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 2017-12-01

00:06:46jhMikeSit's not since it's triggering DMA on demand
00:07:05jhMikeSI also don't get why it has to worry about alignment if it's just copy in memory
00:07:57jhMikeS....I suppose for the dma on demand
00:08:15jhMikeSbut that should just be going and refreshing the screen continually
00:08:28jhMikeSwhile it's on anyway
00:09:34jhMikeSthe LCD looks set up that way but then lcd_update_rect fiddles with it
00:11:02_Bilgusreally the datasheet only covers LCDIF for the chip I didn't even consider the LCD controller may have its own limitations but I wouldn't have expected that to hang the device
00:12:36jhMikeSthe controller is definitely some Renesas model. the e200v1 has a renesas controller and a continuous DMA auto refresh
00:16:46jhMikeScan't say "definitely". it just has that look. more registers than e200's
00:24:42_BilgusNote2. Data are written to GRAM in four-words when operating in high speed mode, the dummy write
00:24:42_Bilgusoperations should be inserted depending on the window address area. For details, see the High-Speed RAM
00:24:42_BilgusWrite Function section.
00:28:21jhMikeSIf it's that ilitek thing, the datasheet is a dead ringer for a renesas controller, down to the register names, addresses and functions
00:28:56jhMikeSI'd change the setup altogether to do auto refresh
00:29:40jhMikeSthere'd be no register fiddling or DMA waits in the update function then
00:29:56_Bilgusyeah thats the ilitek one the sitronix data sheet was written with crayon and translated from chinese or something
00:31:40_Bilgusthough the st datasheet states: Notes1: The ST7781 requires no dummy write operation in high-speed write operation.
00:33:42_BilgusI wonder if they are similar enough in operation to work like you suggest
00:35:50jhMikeSthey both appear intended for that. the set up in the driver included front and back porch blanking intervals for vsync
00:36:31_Bilgusyeah there is also some pretty neat built-ins for video/still/resizing
00:36:53_BilgusI doubt they are utilized in the driver though
00:37:02jhMikeSthis shouldn't need more than lcd-memframe.c and a write-buffered framebuffer
00:37:02_Bilguswell in OUR driver
00:40:19jhMikeSsome targets can't really reach their full hw potential because of the simplistic lcd interface
00:41:34jhMikeSgigabeat S has a pretty hefty image processing unit that can interesting things and I'm not sure how i'd take advantage of it atm
00:52:07_Bilgusyep pretty sure its over my head I get the idea but can't even tell what needs to be removed vs added to implement lcd-memframe
00:54:56 Quit duncan^ (Quit: WeeChat 1.9.1)
00:56:15 Quit Moarc (Ping timeout: 248 seconds)
00:58:10 Join Moarc [0] (
01:01:14 Join MrZeus_ [0] (~MrZeus@2a02:c7f:7066:fb00:cda2:f27e:7272:1e90)
01:07:28[Saint]OK. I now have the start of the dynamic overflow menu working.
01:07:46[Saint]I also discovered a bug in the display logic for preloaded images.
01:07:53[Saint]But I can work around this.
01:09:56 Part chrisb ("rcirc on GNU Emacs 27.0.50")
01:10:39 Quit dandels_ (Ping timeout: 248 seconds)
01:13:21***Saving seen data "./dancer.seen"
01:15:27jhMikeS_Bilgus: set up DVI mode, it appears. start the dot clock and forget it. don't mess with controller registers after setup. make sure frame buffer isn't cached.
01:16:04_Bilgusregion `IRAM' overflowed by 1828 bytes
01:16:11jhMikeSlooks fairly clear from the LCDIF dics
01:16:26 Quit michaelni (Ping timeout: 240 seconds)
01:28:41_Bilgusnm it already has the frame buffer I think I have a decent idea now
01:28:58 Join michaelni [0] (
01:47:26 Quit duo8 (Ping timeout: 248 seconds)
01:49:48 Join dandels_ [0] (~dandels@unaffiliated/dandels)
01:50:25 Quit almog1006 (Quit: Page closed)
02:21:03 Quit dandels_ (Ping timeout: 240 seconds)
02:51:04 Join duncan^ [0] (UNKNOWN@unaffiliated/yulax)
02:55:16 Quit jhMikeS (Ping timeout: 268 seconds)
03:07:32 Quit MrZeus_ (Ping timeout: 264 seconds)
03:13:23***Saving seen data "./dancer.seen"
03:13:28 Join Huntereb_ [0] (
03:18:52 Join prof_wolfff_ [0] (
03:22:40 Quit Huntereb (*.net *.split)
03:22:40 Quit prof_wolfff (*.net *.split)
03:22:42 Quit ranmachan (*.net *.split)
03:26:21 Quit Huntereb_ (Quit: See ya!)
03:26:29 Join Huntereb [0] (
03:27:13 Join ranmachan [0] (
03:39:54 Join mendelmunkis [0] (
03:41:36 Quit mendelmunkis (Remote host closed the connection)
04:50:27 Join jhMikeS [0] (
04:54:56 Quit jhMikeS (Ping timeout: 260 seconds)
05:13:24***Saving seen data "./dancer.seen"
07:13:25***No seen item changed, no save performed.
07:19:07 Join edhelas [0] (
07:22:25skinkittenhey, here's an update :
07:23:05skinkittenwhen do I perform the hard-resets? is it at the end or can I intervene now? I'm impatient
07:23:31skinkittendo those hard resets help recover the disc?
07:28:47 Join almog1006 [0] (4d8bf186@gateway/web/freenode/ip.
07:49:01skinkittenthis thing is fried
07:49:05skinkitteni'm trashing it
07:53:57 Quit skinkitten (Quit: Leaving)
07:54:08 Join ender` [0] (
08:07:39 Join petur [0] (~petur@rockbox/developer/petur)
09:00:22 Join wodz [0] (
09:12:30 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:13:26***Saving seen data "./dancer.seen"
09:24:32_BilgusI have as yet been unable to get dotclk to work properly but that may be my inexperience, I've gotten it to display scattered and mirrored image but mainly just a grey white screen
09:26:04pamaury_Bilgus: dotclk requires extra wires, I am not sure those are wired. Also the controller needs to be switched to dotclk mode, on some controllers this is controlled by pins on the controller itself
09:26:36_Bilgusas far as the multiples of 4 It crashed on the oscope app every time the indicator got to the right of the screen even with it bounded properly so maybe something else is up but it seemed to work properly after that
09:27:46_BilgusI kinda figured there was some reason you didn't implement dotclk on some device but did on others maybe JHMikeS will have better luck
09:28:58pamaury_Bilgus: on the Zen, I don't have a choice, dotclk is the only available mode
09:29:23pamauryand believe me it's the most painful code I had to write because the stmp3700 has hardware bugs in the dma and lcdif
09:30:24pamaurythe imx233 is slightly different since it has a dedicated lcdif dma so it should work better
09:33:35pamauryI can have a try at it this weekend, although I am not convinced dotclk brings much advantage in our case
09:34:11pamaury_Bilgus: actually there is this comment in the file by me:
09:34:21_BilgusI did notice that when I started adding support for memframe I 1 had to move some sections of memory around and related it added a few KB to the image as well
09:34:25pamauryWARNING the B1P22 and B1P24 pins are used by the tuner i2c! Do NOT drive them as lcd_dotclk and lcd_hsync or it will break the tuner!
09:34:47_Bilgusah so they are shared with it then I take it?
09:35:37_Bilguslike literally break the tuner? If so I suppose its a good thing I don't use it for the radio
09:35:47pamaury_Bilgus: no, lcd_vsync and lcd_dotclk don't go to the lcd most probably
09:35:58pamauryso at best you can hope for vsync mode, not dotclk
09:36:21_Bilgussounds like diminishing returns
09:37:47pamauryI think it's very similar to dotclk, it has continuous update
09:38:51_Bilgusnow as for the patch< is there any reason you mod(height, 2) even though it only needs to be even?
09:39:11_Bilgusor does that work out to the same ASM as height & 1?
09:40:45pamauryI'm pretty sure the compiler optimizes this to a bit mask
09:47:11 Join dandels_ [0] (~dandels@unaffiliated/dandels)
09:49:21pamaury_Bilgus: should I push g#1751 ? it can't make it worse anyway
09:49:23fs-bluebot_Gerrit review #1751 at : Fuze PLUS Fix lcd_update_rect() by William Wilgus
09:51:47_BilgusIf you want but if I'm successful with VSYNC its obsolete, I suppose its a decent stop gap
09:53:36pamaury_Bilgus: it documents a problem, that might be useful to know
09:57:02_BilgusI'd like to get someone with the other lcd and see if it crashes for them as well my device has the ilitek controller maybe OP will check in and we can ask him for his lcd model as well [ <lukeoftheaura> ]
09:58:26pamaury_Bilgus: what is your lcd model? I don't remember which one my fuze+ uses but I can check tonight
09:58:57 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:d49:3b44:d479:7b05)
09:59:34pamauryI think mine has st7783
10:00:08_Bilgusthats the other one mine is ili9325
10:00:10 Quit dandels_ (Ping timeout: 276 seconds)
10:00:44_Bilgusthat pretty much rule out if its the device versus the LCD
10:01:29pamaury_Bilgus: if it hangs, it sounds like an lcdif problem
10:03:32_BilgusI'd like to think it wasn't a problem with the imx233 you'd think there would be an eratta by now I mean it is pretty rarely an issue since everything above 4 is a multiple of two but still
10:04:41pamaury_Bilgus: is the full device hanging or just the LCD?
10:04:57pamaurybecause the LCD cannot hang the SoC, there is no feedback from LCD to SoC
10:05:02_Bilgusthe whole device no usb no power off..
10:05:28_Bilgusah its just a dumb interface huh so maybe it is the processor
10:07:28pamauryyeah so sounds like an lcdif problem, the driver waits for the lcdif to be ready before sending anything. If for some reason the interface thinks the dma did not send enough data, this will hang
10:11:06_BilgusUnfortunately it didn't solve the other bug OP reported not that I figured they were related
10:12:24pamaurywhat is the other bug?
10:13:38_Bilguspeak meter in the voice recorder
10:14:26pamaurywhat is the problem?
10:15:52_Bilgusit doesn't work, My WAG is its probably reading the wrong channel or wrong interface but I haven't looked into it yet besides verifying that it exists in DEV and 3.14
10:16:20_Bilgushe said it still records fine so its really not a detriment
10:16:37pamauryjhMikes noticed in his pcm rework that the current imx233 code to get the peak buffer is wrong, I forgot to change it when I did some changes in the rest of the pcm driver
10:16:58pamauryit does not return the whole uffer
10:19:36pamauryat least for DAC it is wrong, not sure for ADC and how the peakmeter in voice recorder works
10:24:25*pamaury has to go
10:28:36 Quit pamaury (Ping timeout: 246 seconds)
10:39:31 Join jhMikeS [0] (
10:39:33wodzpamaury: (log) So I dived in atj213x boot rom. It has quit a few booting options. It first tries to boot from NAND, SD, SPI flash, ATA, I2C eeprom and finally jumps to USB ADFU handler if all fails
10:41:17 Part almog1006
11:03:05 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:11:11 Quit paulk-gagarine (Ping timeout: 248 seconds)
11:11:50 Join dovber [0] (~dovber@2600:8801:3180:19:beee:7bff:fee3:335f)
11:13:28***Saving seen data "./dancer.seen"
11:13:41 Quit dovber_ (Remote host closed the connection)
11:13:51 Join paulk-gagarine [0] (
11:30:14 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
11:58:40 Quit pamaury (Ping timeout: 248 seconds)
12:22:34 Quit Riku (Ping timeout: 240 seconds)
13:13:32***Saving seen data "./dancer.seen"
13:35:18 Quit petur (Quit: Leaving)
14:00:03 Quit _meg (Ping timeout: 250 seconds)
14:02:16 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
14:04:37 Join _meg [0] (~notsure@
14:17:54 Join xorly [0] (~xorly@
14:28:57 Join chrisb [0] (~chrisb@
15:03:18 Join smoke_fumus [0] (~smoke_fum@
15:13:36***Saving seen data "./dancer.seen"
15:19:59 Quit wodz (Ping timeout: 258 seconds)
15:33:56 Quit xorly (Ping timeout: 240 seconds)
15:55:05 Join xorly [0] (~xorly@
16:05:48 Join johnb2 [0] (
16:25:06 Join Acou_Bass [0] (
16:26:25 Join amayer [0] (
16:35:10 Quit xorly (Ping timeout: 250 seconds)
16:52:27 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:d49:3b44:d479:7b05)
17:10:07 Quit Acou_Bass (Ping timeout: 248 seconds)
17:13:39***Saving seen data "./dancer.seen"
17:15:05 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
17:21:59 Quit MrZeus (Read error: Connection reset by peer)
17:24:37 Join MrZeus [0] (~MrZeus@
17:25:53 Join xorly [0] (~xorly@
17:33:31 Join Acou_Bass [0] (
17:40:00 Quit xorly (Ping timeout: 248 seconds)
17:42:33 Quit MrZeus (Quit: Leaving)
17:47:45 Join PimpiN8 [0] (~textual@2a02:a454:38ea:1:d49:3b44:d479:7b05)
17:58:28 Quit johnb2 (Ping timeout: 255 seconds)
18:02:06 Quit Acou_Bass (Remote host closed the connection)
18:10:20 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
18:22:47 Join johnb2 [0] (
18:25:12 Quit sielicki (Ping timeout: 268 seconds)
18:30:24johnb2_Bilgus (logs): For the record, my (son's) read-only clip+ has now fully gone bad. It is usually just showing the 32MB partition. I tried getting it into Bootload USB mode (UP key pressed while connecting to USB). If briefly showed the Rockbox logo, then the "Debug->View I/O Ports" screen, then an ATA error.
18:32:19_Bilgusjust a matter of time I suppose I'm sure the extra fiddling didn't do it any favors
18:32:45johnb2Might be true.
18:33:17_Bilgusinteresting that it has 32 mb partition and recovers from it though
18:33:48johnb2That is not reliably repeatable.
18:33:58_Bilgusmakes me wonder if its a physical issue on the board or the chip
18:34:58johnb2I was thinking about that too. My son surely gave it a hard time. The display also dimmed or was dark every now and then.
18:35:14johnb2I will open it and check if I can spot anything.
18:35:19_Bilgusget someone to slap some flux on there and a good drag soldering with some real lead solder
18:36:09_Bilgusthe wholde lead free solder thing is a detriment to electronics not to mention the terrible atmosphere these players endure
18:37:07johnb2Different topic: I tried adding the MB patch on top of the Power Savings. I doesn't merge automatically.
18:37:24_Bilgusbut it could just be that the chip is kaput and the cells sometimes reset enough to pass muster to the controller
18:38:15_Bilgusno I would suspect it wouldn't lots of changes and MB hasn't been rebased in a while
18:39:33_Bilgusspeaking of have you tried out the latest PS patch on gerrit? I changed a few thing but it should be good to go even did a preliminary manual entry
18:40:23johnb2yes, I compiled today for fuze, clipp and zip.
18:40:33johnb2Fuze seems just fine.
18:41:20johnb2The other I need to test further.
18:41:42johnb2FuzeV2 I still have to compile.
18:43:51_Bilgusthe zip is still a bit sluggish on the display when cpu boost is off like with the cube demo but Its much better than it was and perfectly suitable for playing music and browsing menus
18:44:31johnb2Another hardware question: In the lot of fuzes I bought on *bay there are 2 Fuze v1 devices, that do not connect to usb at all. Charging however works, also the SD card. Is this more likely a mechanical problem or electronic?
18:45:08johnb2It is similar for the clip+ at first glance.
18:45:53johnb2Like you see the screen refresh from top to bottom.
18:46:35johnb2Have you done any tests what are the respective power savings of each feature?
18:48:27_Bilgusthat sounds mechanical but It could be that someone fried them too
18:49:04_BilgusMB should be good now I rebased it and its parent commit
18:51:15_Bilgusnot beyond some initial stuff 100- 50% battery and between voltage (low) and the second run with all optimization it increased by ~40%
18:52:29_Bilgusbut did one optimization save 37% and the other two save 3% idk something that I need to test over a few days
19:07:54 Join lebellium [0] (
19:13:40***Saving seen data "./dancer.seen"
19:28:40johnb2: the patch still doesn't merge on master
19:29:27_BilgusMB onto master or PS onto MB?
19:30:19johnb2I tried MB on PS first and get compile errors.
19:30:42johnb2then MB on master, git errors
19:31:07_Bilgusok try this first go back to master do git pull −−rebase
19:32:06_Bilgusactually better would just be to review the whole thing
19:33:02_Bilgusdo git review -d I697b3d0499f85e789c3020bc2133fbe0023f72a2 && git branch -m multiboot_12_1
19:34:28johnb2Hm, I have added the review feature a while ago, but failed to specify the repository.
19:34:29johnb2No '.gitreview' file found in this repository.
19:34:29johnb2We don't know where your gerrit is. Please manually create
19:34:29DBUGEnqueued KICK johnb2
19:34:29johnb2a remote named gerrit and try again.
19:34:46johnb2Any hint how to set this?
19:35:42_Bilgusgive me a second i'll have to look it up its been a long while
19:38:37_Bilgusyou already have a git account and ssh key correct?
19:39:21_Bilgusgit remote set-url −−push origin ssh://
19:41:16johnb2I still get the "No '.gitreview' file found in this repository." after that.
19:41:18_Bilgusand I think fetch is the same git remote set-url −−fetch origin ssh://
19:42:36_Bilgusnope fetch would be git remote set-url −−fetch origin git://
19:43:06_Bilgusthen type git remote -v and it should list the entries for push and fetch
19:43:18_Bilguspush should be the ssh your username
19:43:27_Bilgusand fetch should list git
19:45:31johnb2I does, but still review doesn't work:
19:47:51johnb2I don't want to bother you too much, but it is strange that fetching MB onto master doesn't work. I can try and clone the whole repo again.
19:50:04 Quit dys (Ping timeout: 255 seconds)
19:50:23_Bilgusit doesn't have the gerrit ones in there try git remote add gerrit ssh://
19:51:00_Bilgusthen when you type git remote -v it should have two gerrit entries and the two origin entries
19:51:04 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:51:19johnb2yes, it now does
19:51:31_Bilgusnow it should work
19:52:16johnb2yep, it did something. So now this is the MB and I have to add the pOwSav?
19:53:41johnb2Can I still do this with cherry pick?
19:54:04_Bilgusgit review -d I04137682243be92f0f8d8bf1cfa54fbb1965559b && git branch -m powersave
19:54:47_Bilgusnext type git checkout multiboot_12_1
19:55:13_Bilgusand finally git rebase powersave
19:55:29_Bilgusand it should allpy everything for you
19:57:08johnb2git checkout multiboot_12_1 or multiboot_12? the first gives me: error: pathspec 'multiboot_12_1' did not match any file(s) known to git.
19:58:14_Bilgusthe first review you did
19:58:44_Bilgusif you can't find it it might have not renamed type 'git branch'
19:59:38_Bilgusit'll bring up a list if it didn't rename it'll be something like review/gerrit/bleh/williamwilgus/I697b3d0499f85e789c3020bc2133fbe0023f72a2
20:00:41_Bilgusif so checkout that branch and do git branch -m multiboot_12_1
20:01:04_Bilgusand finally finish with git rebase powersave
20:01:31_Bilgusyou could use that big long ID too without rename any branch but that gets tedious
20:01:31johnb2it is compiling now (with multiboot_12).
20:02:05johnb2Thanks for walking me there ;-)
20:03:00_Bilgusso to review you, reviewed + renamed multiboot, you reviewed + renamed powersave, switched to multiboot and rebased it with powersave
20:03:38_Bilgusthats the same way I learned how to use it too from [Saint]
20:04:03pamaury_Bilgus: I can confirm the oscilloscope bug on fuze+ with st display
20:04:07_Bilgusso you are welcome
20:04:09johnb2I will save that conversation and try and understand it later ;-)
20:04:50_Bilguspamaury then I'd say its for sure a bug in LCDIF thanks for checking
20:05:56pamaury_Bilgus: I'll have a try at vsync mode tonight/this weekend
20:06:51_Bilguscool I'm sure your code will be a lot more likely to work :p otherwise I plan to mess with it monday or tues
20:07:14 Join Acou_Bass [0] (
20:07:23pamauryit requires some care, one need to change not only the driver's code but also some LCD registers
20:08:39_Bilgusyeah i was reading the data sheets the timing seems like a PITA
20:10:34_Bilgusmight be a good idea to do most of the leg work in a macro and have it use the actual pixclk so it is at least easily changed
20:14:42johnb2_Bilgus : build is running now on the clip+. Thanks.
20:15:59_Bilgusnice I haven't tried on my main clip+ yet since I only run sd builds there but I do love the way multiboot allows switching between builds
20:16:33_BilgusI think a plugin that searches and allows selection for next boot might be a nice addition
20:19:16johnb2Specifying the path still seems to be picky. I only use Notepad++. I had "/rb/clipp" and it complained about No .rockbox directory. I changed it to "/rb/clip" with Notepad++ and it worked. No additional chars or linebreaks in both cases.
20:20:59johnb2additional = trailing
20:24:40_BilgusIf you use a trailing / I think it should work properly
20:25:18 Join dandels_ [0] (~dandels@unaffiliated/dandels)
20:26:42_Bilgusits picky though because I was going for the absolute smallest code size possible since a few of the targets were down to not fitting by 10 bytes and such, JhMikeS and I have fixed that now though so I might revisit it later to make it slightly more robust
20:27:15johnb2I will try the trailing /.
20:29:48 Join xorly [0] (
20:30:38 Quit Acou_Bass (Remote host closed the connection)
20:31:37 Quit C-Keen (Quit: WeeChat 1.9.1)
20:33:47 Join Acou_Bass [0] (
20:34:08 Join dys [0] (
20:35:07 Quit Acou_Bass (Client Quit)
20:39:51 Join Acou_Bass [0] (
20:39:55 Quit Acou_Bass (Remote host closed the connection)
20:42:13 Join Acou_Bass [0] (
20:52:14 Join sielicki [0] (~sielicki@unaffiliated/n1cky)
20:52:45 Quit johnb2 (Ping timeout: 260 seconds)
20:53:58 Join lukeoftheaura [0] (~lukeofthe@
20:54:39lukeoftheaurais it possible to set the database to only scan a specific fonder, and then have a separate folder for non-music audio files that isn't scanned?
20:56:53pamaurylukeoftheaura: see
20:57:09pamauryyou can create file database.ignore in directories you want the database to ignore
21:04:12 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:584:1b0e:3be5:84bf)
21:09:21 Quit Jon (Ping timeout: 264 seconds)
21:10:27 Join Jon [0] (
21:10:28 Quit lukeoftheaura (Quit: Leaving)
21:11:38 Join Jinx [0] (Dojo@unaffiliated/jinx)
21:11:56 Quit sielicki (Ping timeout: 240 seconds)
21:13:38 Join sielicki [0] (~sielicki@unaffiliated/n1cky)
21:13:41***Saving seen data "./dancer.seen"
22:01:44 Join _jhMikeS_ [0] (
22:01:44 Quit jhMikeS (Disconnected by services)
22:01:45 Nick _jhMikeS_ is now known as jhMikeS (
22:05:42 Quit sielicki (Read error: Connection reset by peer)
22:05:52 Quit amayer (Quit: Leaving)
22:09:17 Join sielicki [0] (~sielicki@unaffiliated/n1cky)
22:13:34 Quit dandels_ (Ping timeout: 248 seconds)
22:20:50 Quit dys (Ping timeout: 246 seconds)
22:33:19__builtin_Bilgus: I don't quite understand your code in g#1751
22:33:20fs-bluebot_Gerrit review #1751 at : Fuze PLUS Fix lcd_update_rect() by William Wilgus
22:33:36__builtinwhy would you use the comma operator here? "w += x, x = 0"
22:34:12_Bilgusthat was actually already in from the code mike gave me for it but it is valid c
22:35:14__builtinI know it's valid, but why not make it less convoluted?
22:36:20 Join bp0 [0] (~bp@
22:36:20 Quit bp0 (Changing host)
22:36:20 Join bp0 [0] (~bp@unaffiliated/bp0)
22:45:11jhMikeSgeez, the fuck in convoluted about a comma operator?
22:46:14__builtinmaybe confusing would be a better word
22:46:18_Bilgusit is easy to overlook scanning the code
22:47:29jhMikeS__builtin: to whom, those who don't know C?
22:48:08__builtinand those trying to learn, yes
22:48:23__builtinI just think a block with two statements would be more clear in this case
22:48:34jhMikeS__builtin: then they should see those things if they want to learn
22:49:10jhMikeS__builtin: meh, people object to weird things for some reason, often because someone said they should
22:50:11__builtinI mean, either one works, but one is just more common
22:50:28jhMikeSthough I'll admit I put it there about 10 years ago just so someone would complain about it.
22:50:36__builtinif you insist on keeping it, go ahead
22:51:11jhMikeSI don't
22:52:45jhMikeSI'm just amused a bit
23:01:44 Quit ender` (Quit: It's not a dictatorship. It's an alternate democracy.)
23:02:31pamauryif C wasn't the terrible language it is, I wouldn't mind the comma operator, but C is a terrible language, the comma operator does not help with that. At least modern compilers warn about misleading indentation, but we don't use modern compilers either ;)
23:04:34jhMikeSIt's just slightly less terrible than others
23:13:42***Saving seen data "./dancer.seen"
23:38:22 Quit chrisb (Excess Flood)
23:46:46 Quit _meg (Ping timeout: 255 seconds)
23:47:40 Join _meg [0] (~notsure@
23:48:06_BilgusOk __builtin I spread it out a bit and even knocked off a few instructions in the assembly while I was at it
23:50:26 Quit ZincAlloy (Quit: Leaving.)
23:52:09_Bilgusoh and pamaury h &1 was quite a bit shorter than h%2 i'm guessing since its not unsigned but I would have figured the compiler would have figured it out

Previous day | Next day