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 2021-08-05

00:40:05***Saving seen data "./dancer.seen"
02:20:34 Quit F3l1x_10m (Ping timeout: 240 seconds)
02:40:09***Saving seen data "./dancer.seen"
03:13:04 Join ideasman42 [0] (~ideasman4@2402:b801:2844:2100::1)
03:15:01 Quit ideasman42 (Client Quit)
04:14:49 Join ZincAlloy [0] (
04:19:17 Quit ZincAlloy (Ping timeout: 245 seconds)
04:27:34 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:f49f:88f4:4840:d1fb)
04:32:28 Quit ZincAlloy (Ping timeout: 272 seconds)
04:40:12***Saving seen data "./dancer.seen"
04:50:01 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:c53:bc06:23e3:720a)
04:50:17 Join Moriar [0] (
05:19:10 Join johnb2 [0] (
05:21:15 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
06:05:47 Quit Moriar (Ping timeout: 258 seconds)
06:26:53rb-bluebotBuild Server message: New build round started. Revision 03a6eb63f1, 303 builds, 9 clients.
06:40:16***Saving seen data "./dancer.seen"
06:44:00 Quit F3l1x_10m (Ping timeout: 276 seconds)
06:44:19_bilgusuh oh
06:44:34_bilgusmight be stuck
06:45:31rb-bluebotBuild Server message: Build round completed after 1118 seconds.
06:45:34rb-bluebotBuild Server message: Revision 03a6eb63f1 result: All green
06:45:43rb-bluebotBuild Server message: New build round started. Revision 848633f921, 303 builds, 9 clients.
06:46:19_bilgusspoke too soon
06:50:20 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
06:51:48 Quit johnb2 (Quit: Nettalk6 -
07:01:04rb-bluebotBuild Server message: Build round completed after 921 seconds.
07:01:05rb-bluebotBuild Server message: Revision 848633f921 result: All green
07:01:07rb-bluebotBuild Server message: New build round started. Revision d57900ae12, 303 builds, 9 clients.
07:14:59rb-bluebotBuild Server message: Build round completed after 831 seconds.
07:15:01rb-bluebotBuild Server message: Revision d57900ae12 result: All green
07:15:04rb-bluebotBuild Server message: New build round started. Revision cb6b0d2c0e, 303 builds, 9 clients.
07:27:36rb-bluebotBuild Server message: Build round completed after 752 seconds.
07:27:38rb-bluebotBuild Server message: Revision cb6b0d2c0e result: All green
08:07:57 Nick yang-idl1 is now known as yang-idle (~yang@
08:08:18 Quit yang-idle (Changing host)
08:08:18 Join yang-idle [0] (~yang@fsf/member/yang)
08:22:56 Join cereal_eater [0] (
08:23:35 Join dconrad [0] (~dconrad@
08:23:41 Quit dconrad (Client Quit)
08:23:57cereal_eaterAny comments on the issue with the colour selector and the anti alieased fonts?,53937.msg249012.html#msg249012
08:29:09_bilgusbraewoods iriver.c iriver_encode/decode
08:29:38_bilgusany clue why no free is called?
08:30:32_bilgusI was thinking if you were in a rolo like situ it wouldn't matter but this file runs on the host side
08:32:04_bilguscereal_eater, those are macros
08:32:10_bilgusgve me a moment
08:38:26 Join massiveH [0] (
08:40:17***Saving seen data "./dancer.seen"
08:42:10_bilgusit appears the device shines through in that color picker
08:45:10cereal_eaterthanks for the information, unfortunately this is way above my head and I don't understand the range from 0 to 31 instead of 0 to 255.
08:45:20_bilgusmight be a bug that 31 i thingk it should be R & 0x3F
08:46:52cereal_eaterfor green it is 0 to 63
08:46:58_bilgusits 565
08:47:05_bilgusor reverse 565
08:47:39_bilguswhat is the rgb color you want
08:47:46_bilgusI'll do the math for you
08:49:00cereal_eaterplease show me the math as in teach me to fisch :-)
08:51:30cereal_eaterthe RGB is 48, 79, 254
08:55:53speachy_bilgus: the pile of scanner emails is.. confusing. it's acting almost like they are radically different projects being built under the same umbrella.
08:56:25_bilgusI think that might make it a do one device get it clean switch to another
08:56:45speachy_bilgus: btw, I didn't intend to push coverity stuff via the buildfarm; instead have it batch one or two "representative" builds on a nightly basis. eg one hosted, one native
08:56:46_bilgusbut at least it does have the bugs in the snapshots
08:57:47_bilgusI assumed you would have one builder that did that in the pool but I think it being so slow itd probably be wise to not do that
08:57:59cereal_eater_bilgus: is it like 31 is 255 and all numbers in between correspond linearly?
08:58:20_bilguseverything gets anded with 3f
09:01:57braewoods_bilgus: if i had to guess, it's lazy C programming. i've seen this crap in short lived utilities. they let the kernel clean up their mess.
09:02:57speachy"we're terminating anyway, who cares about cleaning up first?"
09:03:24speachyone bit of code I'm responsible for only leaks because libusb does.
09:03:26braewoods_bilgus: why ping me about it though? i don't recall touching that code.
09:06:52braewoodshm. interesting. i wonder if any program uses the extra fields of gzip containers.
09:06:56braewoodsgzip itself only uses filename
09:07:27_bilgusmaybe not that function in particular? but the file so you've seen more of it than I :)
09:07:48braewoodslet me look at it
09:07:52_bilguscereal_eater, give me a few i'll do a write up
09:10:01braewoods_bilgus: looks like you could free pChecksums. it retains the original pointer and doesn't appear to be modified.
09:10:11braewoodsit'd need to be added to each exit point
09:10:15braewoodsshall I prepare a patch?
09:11:09braewoodsright now i've been testing my gzip container code
09:13:38speachyoh, one request −− if you mark a coverity issue as "ignore" please add a comment why.
09:13:46_bilgusbraewoods, sure if you want
09:13:57speachy(and classification as "intentional" or whatever else is appropriate)
09:14:38_bilgussure hopefully the modeling file makes that less frequent
09:15:18speachyI've been focusing on "control flow" issues; a lot of them are due to #ifdef chains from SIMULATOR builds or something like that
09:15:44speachyso it's not a "false positive" per se; the analysis is correct but we don't care.
09:16:11speachy(eg peak meters in sim builds don't work, because we don't have working recording in the simulator)
09:23:11braewoods g#3655
09:23:14rb-bluebotGerrit review #3655 at : tools/iriver: fix resource management in encode/decode functions by James Buren
09:23:16braewoodswhat you asked for
09:23:30braewoodsi ended up redoing the resource management to handle all the exit paths easier
09:24:13braewoodsaka, add a single exit point that is jumped to via goto
09:38:42 Quit cereal_eater (Ping timeout: 252 seconds)
09:42:47 Join cereal_eater [0] (
09:44:59rb-bluebotBuild Server message: New build round started. Revision 9f0f2c6658, 303 builds, 9 clients.
09:46:42_bilgushopefully that makes it clear enough
09:47:14_bilgusbasically each field gets converted t 0-31 or 0-63(GREEN) then you shift them into place
09:47:52_bilgusbraewoods, thank you!
09:48:34braewoodsi'm close to being ready to push inflate code to gerrit
09:49:08cereal_eaterwow, cool thanks for the explanation. I'm gonna wrap my head around this and customize some themes.
09:49:56_bilgusbe aware that the range is smaller
09:50:13_bilguslike they get muddy
09:50:28cereal_eaterOK, I see.
09:50:31_bilgusso sometimes its best to tweak it by eye
09:51:09_bilgusI'm not familar enough with it enough to tell you the math for that but by eye works for me :)
09:58:19rb-bluebotBuild Server message: Build round completed after 800 seconds.
09:58:24rb-bluebotBuild Server message: Revision 9f0f2c6658 result: All green
10:03:43speachyso far both high/memory corruption issues in the mp4 metadata parser are false positives.
10:04:51cereal_eaterby the way, is there interest in the anti aliased fonts? Could I upload a pack to the wiki?
10:07:53speachyoh, _bilgus −− When I get this server upgrade dealt with I'd like to pick up your coverity build/submission scripts so this stuff can all proceed automagically.
10:22:09 Quit massiveH (Quit: Leaving)
10:23:00rb-bluebotBuild Server message: New build round started. Revision a20755e9ef, 303 builds, 9 clients.
10:35:16rb-bluebotBuild Server message: Build round completed after 736 seconds.
10:35:18rb-bluebotBuild Server message: Revision a20755e9ef result: All green
10:37:56 Quit cereal_eater (Quit: Connection closed)
10:40:18***Saving seen data "./dancer.seen"
10:40:22speachySo, straw poll −− how does Monday morning (US eastern time) sound for taking things down for the upgrade?
10:54:33braewoodsok... fixed a few bugs I found in my gzip container code (mostly due it not being tested before)
10:55:22braewoodsi think that's as far as i can get just using my desktop to test...
10:55:55braewoodsnow i need to write some code to test it on an actual target so i can see if there's any obvious issues
11:29:42 Quit akaWolf (Ping timeout: 258 seconds)
11:32:02 Join akaWolf [0] (
11:36:42braewoodsunsurprisingly it works
11:36:56braewoodsnow for the next testing phase; how much does cpu boosting help?
11:37:11braewoodsit took like 68s for a 16MB file
11:37:14braewoodsw/o i t
11:41:02braewoodshuh. much faster but still slow.
11:41:13braewoodstime to see how much is due to IO writes
11:43:30braewoodsof course IO is going to suffer some due to the limited RAM
11:45:46braewoodshuh. not much difference. it really is that slow
11:46:17braewoodstakes 27s for my iriver h120 to decompress 16MB
11:46:23braewoodsit does work though
11:46:42braewoodsbut i can't see this being practical for large payloads
11:56:00 Quit tomato (Ping timeout: 268 seconds)
11:58:41braewoods g#3660
11:58:43rb-bluebotGerrit review #3660 at : inflate: import initial module for deflate decompression by James Buren
11:58:47braewoodswhoever wants to review it
11:59:11_bilgusspeachy sounds great
11:59:38_bilgusbraewoods I can't sit with it rn but will tonight or tomorrow if no one else does
11:59:51_bilgusspeachy there isn't much going on there
12:00:06speachyI'd love to do it tomorrow but it conflicts with the meet&greet for the kiddo's school
12:00:27_bilgusI meant with the coverity builds
12:00:40_bilguscurl command is the worst of em
12:00:44braewoods_bilgus: we can still go with your ZIP idea but i suspect the slow decompression will mean we'd want to still use STORE method (uncompressed)
12:01:20braewoodsit's fast on my desktop probably mainly due to more RAM and clock speed
12:02:09braewoodsdoes work but it is slow. i'll work on a ZIP UI extractor later on.
12:02:31_bilgusbraewoods, I don't know that we will do them on the fly it would be nice from a read-only FS standpoint
12:03:06_bilgusbut even uncompressed its not like these are substantial space hogs
12:03:08braewoodsi ended up adapting the plan9 inflate algorithm
12:03:43_bilguswas that the single byte at a time one?
12:03:54braewoodsyes but i modified it with new IO routines
12:04:04braewoodsit still is but it reads from an IO buffer now
12:04:17braewoodsso it doesn't have as much callback overhead
12:04:24braewoodsit should be faster on our beefier ports
12:05:04braewoodsi'm using the minimum buffers i could get away with
12:05:07braewoodslike 32K each
12:05:15braewoodsaround 70K of non-stack memory is needed
12:05:25braewoodsi moved most of the heavy stuff off the stack
12:05:37braewoodsleaving only the small state variables
12:06:30braewoodsi wouldn't be surprised if the main reason it's slow on coldfire is the small amount of "fast ram" it has
12:07:38speachyIIRC the coldfire also lacks a data cache too
12:08:56braewoodsit may not be worthwhile to add that firmware extracting feature if it's THIS slow
12:09:04braewoodsi'll test it out on more targets
12:09:57braewoodsbut it was kinda sad seeing data rates of like 500 kilobytes a second
12:10:00_bilgusthats the rocker
12:10:49_bilgusive done the 3 I had cross compilers for on this laptop I think arm mips and m68k?
12:11:07_bilgusarm was the only one that I removed the cache on
12:12:12_bilgusoh sorry wrong convo on the cache
12:40:22***Saving seen data "./dancer.seen"
12:46:57 Part edhelas
12:47:31 Join edhelas [0] (
13:29:24 Quit rogeliodh (Quit: The Lounge -
13:30:47 Join rogeliodh [0] (
14:40:24***Saving seen data "./dancer.seen"
14:50:02 Part edhelas
14:50:41 Join edhelas [0] (9d94237298@2a03:4000:51:f44:4e1:2ff:fe00:4257)
16:40:28***Saving seen data "./dancer.seen"
17:52:19 Join Moriar [0] (
18:40:31***No seen item changed, no save performed.
18:45:46 Quit ZincAlloy (Quit: Leaving.)
18:47:34 Quit _bilgus (Quit: Leaving)
20:34:52speachy_bilgus: so we should ideally submit one per toolchain then? Including hosted and SDL/sim...)
20:35:25speachywe're in the only 2 per day bucket though
20:37:15speachyonly a handful of issues have been under firmware/target, and even firmware/ has been pretty sparse. plugins and codecs are where it's all at..
20:39:48braewoodsspeachy: wasn't coverity mainly intended for hosted programs?
20:39:54braewoodsnot sure how useful it is for freestanding
20:40:10speachysure, but C is C..
20:40:26speachyit's not the different toolchains we care about per se, but the coverage from the different targets they represent
20:40:32***Saving seen data "./dancer.seen"
21:08:09 Quit Moriar (Quit: Leaving.)
21:53:19 Join advcomp2019_ [0] (~advcomp20@user/advcomp2019)
21:56:58 Quit advcomp2019__ (Ping timeout: 240 seconds)
22:06:24 Quit edhelas (Remote host closed the connection)
22:20:06 Join _bilgus [0] (~bilgus@2600:1702:3810:17f0:18b6:b7d6:6d21:7004)
22:40:36***Saving seen data "./dancer.seen"
23:08:13rb-bluebotBuild Server message: New build round started. Revision da45b37fac, 303 builds, 9 clients.
23:19:20rb-bluebotBuild Server message: Build round completed after 668 seconds.
23:19:23rb-bluebotBuild Server message: Revision da45b37fac result: All green

Previous day | Next day