advcomp2019_ (Ping timeout: 250 seconds)
19:18:14 Join advcomp2019 [0] (~advcomp20@65-131-145-69.sxct.qwest.net)
19:18:14 Quit advcomp2019 (Changing host)
19:18:14 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
19:18:24***Saving seen data "./dancer.seen"
19:21:40 Join smoke_fumus [0] (~smoke_fum@188.35.176.90)
19:27:35 Join lebellium [0] (~hexchat@89-93-177-206.hfc.dyn.abo.bbox.fr)
19:51:20 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:52:59 Join this_is_a_nick [0] (~amofiuhr_@201.199.104.169)
19:53:58 Quit advcomp2019 (Read error: Connection reset by peer)
19:58:22 Join advcomp2019 [0] (~advcomp20@65-131-179-3.sxct.qwest.net)
19:58:22 Quit advcomp2019 (Changing host)
19:58:22 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
20:00
20:29:50 Quit ernestask (Quit: ernestask)
20:30:54 Quit ender` (Read error: Connection reset by peer)
20:31:47 Join ender` [0] (krneki@foo.eternallybored.org)
20:43:21 Join JannF [0] (~jann@2a02:8071:198:9f00:5ee0:c5ff:febe:60e0)
20:54:50 Quit prg318 (Ping timeout: 250 seconds)
20:57:27 Join prg318 [0] (~prg@deadcodersociety/prg318)
20:57:56 Quit JannF (Ping timeout: 246 seconds)
21:00
21:06:38 Join amofiuhr_ [0] (~amofiuhr_@201.199.104.4)
21:07:09 Quit this_is_a_nick (Ping timeout: 272 seconds)
21:07:45 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:7d20:7795:1d8b:94f7)
21:14:33 Join dys [0] (~dys@tmo-098-204.customers.d1-online.com)
21:16:35 Join johnb2 [0] (~johnb2@p5B3AFF83.dip0.t-ipconnect.de)
21:18:20 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:a44b:4e64:1e1d:5416)
21:18:28***Saving seen data "./dancer.seen"
21:27:04 Quit duncan^ (Changing host)
21:27:04 Join duncan^ [0] (UNKNOWN@fez.tardis.ed.ac.uk)
21:27:43 Quit duncan^ (Quit: WeeChat 1.9.1)
21:29:30 Join duncan^ [0] (UNKNOWN@fez.tardis.ed.ac.uk)
21:58:05 Quit johnb2 (Ping timeout: 240 seconds)
22:00
22:07:05 Quit krabador (Remote host closed the connection)
22:08:39 Join krabador [0] (~krabador@unaffiliated/krabador)
22:11:32 Join JannF [0] (~jann@HSI-KBW-046-005-019-085.hsi8.kabel-badenwuerttemberg.de)
22:29:05 Quit JannF (Ping timeout: 240 seconds)
22:35:30pamauryGNU linker script never stop to amaze me:
22:35:30pamaury_relocstart = . ; // -> yields 0xf0000000
22:35:30pamaury_relocstart = (. & 0x3fffffff) | 0x80000000; // yields 0x700000
22:35:30DBUGEnqueued KICK pamaury
22:35:30pamauryunless I got my math wrong, this is not what I expected!
22:39:04pamauryunless the gnu *dis*assembler produces the wrong output and the actual code is correct, damn something is very wrong
22:42:21 Quit amayer (Quit: Leaving)
22:46:59 Quit scorche (Disconnected by services)
22:47:02 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
22:51:32 Join n17ikh_ [0] (~n17ikh@unaffiliated/n17ikh)
22:56:25 Quit shrizza (*.net *.split)
22:56:26 Quit n17ikh (*.net *.split)
22:56:43 Quit dan- (*.net *.split)
22:56:43 Quit rogeliodh (*.net *.split)
22:56:43 Join rogeliodh [0] (~rogeliodh@ec2-52-90-241-120.compute-1.amazonaws.com)
23:00
23:00:03 Join shrizza [0] (~shrizza@oki-dc-urasoe01-187.glbb.ne.jp)
23:02:03 Join dan- [0] (~d@101.165.131.18)
23:02:03 Quit dan- (Changing host)
23:02:03 Join dan- [0] (~d@freenode/corporate-sponsor/privateinternetaccess.com/doaks)
23:02:38jhMikeSdysassembler
23:09:41 Quit lebellium (Quit: Leaving)
23:10:01 Nick SovietShaman is now known as CommunistWitchDr (quasselcor@97-87-177-85.dhcp.stls.mo.charter.com)
23:14:56pamauryok so linker is definitely doing broken math, although it's so poorly documented that I'm sure it's correct in some weird sense
23:15:39pamauryalso what is section .MIPS.abiflags and why does it randomly pops up depending on somehow unrelated changes in the linker script ?!
23:18:32***Saving seen data "./dancer.seen"
23:31:03 Join petur [0] (~petur@rockbox/developer/petur)
23:32:49 Quit amofiuhr_ (Ping timeout: 272 seconds)
23:42:06 Quit ZincAlloy (Quit: Leaving.)

Previous day | Next day