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 2021-07-11

01:00
01:00:16DEBUGEOF from server (Success) (snapshot: netstuff.c line 545)
01:00:16***Cleanup
01:00:16***Cleanup
01:00:16***No seen item changed, no save performed.
01:00:16***Exit
01:00:16***Started Dancer V4.16
01:00:16***Connected to irc.libera.chat on port 6667
01:00:16***Logfile for #rockbox started
01:00:18Mode"rb-logbot :+i" by rb-logbot
01:00:24***Server message 501: 'rb-logbot :Unknown MODE flag'
01:00:24 Join rb-logbot [0] (~rockbox@stuffed.shaftnet.org)
01:00:24 Join funman_ [0] (~fun@chui-pas.net)
01:00:24 Join munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
01:00:24 Join akaWolf [0] (~akaWolf@akawolf.org)
01:00:24 Join Romster [0] (~romster@user/romster)
01:00:24 Join advcomp2019 [0] (~advcomp20@user/advcomp2019)
01:00:24 Join tchan [0] (~tchan@c-98-206-141-238.hsd1.il.comcast.net)
01:00:24 Join Maxdamantus [0] (~Maxdamant@user/maxdamantus)
01:00:24 Join Retr0id [0] (~Retr0id@user/retr0id)
01:00:24 Join Airwave [0] (~Airwave@ti0006a400-1599.bb.online.no)
01:00:24 Join Galois [0] (djao@efnet.math.uwaterloo.ca)
01:00:24 Join bertrik [0] (~bertrik@revspace/participant/bertrik)
01:00:24 Join bluebrother [0] (~dom@55d4dc16.access.ecotel.net)
01:00:24 Join amiconn [0] (jens@p2e55b6bb.dip0.t-ipconnect.de)
01:00:24 Join aquijoule__ [0] (~richbridg@213-225-7-126.nat.highway.a1.net)
01:00:24 Join JanC [0] (~janc@user/janc)
01:00:24 Join pablocastellanos [0] (~pidgin@user/pablocastellanos)
01:00:24 Join j-r [0] (~j-r@p5df2ecee.dip0.t-ipconnect.de)
01:00:24 Join ufdm [0] (~ufdm@c-73-164-63-214.hsd1.mn.comcast.net)
01:00:24 Join yosafbridge [0] (~yosafbrid@static.38.6.217.95.clients.your-server.de)
01:00:24 Join desowin [0] (~linux@rockbox/developer/desowin)
01:00:24 Join pixelma [0] (marianne@p2e55b6bb.dip0.t-ipconnect.de)
01:00:24 Join F3l1x_10m [0] (~Al3x_10m@user/f3l1x-10m/x-3393542)
01:00:24 Join spork [0] (topic@31-151-2-135.dynamic.upc.nl)
01:00:24 Join dbohdan [0] (~dbohdan@user/dbohdan)
01:00:24 Join SammysHP [0] (~SammysHP@faol.sammyshp.de)
01:00:24 Join CasBot [0] (~matrixbot@connolly.tech)
01:00:24 Join ats [0] (~ats@cartman.offog.org)
01:00:24 Join Topy44 [0] (kdqScR72dZ@bellatrix.uberspace.de)
01:00:24 Join +speachy [0] (~speachy@rockbox/developer/speachy)
01:00:24 Join reductum [0] (~reductum@cpe-72-134-86-174.natsow.res.rr.com)
01:00:24 Join emacsomancer [0] (~emacsoman@c-174-52-88-123.hsd1.ut.comcast.net)
01:00:24 Join kadoban [0] (~kadoban@user/kadoban)
01:00:24 Join Natch [0] (~natch@c-e070e255.014-297-73746f25.bbcust.telenor.se)
01:00:24 Join olspookishmagus [0] (~pookie@snf-137798.vm.okeanos.grnet.gr)
01:00:24 Join braewoods [0] (~braewoods@user/braewoods)
01:00:24 Join gsora [0] (~gsora@140.238.174.213)
01:00:24 Join ddevault [0] (znc@sourcehut/staff/ddevault)
01:00:24 Join XDjackieXD [0] (~jackie@banana-new.kilobyte22.de)
01:00:24 Join Guest9931 [0] (~leachim6@neptune.cosmicforge.org)
01:00:24 Join Guest5135 [0] (~scorche@184-98-25-134.phnx.qwest.net)
01:00:24 Join TorC [0] (~Tor@fsf/member/TorC)
01:00:24 Join gevaerts [0] (~fg@user/gevaerts)
01:00:24 Join user890104 [0] (~Venci@freemyipod/user890104)
01:00:24 Join Arsen [0] (~arsen@managarm/dev/Arsen)
01:00:24 Join Rondom [0] (~rondom@user/rondom)
01:00:24 Join rudi_s [0] (~simon@user/rudi-s/x-7673890)
01:00:24 Join aevin [0] (eivindsy@microbel.pvv.ntnu.no)
01:00:24 Join jbgg_ [0] (~jbgg@95.179.159.229)
01:00:24 Join Xeha [0] (~Xeha@dynamic-82-220-88-142.ftth.solnet.ch)
01:00:24 Join vup [0] (~~~~@46.101.193.235)
01:00:24 Join hook54321 [0] (sid149355@user/hook54321)
01:00:24 Join wisperwind [0] (~quassel@user/wisperwind)
01:00:24 Join _bilgus [0] (~bilgus@162.154.213.134)
01:00:24 Join benjaoming [0] (~benjaomin@37.139.19.237)
01:00:24 Join toruvinn [0] (~toruvinn@77-255-90-179.adsl.inetia.pl)
01:00:24 Join rasher [0] (~rasher@user/rasher)
01:00:24 Join markun [0] (~markun@178-84-100-63.dynamic.upc.nl)
01:00:24 Join rogeliodh [0] (~rogeliodh@rogeliodh.dev)
01:00:24 Join __builtin [0] (~quassel@rockbox/developer/builtin)
01:00:24 Join michaelni [0] (~michael@213-47-68-29.cable.dynamic.surfer.at)
01:00:24 Join Bobathan [0] (~admin@cpe-65-29-248-157.wi.res.rr.com)
01:00:24 Join dys [0] (~dys@user/dys)
01:00:24 Join f1refly [0] (~f1refly@pd954c410.dip0.t-ipconnect.de)
01:00:24 Join asaba [0] (~asabas@103.113.159.184)
01:00:24 Join rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
01:00:24 Join blbro[m] [0] (~blbrostra@2001:470:69fc:105::8f7)
01:00:24 Join tomato [0] (~tomato@user/tomato)
01:00:24 Join Piece_Maker [0] (~Piece_Mak@cpc95746-bolt17-2-0-cust360.10-3.cable.virginm.net)
01:00:24 Join jschwart [0] (~quassel@2001:985:2c6e:0:b00b:32ff:fe28:5567)
01:00:24 Join danwellby [0] (~danwellby@88.97.8.245)
01:00:24 Join Ckat [0] (~Ckat@xn--z7x.xn--6frz82g)
01:00:26 Join ParkerR [0] (~ParkerR@znc.withg.org)
01:00:34 Join kirvesAxe [0] (kirvesaxe@user/kirvesaxe)
01:17:07braewoods g#3567
01:17:09rb-bluebotGerrit review #3567 at https://gerrit.rockbox.org/r/c/rockbox/+/3567 : zip: implement zip extraction support by James Buren
01:17:46braewoodsit tests ok so far. i think it will cause some problems for ports where modtime isn't yet available but that should be easily remedied.
02:00
02:35:23 Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
02:52:51 Quit munkis (Remote host closed the connection)
02:53:15 Join munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
03:00
03:00:18***Saving seen data "./dancer.seen"
03:16:13 Quit ZincAlloy (Quit: Leaving.)
03:19:43 Join lebellium [0] (~lebellium@2a01:cb10:2e:2000:a5f1:9e7:7890:b2d2)
04:00
04:25:44 Quit blbro[m] (Quit: Bridge terminating on SIGTERM)
04:25:47 Quit kadoban (Quit: Bridge terminating on SIGTERM)
04:28:19 Join kadoban [0] (~kadoban@user/kadoban)
04:34:37 Join blbro[m] [0] (~blbrostra@2001:470:69fc:105::8f7)
04:45:43 Join mendel_munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
04:45:58 Quit munkis (Killed (lithium.libera.chat (Nickname regained by services)))
04:45:58 Nick mendel_munkis is now known as munkis (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
05:00
05:00:22***Saving seen data "./dancer.seen"
05:47:03 Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
05:47:23 Quit ZincAlloy (Client Quit)
05:47:32 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:f18b:877:3b86:12d7)
07:00
07:00:23***Saving seen data "./dancer.seen"
07:15:06 Join amachronic [0] (~amachroni@user/amachronic)
07:30:00_bilgusbraewoods you ready to push that?
07:54:16 Quit f1refly (Quit: see ya in hell)
07:56:37 Join f1refly [0] (~f1refly@pd954c410.dip0.t-ipconnect.de)
08:00
08:01:09braewoods_bilgus: more or less. i just mentioned the one compile issue that may arise. for unix ports they may need an implementation of modtime. easiest option is to wrap utime and set both fields to the mtime value provided.
08:01:44_bilgusk, guess we will see where it falls
08:03:16_bilgusoh you have a comment from amachronic?
08:03:43braewoodsyea, stuff i can address in another commit.
08:04:17braewoodsi was aliasing path and name so i could append to the proper position
08:04:26braewoodsit was a small time optimization
08:04:40braewoodsavoid recopying the same initial segment
08:04:47braewoodsthat doesn't change
08:05:30_bilgusok so good then?
08:05:33braewoodsas for the error codes? i don't see much value in documenting them because i was leaning towards using a unique one
08:05:35braewoodsafaik
08:05:50braewoodsfor each one so i can know which one triggered the return
08:06:02amachronic^^^ that's why I don't care so much about naming them
08:06:21braewoodsit seemed a standard rockbox practice
08:06:40braewoodsin any case, negative values have a special case in callbacks here
08:06:54braewoodsthey mean "skip to next file" or so.
08:07:08braewoods0 means continue as normal
08:07:18braewoodspositive are error codes for the callbacks
08:07:29braewoodsnegatives are error codes for the regular zip code
08:07:55amachronicokay, so that's why that one call to zip->cb ignores the negative case. makes sense now
08:08:26braewoodsi was trying for a trinary system of sorts
08:08:41braewoodswhile still giving some kind of meaning to them
08:09:12rb-bluebotBuild Server message: New build round started. Revision 8846e087c0, 297 builds, 8 clients.
08:09:29_bilgusgot it I put your comment into the reply
08:10:35braewoodsi plan to enlarge my data buffer for the data reading later to better support DEFLATE
08:10:45braewoodsmain issue is figuring out the ratio
08:10:52braewoodsof the in/out buffer
08:11:06braewoodsso worst case won't be too big for out buffer
08:11:53braewoodsi'm already using ~65k for the first major read i do
08:11:57braewoodstemporarily
08:12:02_bilguswont processor speed be the biggest factor for ratio?
08:12:28_bilguslike a slower processor would o better with a larger ratio to the out buffer?
08:12:30braewoods_bilgus: no, as in, given X input how much space will the Y output consume
08:12:56braewoodsi found one weakness of the tinflate option is
08:13:15braewoodsif your output buffer isn't large enough, it'll just discrd the rest and finish processing the input
08:13:40braewoodsso if i want to use it, i need to size the buffers so that doesn't happen
08:14:09braewoodsthing is i never stopped to think about the best case ratio for deflate is
08:14:27braewoodsi'd guess 90% is a bit high
08:14:28_bilgusah so that would be rather deleterious to the whole thing
08:14:51braewoodsindeed and there's some issues with inflate algorithms for embedded
08:15:03braewoodsmany of them require the whole thing to be in ram, not ideal...
08:15:24braewoodsothers have limits like this
08:15:25_bilgusso is it such a state where you could make due with a circular buffer
08:16:01_bilgusstart in buffer fill it use the remain for out then roll over if needed
08:16:01braewoodseh, probably if i could trick it into flushing the buffer instead of cannibalizing it
08:16:24braewoodsthe trouble is finding a way to make the inflate algorithm resumeable
08:16:30braewoodsor so
08:16:47braewoodsi either need to use an in memory one i can force refill or flush with as needed or
08:17:19braewoodsa resumeable one
08:17:42braewoodsjust tinflate as written only allows the input this luxury
08:17:59braewoodsi realized i could reset the output position between calls so reuse the same buffer
08:18:05braewoodsbut it would still discard stuff
08:18:22braewoodsunless i could size them such that overflow never happens
08:18:26braewoodsor gets discarded
08:18:29_bilgusha
08:18:32_bilgus:P
08:18:37braewoodsyea, fat chance.
08:18:43braewoodsi could handle the typical
08:18:45braewoodsbut
08:18:49braewoodsit's still just guessing.
08:19:12_bilguswell given the crc you could at least know when it failed
08:19:12braewoodsi would prefer same sized buffers and just flush when i run out
08:19:33braewoodswell it returns the actual size is would have written
08:19:39_bilgusah ok
08:19:42braewoodsyou can compare to know if it discarded
08:20:10_bilgusthat would be decent enough
08:20:26braewoodsi've been running out of options so this hack was looking like my only option
08:20:40braewoodsi was going to test out zopfli and see the best ratio i can get on some basic datais
08:20:42braewoodsis
08:20:46_bilgusits what makes embedded fun
08:20:56_bilgusor 'fun'
08:21:14rb-bluebotBuild Server message: Build round completed after 723 seconds.
08:21:19rb-bluebotBuild Server message: Revision 8846e087c0 result: 0 errors 180 warnings
08:21:26braewoodswhat's the warning
08:21:42braewoodsi mean where's the log
08:22:27amachronichttps://build.rockbox.org/dev.cgi
08:23:45braewoodsapp_modtime
08:23:54braewoodswhy is strlcpy not defined?
08:24:05braewoodshuh
08:24:10braewoodsanyway let me write app_modtime
08:26:48_bilgussim_modtime too
08:27:10_bilgusoh what was it with strlcpy uh strncpy
08:27:53braewoodsstrlcpy. it's different.
08:28:02braewoodsit's defined on native yet not these other?
08:28:07amachronicit's not quite standard afaik
08:28:22_bilgusidk I just rem thats what was in already
08:28:52braewoodsit's available on BSD
08:28:56braewoodsbut not Linux
08:29:12braewoodsi chose it because it's faster than snprintf
08:29:27braewoodsfor simple stuff
08:29:50braewoodswell one problem at a time
08:31:28_bilgusi think strn is more common than strl
08:31:50braewoodsit is... but it has some nasty gotchas
08:32:07braewoodsi'll look into a solution in a bit.
08:32:16braewoodsi just thought rockbox already has a solution to this issue
08:32:52_bilgusI think we just used n with some 'careful' handling
08:33:16braewoodsit also has one major inefficiency
08:33:30braewoodsunused parts of the string are zero padded beyond the null
08:33:34braewoodswhy is beynd me
08:34:25_bilguswe know why
08:34:37braewoodswhy?
08:35:48_bilgusto try and prevent unbounded strings
08:40:57braewoods g#3568
08:40:59rb-bluebotGerrit review #3568 at https://gerrit.rockbox.org/r/c/rockbox/+/3568 : filesystem: implement os_modtime for unix by James Buren
08:41:02braewoodsthis should fix some of those
08:41:28braewoodsit may also need win32, i'm not sure yet how important that is
08:41:43braewoodssince most dev work is unix based
08:50:35rb-bluebotBuild Server message: New build round started. Revision fa743258ea, 297 builds, 8 clients.
08:51:00braewoodsbut i think i found my way to do that
08:51:10braewoodsi'll see what this next round produces
08:51:56braewoodsif i convert the times back to dostime i can use the win32 api to convert it to the modern times it uses
08:52:13braewoodsi don't know what format it uses so this is probably the fastest solution
08:52:54_bilguswindows should have an api for unix time directly
08:53:19braewoodsit does? where i wonder
08:53:46braewoodsAll I found was SetFileTime()
08:54:32_bilgusah its just trivial, https://stackoverflow.com/questions/3585583/convert-unix-linux-time-to-windows-filetime
08:55:47_bilgusthat windows 100 ns comes up alot
08:56:13_bilgustheir audio API WASAPI uses that as the timing base
09:00
09:00:25***Saving seen data "./dancer.seen"
09:00:40sporkamachronic: i just wrote that i have had no sd card issues with the q1, but i have it too
09:01:35sporkon a 32GB samsung evo card
09:02:12sporkselecting 'start rockbox' from the bootmenu works fine afterwards
09:02:34amachroniclol
09:02:37rb-bluebotBuild Server message: Build round completed after 723 seconds.
09:02:40rb-bluebotBuild Server message: Revision fa743258ea result: 0 errors 145 warnings
09:02:43amachronicis it intermittent or does it happen every time?
09:04:56sporkintermittent
09:05:24sporkso far only after i have claimed not to have the problem
09:07:14 Join tomato5 [0] (~tomato@user/tomato)
09:07:38_bilgusah heisenbugs
09:08:03sporkschrodingers sd csrd
09:08:05amachronici wonder what is causing it... short of block-level trace and verifying each data transfer I have no idea how to debug it further
09:08:25 Quit tomato (Ping timeout: 256 seconds)
09:08:26 Nick tomato5 is now known as tomato (~tomato@user/tomato)
09:08:30amachronicit at least happens almost every time with my 256gb card.
09:08:39braewoods_bilgus: g#3569
09:08:41rb-bluebotGerrit review #3569 at https://gerrit.rockbox.org/r/c/rockbox/+/3569 : zip: include string-extra.h by James Buren
09:09:00sporkmaybe boost power during boot ?
09:09:03braewoodsno wait...
09:09:07sporkif there is such a thing
09:09:29amachronicweirdly, if I re-read the firmware twice, the problem goes away
09:09:42amachroniclike the first read fails but the 2nd read will succeed
09:09:55amachronicbut there's no low-level error from the 1st read...
09:10:41braewoodsthere we go
09:10:43braewoodsthat should do it
09:10:49_bilgusbraewoods, what are you using from string-extra.h?
09:10:56braewoods_bilgus: strlcpy?
09:11:04braewoodsthe source for it is compiled
09:11:09braewoodsi checked the build log
09:11:18braewoodsso
09:11:24braewoodsall i should need to do is include the header
09:11:33_bilgusstrlcpy.h
09:11:37_bilgus?
09:11:47braewoodsyea but everything else includes string-extra.h
09:11:53_bilgusah ok
09:12:28braewoodswell i stand corrected, some include i some don't
09:12:31braewoodsit's a bit weird
09:12:36_bilgusis anything in core doing that though?
09:12:45braewoodsno
09:12:55braewoodsonly string-extra.h headr
09:13:00_bilgusin core I'd think you'd only want to tie the v. minimum
09:13:39braewoodsin general though it makes no difference to runtime
09:13:43_bilguslet me look at the differences in implementation maybe we can bring in strlcpy to core and make it also strncpy
09:14:03braewoodserm, they work fundamentally differently
09:14:11braewoodsdifferent return types as well
09:14:28_bilgusdepends on how much code..
09:14:34braewoodsstrncpy will leave the buffer non-terminated if there's insufficient space
09:14:50braewoodsand it zero-pads
09:15:09braewoodsok... this was just the quick fix i found
09:15:42_bilguseh nm its not much code we can afford it
09:16:31_bilgusbut just do strlcpy.h we don't want to encourage all that other in core, well its already there so I used it
09:16:54braewoodswell i was just following what i saw.
09:16:56*braewoods shrugs.
09:17:04braewoodsas you wish
09:17:29_bilgusjust trying to save the next you :)
09:18:22braewoodsthere
09:18:48braewoods_bilgus: i'm considering adding fast versions of crc algorithms we have in core for ports that can afford to spend the space
09:19:05braewoodsat least in the main firmware. bootloader probably should stick to the space optimized onee.
09:19:06_bilgusnow that strlcpy is there I will be adding it to the damned plugin API :)
09:19:25braewoodsyou can also add modtime if you want
09:19:35braewoodsi'm not ready for the zip api to be added yet
09:19:37 Join dconrad [0] (~dconrad@208.38.228.17)
09:19:52rb-bluebotBuild Server message: New build round started. Revision 1298e47134, 297 builds, 8 clients.
09:20:31braewoodsi just try to optimize my string handling where i can do so; it's one of the most common recurring inefficiencies :D
09:20:35_bilgusprobably not modtime not general enough
09:21:32braewoodsone example, using information from previous string runs can save time
09:21:50braewoodsone thing i always hated about fgets is it gives no clue where the index of the newline is
09:22:05braewoodsso you need to scan it yourself to remove it if that matters
09:22:19braewoodsthat's about the only flaw of the function
09:22:34braewoodsone reason to use getline.
09:23:31braewoods_bilgus: how does one fix the simulator? i was assuming it would be using the OS apis.
09:23:56_bilgusi think it uses stubs?
09:24:46_bilgusFS_PREFIX(modtime)
09:25:03_bilgusyou need to use that macro
09:25:27_bilgusit adda app_ and sim_
09:25:32_bilgusadds*
09:26:46 Quit tomato (Read error: Connection reset by peer)
09:26:57braewoods_bilgus: that was added in the initial implementation
09:28:15_bilgusdid you also use it in the function def?
09:28:51braewoodsno, because that's only used in the header macro to change which one is used.
09:29:08braewoodsthe unix one is called os_modtime
09:29:14braewoodsthe native one is modtime
09:29:19braewoodsetc
09:29:48braewoodsi try to follow established usage because that's usually correct
09:29:50_bilguseh give me a few
09:31:21braewoodsjust grep for FS_PREFIX
09:31:36rb-bluebotBuild Server message: Build round completed after 704 seconds.
09:31:38rb-bluebotBuild Server message: Revision 1298e47134 result: 0 errors 57 warnings
09:31:39braewoodsit's not used in any source other than the plugin one so it the functions get mapped correctly
09:33:28braewoodsok looks like simulator is the last one to fix
09:35:46_bilgushttps://github.com/Rockbox/rockbox/blob/master/uisimulator/common/filesystem-sim.h
09:36:12braewoodsyea i see it
09:37:16_bilgusnow do you need to map it in the c file too IDK yet
09:37:59braewoodsi will unless i'm willing to reuse os_modtime
09:38:28braewoods... this is the weirdest while loop conditional i've ever seen
09:38:32braewoodswhile (8675309)
09:38:38braewoodsline 1147, dircache.c
09:38:56braewoodswhy 8675309? most use 1 or true for infinite loops
09:39:28_bilguswtf
09:39:46braewoodsit appears to be a reference to this
09:39:52braewoodshttps://en.wikipedia.org/wiki/867-5309/Jenny
09:41:07_bilgusha probably a way to see who has copied it
09:41:27_bilgusfound your number on a bathroom stall
09:42:00_bilgus8675309inine
09:42:19sporkcodepilot is calling jenny right now
09:43:17_bilgusif that thing could actually integrate disparate codebases itd be something, rn its a copypasta machine
09:43:34braewoods_bilgus: i can't see anything in here that would require us to wrap the os version.
09:43:48braewoodssince i don't open the file handle
09:44:00braewoodsbut i guess for convention we should
09:44:03_bilgusthe sim has no need for space considerations
09:44:23_bilgusidk your choice
09:44:43braewoodsworking on it
09:44:44_bilgusprobably more portable to lock in your own
09:45:08 Join tomato [0] (~tomato@user/tomato)
09:46:44braewoodslooks like it wraps it for mapping paths
09:48:56braewoods_bilgus: g#3570
09:48:58rb-bluebotGerrit review #3570 at https://gerrit.rockbox.org/r/c/rockbox/+/3570 : uisimulator: implement sim_modtime by James Buren
09:50:22rb-bluebotBuild Server message: New build round started. Revision d6dcb99684, 297 builds, 8 clients.
09:50:40braewoods_bilgus: i noticed crc32 is also used during normal runtime by rockbox by buflib. so there's some benefit to speeding it up.
09:51:10braewoodsdepends how often it's called but dynamic allocations tend to be common
09:51:16_bilgusin normal rb sure but be careful of bootloaders
09:51:38braewoodsindeed, i was planning to keep the slower version for a default and only enable a faster one where we know it's safe
09:51:46_bilgushave you compiled with test plugins yet?
09:51:55braewoodsnot yet
09:52:16_bilgusit'll allow you to quantify that
09:52:25braewoodsi'll be benchmarking it to see how much of a difference it makes
09:52:53braewoodsbut if the full lookup table that takes ~900 more bytes adds a significant boost
09:53:01braewoodswe should probably enable it for targets that can afford it
09:53:01_bilgusI do add stuff to those plugins when testing but rarely update them unless its warranted
09:53:30braewoodsand for a fair comparison i'll also need to cpu boost
09:53:31_bilguslike shortcutting to specific tests
09:53:58_bilgusmost of them have the option or do boosted/unboosted
09:54:34braewoodsi noticed that some BLs have no effective limit on space
09:54:39braewoodsnotably PP since it resides on disk
09:54:52braewoodson the regular FS with rockbox
09:55:03braewoodsso for those we could probably enable it under all circumstances
09:55:35_bilgusI try not to touch bootloaders without compelling reason
09:55:38braewoodsbut by default i plan to keep the SPACE optimized version unless we know otherwise
09:56:05braewoodsi generally don't but all changes made to core have this potential to impact them
09:56:23_bilgusindeed :)
09:56:41braewoodshopefully LTO will mitigate that by being able to omit unused code
09:56:49braewoodsremoved at link time
09:57:40braewoodswe still need a win32 modtime but it's pretty rare we build that
09:58:02braewoodsi'll work on it later when i can actually test it properly
09:58:08braewoodsi know unix well enough to know this is sufficient
09:58:21_bilgusits been so long that I have programmed in windows
09:58:32braewoodsi can make it work.
09:58:40_bilguseven used it for that matter
09:58:40braewoodsi'll just need to test in wine
10:00
10:05:43rb-bluebotBuild Server message: Build round completed after 921 seconds.
10:05:46rb-bluebotBuild Server message: Revision d6dcb99684 result: All green
10:07:52_bilguswoot
10:08:39braewoodsnow we can begin using it, for non-compressed ZIP files
10:09:08braewoodsthey say you can use zlib if you strip it...
10:09:15braewoodsbut is that really true?
10:13:42braewoodshuh with -Os, normaly zlib takes up 80KB
10:13:59braewoodsstill pretty fat
10:40:23rb-bluebotBuild Server message: New build round started. Revision 84362141a0, 297 builds, 9 clients.
10:52:28rb-bluebotBuild Server message: Build round completed after 725 seconds.
10:52:30rb-bluebotBuild Server message: Revision 84362141a0 result: All green
11:00
11:00:28***Saving seen data "./dancer.seen"
11:43:24dconradamachronic, I'm going to try out the bootloader, I think I should be able to use usbboot to try it out before writing it to nand, is that right?
11:43:58amachronicyeah try with USB boot and use the 'bootloader backup' first
11:44:12amachronicdid you try doing a nanddump from OF linux yet?
11:44:25dconradshould be stage 1 spl.erosq, stage 2 bootloader.bin?
11:44:29amachronicyep
11:44:30dconradno, I haven't tried that
11:44:44dconradwhats the process on that again?
11:45:31dconradI've been completely avoiding anything to do with bootloaders haha
11:45:35amachronicyou'd have to rig up a script to call nanddump
11:45:38amachronichttps://www.rockbox.org/wiki/ShanlingQ1Port
11:45:47amachronicwiki page has an example of nanddump script
11:45:49dconradoh I see
11:46:40amachronicidk can you run scripts from the "bootloader" on the Linux version of rockbox? like the m3k xvortex bootloader
11:47:14dconradhuh, it claims you can
11:47:23dconradI suppose I'll try that
11:47:42amachronicit's a very good idea, just in case I screwed up the flash code yet again
11:48:02dconradI mean at this point it seems like a low probability, but yeah safety first
11:48:30dconrad... apparently it has adb as well? that's neat
11:48:49amachronicplus, you need to know the NAND layout to figure out where you can flash the Rockbox bootloader.
11:49:27dconradoh, yeah there's no guarantee it's the same, is there
11:50:28amachronicgiven the eros q uses u-boot, I highly suspect the "empty space" that's normally used is occupied by u-boot.
12:00
12:01:42dconradit looks like the bootloader can't run those scripts correctly for some reason - I might try the method you list for the Q1
12:02:16dconradwhat's that magic a35c2bg0suw= file? where on earth does that string come from?
12:03:44amachronicoh, that comes from the Q1's firmware
12:04:04amachronicit's their super secret magic word :P
12:04:14dconraddarn
12:04:41dconradhmm, suppose I should try to figure out why the scripts aren't working in the bootloader then
12:04:42amachronicwell are you able to run ANY scripts? it's possible you're missing the nanddump command
12:04:58dconradthey claim they run, but just return "2" immediately
12:05:14dconradI wonder if I can echo stuff to the screen
12:05:29amachronicthat won't work but you can echo foo > /path/to/file
12:08:23dconradyeah, even that doesn't seem to work. I wonder if I have the wrong path to the sd card...
12:14:33 Quit Natch (Remote host closed the connection)
12:17:33dconradoh, is this a permissions thing I wonder
12:18:24dconradI mean... it's a fat formatted card, and the rockbox binary obviously runs from it, so it shouldn't have any issue, right?
12:19:24amachronicit seems the firmwares tend to mount FAT with RWX permissions, but maybe yours is an exception
12:20:50amachronicusing ext2/3/4 might work
12:21:06dconradyeah, I might have to try that
12:21:57dconradpull another card out of the pile haha
12:32:56dconradhmm, it actually looks like it doesn't support ext3...?
12:33:25dconrador, at least it can't read the files I put on there
12:37:18dconraddconrad.neocities.org/assorted_files/helloworld.sh">https://dconrad.neocities.org/assorted_files/helloworld.sh, https://dconrad.neocities.org/assorted_files/kernelmessages.sh, and https://dconrad.neocities.org/assorted_files/nanddump.sh
12:38:07dconradspeachy, do you know anything about running scripts on the eros q bootloader? any gotchas?
12:38:31dconradI'm not actually sure what the path to the card should be either
12:40:51amachronicif all else fails, you could build a custom rockbox and call system() to run each command
12:42:38dconraddoes rockbox run as root? I think it would need to, right?
12:43:05amachronicit should
12:44:31dconradhm, that might be the thing to do then
12:47:09 Join Natch [0] (~natch@c-e070e255.014-297-73746f25.bbcust.telenor.se)
13:00
13:00:29***Saving seen data "./dancer.seen"
13:02:59dconradaha, it's located at /mnt/sd_0/
13:10:58dconradthere we go, now we're makin progress
13:11:10dconradgot the rootfs and dmesg dumped
13:15:52dconradamachronic, are there really 9 /dev/mtd*? this only seems to list mtd0 - mtd5 in my filesystem listing
13:19:26dconradI would have figured for the same chip, it should have the same number of mtd*
13:28:40dconradtotal size of 128 MiB
13:32:43 Quit jschwart (Read error: Connection reset by peer)
13:33:00 Join jschwart [0] (~quassel@2001:985:2c6e:0:b00b:32ff:fe28:5567)
13:50:16 Join mendel_munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
13:50:47amachronicdconrad: the number of mtd partitions is all up to the OF
13:51:09dconradok, just seems weird it wouldn't match, given its the same hardware
13:51:38dconradbut that's all that's there, so I suppose it should be correct
13:51:59amachronicit's not much different to partitioning a hard disk, it's all software. As long as what you got adds up to 128MB it has to be correct.
13:52:11dconradok, cool
13:53:23 Quit munkis (Ping timeout: 268 seconds)
13:57:06dconradthanks for the review btw, I think I'll get that stuff changed since it's all small stuff, and then try to rebase/get a bootloader working
14:00
14:22:17sporkamachronic: that yes/no thing works ok everywhere for me, except the prompt to create a bookmark
14:22:25sporkwhere else does it not work for you ?
14:24:17amachronicat least the first 'create database' prompt, I had it a handful of other times. I think it may be due to some absolute pointing / 3x3 mode conflict.
14:24:45amachronicmaybe your problem is different then, like a queued button or touch.
14:24:52sporkthe yes/no thing has its own logic
14:25:06sporkbottom half of the screen and left/right half for yes and no
14:25:28sporkpretty much like those pretty rectangles it draws
14:25:31amachronicyeah, and I had a very hard time pressing *either* of the buttons, which is a bit opposite to your issue.
14:26:02sporkok, i will poke it a bit. gui stuff at least i understand
14:26:27amachronicanyhow I uploaded an 'official' bootloader and I'm going to work on some other odds and ends so this can finally be merged
14:27:31sporkuploaded to be used in install instructions later ?
14:27:54amachronicnot a binary yet, just the source
14:28:04sporksaw that commit yes
14:29:02sporkjztool needs some help before that can be used
14:29:25sporkbut i got your point about non-official bootloaders
14:30:43amachronicthat's also because I am trying to maintain a semblance of forward and backward compatibility.
14:31:09amachronicso daily builds don't suddenly break with an older bootloader, or vice versa
14:32:22amachronicstill, the worst that would happen is you need to use jztool again to reinstall so it's mostly to avoid a nuisance to everyone.
14:33:24sporkmine boots, usually well, so i am quite happy as it is
14:35:26amachronicglad to hear it
14:36:02sporkbeen playing jewels way too much
15:00
15:00:32***Saving seen data "./dancer.seen"
15:19:13 Quit amachronic (Quit: amachronic)
15:30:17 Join mendel_munkis_ [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
15:30:25 Nick mendel_munkis_ is now known as munkis (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
15:33:10 Quit mendel_munkis (Ping timeout: 252 seconds)
15:35:53 Quit f1refly (Quit: see ya in hell)
15:37:43 Join f1refly [0] (~f1refly@pd954c410.dip0.t-ipconnect.de)
15:53:47 Quit dconrad (Remote host closed the connection)
15:57:50 Join dconrad [0] (~dconrad@208.38.228.17)
16:00
16:02:37 Quit dconrad (Ping timeout: 268 seconds)
16:11:26braewoodsspeachy: the front page says the HD300 is stable but the status page doesn't.
16:11:28braewoodswhat gives?
16:30:58 Join GeekShadow [0] (~antoine@82-64-164-139.subs.proxad.net)
16:43:25sporkthat is not unique, zen mozaic is in the same boat
16:44:56sporkHD200 too
16:45:15sporksorry, not the HD200
16:45:29braewoodsi managed to find an HD300 on ebay
16:45:45braewoodsthe mpio stuff is pretty rare
16:48:02sporkcrazy price ?
16:49:00braewoodsno, like $60
16:49:10braewoodsthe mpios aren't in high demand nor desired
16:51:21sporkit looks pretty nice, but it is seriously old now
16:51:36braewoodsi mainly bought it to test the port
16:55:09sporkthat is dedication
17:00
17:00:35***Saving seen data "./dancer.seen"
17:02:45speachybraewoods: The HD300 has been marked as "stable" in the build system since 2012.
17:03:09braewoodsspeachy: i see. the wiki is weird then.
17:03:19speachywe're all only human.
17:07:23 Join dconrad [0] (~dconrad@208.38.228.17)
17:12:14 Quit dconrad (Ping timeout: 272 seconds)
17:16:06 Join skipwich [0] (~skipwich@user/skipwich)
17:41:31 Join jadzia [0] (~jadzia@2604:3d09:1b79:ca00::1e45)
17:55:33 Quit lebellium (Read error: Connection reset by peer)
18:00
18:02:54 Quit akaWolf (Ping timeout: 272 seconds)
18:21:08 Join akaWolf [0] (~akaWolf@akawolf.org)
18:53:43 Quit ZincAlloy (Quit: Leaving.)
19:00
19:00:37***Saving seen data "./dancer.seen"
19:17:31 Join dconrad [0] (~dconrad@208.38.228.17)
19:25:10 Join cockroach [0] (~blattodea@user/cockroach)
19:26:02 Join Solanacean [0] (~solanacea@95.78.202.240)
19:32:35 Quit Solanacean ()
20:00
20:31:01 Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
21:00
21:00:39***Saving seen data "./dancer.seen"
21:55:53 Quit dconrad ()
22:00
22:21:57 Quit cockroach (Quit: leaving)
23:00
23:00:41***Saving seen data "./dancer.seen"
23:15:00 Join Retr0id9 [0] (~Retr0id@user/retr0id)
23:15:54 Quit Retr0id (Ping timeout: 252 seconds)
23:15:54 Nick Retr0id9 is now known as Retr0id (~Retr0id@user/retr0id)
23:47:38 Quit massiveH (Quit: Leaving)

Previous day | Next day