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).

#rockbox log for 2021-03-19

00:24:27***Saving seen data "./dancer.seen"
02:07:39 Quit SovietShaman (Quit: Fuck this, I'm out)
02:08:28 Join SovietShaman [0] (
02:24:28***Saving seen data "./dancer.seen"
02:46:31 Quit SovietShaman (Quit: Fuck this, I'm out)
02:48:11 Join CommunistWitchDr [0] (
04:24:31***Saving seen data "./dancer.seen"
05:23:03 Join eevan [0] (~eevan@
05:23:43 Quit eevan (Client Quit)
05:24:23 Join eevan [0] (~eevan@
05:41:22 Part eevan ("Leaving")
05:43:37 Join eevan [0] (~eevan@
06:05:36 Quit eevan (Quit: Leaving)
06:06:38 Join eevan [0] (~eevan@
06:24:32***Saving seen data "./dancer.seen"
06:27:19 Join jdarnley [0] (
06:28:56 Quit J_Darnley (Ping timeout: 240 seconds)
06:50:38 Quit Barlow (Ping timeout: 260 seconds)
07:27:22 Join Barlow [0] (
07:27:22 Quit Barlow (Changing host)
07:27:22 Join Barlow [0] (~barlow@unaffiliated/barlow)
07:48:12 Join massiveH [0] (
08:24:34***Saving seen data "./dancer.seen"
08:33:29_bilgus_ speach y re bad software makes great communities?
08:35:37_bilgus_I think rockbox is daunting at that first look at the source code but at least we have guides (even if somewhat out of date)
08:36:57_bilgus_Ithink its more just making it accessible and known and a look at gerrit lately says it or maybe we just broke enough stuff
08:40:48 Quit tchan1 (Quit: WeeChat 3.0.1)
08:41:07 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
08:46:48 Join ac_laptop [0] (~ac_laptop@
08:52:39 Quit massiveH (Quit: Leaving)
09:00:08 Join Saijin_Naib [0] (
09:19:23 Quit spectrumanalyst (Quit: Bye!)
09:22:40 Join spectrumanalyst [0] (~admin@
09:25:27 Join tchan1 [0] (
09:26:05 Quit tchan1 (Client Quit)
09:28:40 Join advcomp2019__ [0] (
09:28:40 Quit advcomp2019__ (Changing host)
09:28:40 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
09:29:20 Join J_Darnley [0] (
09:33:52 Quit tchan (*.net *.split)
09:33:52 Quit jdarnley (*.net *.split)
09:33:53 Quit michaelni (*.net *.split)
09:33:53 Quit advcomp2019_ (*.net *.split)
09:33:53 Quit Galois (*.net *.split)
09:37:28 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
09:42:02 Join michaelni [0] (
10:24:38***Saving seen data "./dancer.seen"
10:52:32 Quit ufdm (Remote host closed the connection)
10:58:37 Join ufdm [0] (
11:23:53 Join Galois [0] (
11:26:28 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
12:19:43 Quit Acou_Bass (Ping timeout: 260 seconds)
12:24:41***Saving seen data "./dancer.seen"
12:33:14 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
12:59:46 Join lebellium [0] (
13:15:20 Quit Saijin_Naib (Disconnected by services)
13:15:24 Join Saijin-Naib [0] (
13:37:46 Quit asaba (Quit: Relay server offline)
13:37:57 Join asaba [0] (~asaba@
14:12:50 Join advcomp2019 [0] (
14:12:50 Quit advcomp2019 (Changing host)
14:12:50 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
14:13:47 Join advcomp2019_ [0] (
14:13:47 Quit advcomp2019_ (Changing host)
14:13:47 Join advcomp2019_ [0] (~advcomp20@unaffiliated/advcomp2019)
14:16:09 Quit advcomp2019__ (Ping timeout: 264 seconds)
14:17:12 Quit advcomp2019 (Ping timeout: 246 seconds)
14:24:44***Saving seen data "./dancer.seen"
14:31:02__builtin"It has many subscribers and quite intense traffic at times." -
14:32:57 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
14:33:19 Quit ufdm (Read error: Connection reset by peer)
14:34:17braewoodsthere's a mailing list?
14:34:31braewoodsi think the forums have largely replaced them
14:39:12 Join ufdm [0] (
14:58:04 Join jdarnley [0] (
14:59:06 Quit J_Darnley (Ping timeout: 256 seconds)
15:18:42 Join MrZeus [0] (~MrZeus@
15:49:48 Quit eevan (Read error: Connection reset by peer)
16:12:58 Join petur [0] (~petur@rockbox/developer/petur)
16:24:48***Saving seen data "./dancer.seen"
16:41:20 Quit Saijin-Naib (Read error: Connection reset by peer)
17:00:09 Join Saijin_Naib [0] (
17:04:44_bilgus_We 'try' to run big changes through there
17:05:17_bilgus_a lot more people subscribed there than actually look at the forum I think
17:57:44 Quit Saijin_Naib (Disconnected by services)
17:57:48 Join Saijin-Naib [0] (
18:10:03 Join advcomp2019__ [0] (
18:10:03 Quit advcomp2019__ (Changing host)
18:10:03 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
18:10:10 Join lebellium_ [0] (
18:10:10 Join fmlatghor [0] (~lcoogan@2601:5cd:8100:2890:9220:3aff:fe1a:350d)
18:10:52 Join mendelmunkis [0] (
18:11:36 Join rogeliodh9 [0] (
18:12:07 Quit rogeliodh (Read error: Connection reset by peer)
18:12:07 Nick rogeliodh9 is now known as rogeliodh (
18:12:56 Quit mendel_munkis_ (Read error: Connection reset by peer)
18:13:06 Quit advcomp2019_ (Ping timeout: 246 seconds)
18:13:06 Quit lebellium (Ping timeout: 246 seconds)
18:13:45 Quit benedikt93 (Ping timeout: 264 seconds)
18:14:22 Join benedikt93 [0] (~quassel@unaffiliated/benedikt93)
18:15:33 Quit Huntereb (Ping timeout: 246 seconds)
18:16:09 Quit Galois (Ping timeout: 264 seconds)
18:16:28 Join Huntereb [0] (~Huntereb@2603:9001:5a04:bbe7:21e:37ff:fe4e:f43a)
18:24:49***Saving seen data "./dancer.seen"
18:58:13 Join Galois [0] (
18:59:29 Quit Galois (Remote host closed the connection)
19:00:17 Join Galois [0] (
19:22:09 Join Acou_Bass [0] (
19:26:00 Quit petur (Remote host closed the connection)
19:55:10 Quit Acou_Bass (Quit: ZNC 1.8.2 -
19:58:20 Join Acou_Bass [0] (
20:00:40 Join dconrad [0] (d026e411@
20:23:46dconradI'm looking at the softlock/backlight code in action.c, what's this unlock_combo about? It looks like it's set in do_auto_softlock() to whatever the last button pressed was? What's this intended to do? I don't quite understand
20:24:20dconradas far as I can tell, if the autolock times out and locks, you can hit whatever the last button you pressed was to undo the lock, is that it?
20:24:52***Saving seen data "./dancer.seen"
20:38:25dconraddoesn't seem to work on my Hifi walker H2, so that must not be it, or it doesn't apply to this device
20:42:24 Quit lebellium_ (Quit: Leaving)
20:47:02 Join massiveH [0] (
21:26:42 Quit fmlatghor (Quit: WeeChat 3.0.1)
21:34:13 Quit MrZeus (Ping timeout: 245 seconds)
21:58:43_bilgus_what it does is stores the last button pressed, since we can just scan for that button to unlock
21:59:06_bilgus_otherwiseyou'd have to look for the completed action
22:00:30_bilgus_IIRC it does first lock, first unlock, then auto lock
22:00:49_bilgus_and once in auto lock it will lock on backlight off
22:01:06_bilgus_if lock and unlock works that combo is being used
22:02:46dconradso the unlock_combo is used for all lock and unlock?
22:03:38_bilgus_so the action system works on top of the button system
22:04:08_bilgus_it implements the repeat/release signals as well
22:04:43_bilgus_but we try to shut down the action sys whil screen locked to save power
22:05:27_bilgus_so we store that last button and assume that is the button used to lock
22:06:52dconradI think I understand
22:08:00_bilgus_I can explain more if you like what are you looking to do?
22:08:14dconradso for instance in do_softlock(), the check if ((action is ACTION_STD_KEYLOCK) or (keys are locked and current button is the unlock combo)
22:08:53_bilgus_well the action is usually several keys or a key and a release or repeat
22:08:54dconradthe action_std_keylock is really for when the backlight is currently on, and the other condition is really for when the backlight is off? Or the same but locked/unlocked
22:09:39_bilgus_we just store the last button that was a part of that lock action
22:09:56_bilgus_well the back light is only really used here fr the auto lock
22:10:12_bilgus_you don't need the back light to do any lock or unlock
22:10:18dconradI see
22:10:32_bilgus_it just uses that screen off notice to know to auto lock
22:10:47_bilgus_now selective key lock and backlight do a bit different
22:10:51dconradwell I submitted this on gerrit:
22:10:59_bilgus_same mech
22:11:14dconradand I wasn't sure if my check is sufficient or it should do something with the unlock_combo
22:11:28_bilgus_which number?
22:11:38dconradit works on my device, but I'm sure it won't on some different one
22:12:02_bilgus_ g#3240
22:12:04fs-bluebot_Gerrit review #3240 at : If backlight is off and keys are unlocked, power button wakes screen only by Dana Conrad
22:12:23_bilgus_if the screen was off it shouldn't do that
22:12:33_bilgus_let me check it might be a keymap issue
22:13:19 Quit Xeha (Ping timeout: 276 seconds)
22:13:49_bilgus_what device?
22:13:59dconradhifi walker h2 / eros q
22:14:11dconrada hibylinux device, I think
22:14:30dconradso in that state, the action system is supposed to not even be running?
22:16:17_bilgus_oh it should be runningbut if there is an action mapping to the same as the ky lock then the action system can decide what to do
22:19:57_bilgus_hmm well we can't do a repeat event on the power button
22:20:10_bilgus_hmm let me see how I solved this the last time
22:20:46dconradso my fix shouldn't work, yet it does?
22:22:39_bilgus_oh of course it does but actually there is a setting called first kypress activates backlight
22:24:09_bilgus_settings>general>display>lcd>first button press
22:24:20dconradI thought I had that enabled, and it still did it
22:24:36dconradlet me test it real quick
22:24:40_bilgus_for sure shouldn't have
22:24:53***Saving seen data "./dancer.seen"
22:27:31 Join Xeha [0] (~Xeha@unaffiliated/xeha)
22:28:06_bilgus_with people wanting to place keylock in the main screen your patch prob has merit
22:28:51dconradyeah, I built the master branch real quick and even with "first buttonpress turns on backlight only" turned on, it locks the keys when turning the LCD back on
22:30:39_bilgus_can you try adding this in the keymap?
22:30:53dconradthe first buttonpress feature works for all other keys, but not the lock/pwr button
22:32:38dconradadding what? (was a file supposed to show up or something?)
22:33:00_bilgus_sorry had to type it out
22:33:33dconradoh, instead of BUTTON_NONE as the previous button?
22:33:47dconradI'll give it a shot
22:33:48_bilgus_then the other try BUTTON_POWER|BUTTON_REL, BUTTON_POWER },
22:33:56_bilgus_same pre part
22:35:55_bilgus_if that doesn't work then the patch will do but i needs to be pushed into the current logic if ((action == ACTION_STD_KEYLOCK)
22:35:55_bilgus_ || (last->keys_locked && last->unlock_combo == cur->button))
22:37:29_bilgus_so add brackets and check if ((!last->keys_locked) && (!is_backlight_on(false)))
22:37:54_bilgus_and don't forget the backlight define
22:40:31dconradyeah, neither keymap change makes the behavior change
22:40:52dconradI tried both states of first keypress as well for each change
22:41:33dconradI can pretty easily move the check into the preexisting if statement
22:41:54_bilgus_yes that way its not increasing the amount of code in the hot path
22:42:06_bilgus_also add tour name to credits
22:42:17_bilgus_i'll be back in a it
22:42:23dconradgotcha, thanks
23:02:33 Quit ac_laptop (Ping timeout: 246 seconds)
23:05:13_bilgus_ok I know why it does it Button power is special cased
23:05:21_bilgus_so yeah proceed on
23:06:07_bilgus_I did that with the original patch because it appears as if the player is hung if no buttons bring up the backlight
23:06:50dconradthat makes sense
23:07:07dconradI uploaded a revision there
23:07:25_bilgus_that whole adventure was an exercise in breaking stuff
23:08:11dconradhah, well it got there in the end
23:08:38dconradsome of this code has me chasing my tail trying to figure out how it works tbh
23:08:57dconradI'm starting to "get it"
23:09:27_bilgus_already in the credits?
23:09:46_bilgus_ok looks good to me we'll push it
23:09:55dconradsick, thanks
23:10:54fs-bluebot_Build Server message: New build round started. Revision d5676fcd90, 293 builds, 9 clients.
23:11:16_bilgus_ok stick round long enough to make sure its still geen and all will be well
23:15:09_bilgus_bout 10-12 mins make sure it didn't break anything else overtly
23:15:50dconradyeah, I thought it would be near instantaneous but I suppose when you're building for this many targets...
23:16:10dconradoh, quick question while I've got your ear
23:16:29_bilgus_I did it on my own machine for the selective backlight and softlock stuff since it was so easy to break half the targets
23:16:57_bilgus_but it took like 2 hours
23:17:39dconradwhen you need to revise a gerrit patch and you run the command it gives you to download into detached HEAD state,
23:17:47_bilgus_go head actually always ask questions in chan and someone will answer if they know (eventually)
23:18:00dconraddoes it matter what state your current git tree looks like?
23:18:34dconradlike, if you branch off for a fix and then upload and need to revise, should you checkout master, or leave it in your branch?
23:18:39_bilgus_depends how you pull it back n\let me grab my cheat sheet
23:19:04dconradI thought I was good at git and then this threw a wrench in my understanding, hah
23:19:18_bilgus_if you upload but its not pushed then just continue on that branch and commit −−amend
23:19:34dconradwithout need to download it again?
23:19:50_bilgus_if it got pushed then co master then git pull −−rebase and do a new branch
23:20:06dconradok, yeah I suppose that makes sense
23:20:40_bilgus_yeah just if you mess up and do a commit −−amend on a new branch i pulls you back to the previous commit
23:20:55_bilgus_so in that case delete everything and it will cancel
23:21:01 Join f1reflyylmao [0] (
23:21:06_bilgus_ie all text in the commit
23:21:34dconradoh, huh so that's how you cancel a commit
23:22:02_bilgus_now pulling back stuff in your branch I'd recommend grabbing git-review
23:22:16_bilgus_then you can use the gerrit id
23:22:44 Quit cockroach (Quit: leaving)
23:23:12 Quit f1refly (Ping timeout: 246 seconds)
23:23:12 Nick f1reflyylmao is now known as f1refly (
23:23:32dconradbut a branch should basically be a single commit, because that's all that gets submitted to gerrit, right?
23:23:40dconradthe most recent commit
23:24:05dconrador the whole chain will
23:24:16_bilgus_well I sometimes do a branch either mine or someone elses then do a checkout -b newbranch and that makes it a child of the first
23:24:38_bilgus_then in gerrit it will tie them together
23:24:48_bilgus_so they go lock step on commit
23:25:12_bilgus_or even several off that new 'master' commit
23:25:41dconradgod git can get complicated quick
23:25:56_bilgus_typically people like incremental changes but sometimes its just abig glob of ugly no matter what
23:26:28_bilgus_but one thing we try not to do is have whitespace changes mixed in with code
23:26:44_bilgus_split that into a different commit so is easy to follow
23:27:03_bilgus_then do a child off that with your code
23:27:38_bilgus_also when rebasing gerrit can't alwasy do the stuff git can do
23:27:48fs-bluebot_Build Server message: Build round completed after 1014 seconds.
23:27:50fs-bluebot_Build Server message: Revision d5676fcd90 result: All green
23:27:55dconradI'm not sure I follow the whitespace thing
23:28:10dconradso like you're on master, make a new branch
23:28:29_bilgus_so you run into a file with spaces at EOL or tabs instead of 4 spaces
23:28:49_bilgus_and you want to change that in addition to code
23:29:38_bilgus_so when we review we have to wade through the 50 lines that nolonger end in space and its hard to find the 6 lines of code that were chnagse
23:29:45_bilgus_so instead starting at mastwer
23:29:50_bilgus_git pull −−rebase
23:30:00_bilgus_git checkout-b branchws
23:31:01_bilgus_do your whitespace changes git commit - whitespace changes bleh ->
23:31:45_bilgus_push head ref for master
23:32:00_bilgus_then checkout -b branch code
23:32:07_bilgus_rinse repeat
23:32:23_bilgus_thenj once in gerrit it will subit both
23:32:29_bilgus_when pushed..
23:33:23dconradso the "real" branch is a branch off the fake whitespace branch, and they get pushed separately, but they're somehow tied together still?
23:34:09_bilgus_you could always just push to master do the compile
23:35:14_bilgus_rebase then do code but instead you do it as a ws change commit and a child (or after the first one is submitted if you prefer) then the second one is applied on top
23:35:41dconradhuh, ok
23:35:55_bilgus_in gerrit it will be like 'hey you have a second commit here would you like to push them together?'
23:36:33_bilgus_or you know just don't do a bunch of unrelated ws changes
23:36:48dconradok, so gerrit is smart enough to know when one open commit follows another
23:36:51_bilgus_but thas not the only place it appliues
23:36:53dconradalso, yeah gotcha
23:37:22_bilgus_its comes into play when you have really big shit that touches a bunch of stuff
23:37:52_bilgus_then you can do a bit test and move onto a new branch to keep it organized
23:39:22_bilgus_but git sucks sometimes so i've had to hand rebuild trees when people change things in the middle of your patch -> child -> child chain so try to get them in gerrit to protect your work
23:39:35_bilgus_only an ass would push on a conflict thats WIP
23:41:45_bilgus_actually git is the thing that marks the parent child relations
23:41:57_bilgus_but gerrit picks it up yeah
23:42:34_bilgus_you can also do weird voodoo with rebasing ontop of new parents and it sometimes works
23:42:48_bilgus_(in gerrit)
23:54:25 Join ac_laptop [0] (~ac_laptop@

Previous day | Next day