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 2016-10-17

00:02:14 Quit ungali (Quit: ungali)
00:32:55 Quit paulk-collins (Remote host closed the connection)
00:34:17 Join RBNB [0] (6d4b5cbb@gateway/web/freenode/ip.
00:35:18RBNBHey guys! Does anybody happen to know how I can access the metadata log? I have activated it from System > Debug, but I cannot find the actual log for checking. I've got problems with my database and need to find the file that causes them.
00:35:55[Saint]it'll be sitting inside the /.rockbox directory
00:36:18[Saint]with, from memory, a fairly obvious name.
00:36:42RBNBI've got my device connected to my PC and searched through .rockbox up and down, there's absolute nothing in that direction
00:37:28[Saint]Have you actually init'ed the database since?
00:37:41[Saint]just enabling the feature in and of itself will do nothing.
00:38:17RBNBLet me think....I used the import modifications function, but now that you say it, it's indeed possible that I didn't re-initialize
00:38:41RBNBSo without the database being re-initialized, this file won't exist?
00:39:48RBNBAre there any other possible reasons why this file won't exist? maybe the procedure crashing in a very weird way?
00:40:20[Saint]Massive filesystem corruption is one way.
00:40:33[Saint]Another would be if the init process never even began.
00:41:11[Saint]It isn't really an external feature, it just adds logging to the DB process that is already in place.
00:41:49[Saint]Basically all it does is print off a list of files as they're encountered, and if it screws up, the last file in the list is the file immediately before wherever it fell over.
00:42:20RBNBThat's what I understood from the documentation, good to know
00:42:59RBNBI'm now trying it again, consciously initializing the database and selecting the import function again
00:43:21RBNBI wouldn't be surprised if there was massive filesystem corruption, but i'll just pretend there isn't
00:48:44RBNBJust waiting for the parser to crash, might take some time, haha ^^
01:02:28RBNB[Saint], considering that it is indeed one particular file that causes the parser to crash...what will solve the issue? Deleting the file from the device, editing the database_changelog.txt, or something else?
01:03:46[Saint]Deleting the file would indeed 'solve' the issue. But it would be more interesting to understand why it is failing (it is more than likely corrupt or has corrupt metadata) rather than ignoring it.
01:05:15RBNBHow could I diagnose what causes the issue?
01:06:24 Quit ender` (Quit: If you believe in the Christian story of Jesus, then you can't call Scientology crazy. — Neil deGrasse Tyson)
01:09:15RBNBThe strange thing is that I had transferred the database in the past and it worked without problems (most of the time). Do you know in what order the parser works through the files? Alphabetically, date added, something else?
01:12:22[Saint]unnatural alphanumeric sort.
01:14:55 Quit girafe (Read error: Connection reset by peer)
01:20:34RBNBAlright, procedure finished, once again precisely 2741 files without metadata
01:21:16RBNBand once again NO metadata log to be found anywhere on the device
01:24:06RBNBDoes the metadata log need to be enabled specifically before a database initialization and otherwise it's automatically disabled?
01:24:28RBNBBecause I can't find any reason why the log still appears nowhere on the device
01:26:51[Saint]Make sure it is enabled and reboot the device, and then give it a couple of minutes to make sure that it has finished trying to update the db.
01:29:36RBNBAlright, now the log is there
01:30:00RBNBBut it seems to list all files on the device, so where the hell is the problem now
01:32:51RBNBAre there any other issues that could cause the database import to stop at some point? A missing file maybe?
01:33:41[Saint]A missing file, where? Rockbox ships with everything it requires and it doesn't give a shit about the content or arrangement of your media.
01:34:45[Saint]and if the metadata log lists all the media on the device, and you are confident of this, it means one thing and one thing only - that the database addressed all that media.
01:35:13[Saint]it may be possible that it fell over and couldn't do it all in one lump sum and completed it on the second round.
01:36:01[Saint]I must ask, how do you know there is media missing? How have you verified this?
01:36:25RBNBI have created 32 "numerated" subfolders and, based on what you said earlier about the alphanumeric order, considering that the last file in the metadata log file is the last audio file in folder #32, I assumed that it lists all media.
01:36:34[Saint]Just a plain file count from the toplevel directory isn't going to be indicative for a number of reasons.
01:37:13RBNBMissing media is well possible since sometimes, when I notice a file duplicate, I delete it from within Rockbox via the context menu
01:38:04***Saving seen data "./dancer.seen"
01:38:17[Saint]What I mean is how did you form the opinion that the database init actually crashes and that you have media missing to begin with?
01:38:22[Saint]How was this verified?
01:38:50RBNBAh, sorry about that, this was posted in the thread on the forums and I forgot to mention that here
01:39:39RBNBI use the RATING feature heavily (I know, I know, barely anyone does), and when I exported the database from my other device, only 110 files out of around 7500 were not rated by me yet
01:40:14RBNBAfter importing the database now, it (always) leaves exactly 2741 files without a rating, so I assumed that there was some problem with the parser
01:41:00RBNBsince the number is always the same, I also assumed that it is one and the same file causing trouble everytime
01:43:26[Saint]if the metadata log is listing what you know to be the first and last (anything in the middle is inconsequential for this purpose) files in unnatural alphanumeric sort order then the parser is functioning as intended.
01:45:09RBNBAww crap, I was afraid you would say that, since it leaves me without any clue what is causing the problem here
01:45:17[Saint]I honestly can't speak for why the ratings would or wouldn't be being applied. You're about the third person I have ever talked to that uses the Rockbox rating system since it is such a pain in the tits to use and is proprietary.
01:46:11RBNBI also wish they weren't proprietary, but since it's the best RB got, it's my only choice
01:47:00RBNBI got used to it so much that I couldn't renounce on the feature anymore, as stupid as it might sound
01:47:09[Saint]It is technically possible but at present we have no means to edit media metadata.
01:47:24[Saint](and no one has expressed any desire to change this)
01:48:36[Saint]I would think a more versatile and robust solution would be for you to "rate" your media using playlists.
01:49:46RBNBI've never tried that to be honest...I guess those playlists can be exported somehow?
01:49:52[Saint]Honestly the idea of ratings is kinda weird to me.
01:50:00[Saint]They can, yes.
01:50:22[Saint]If you're rating media as 1-star or whatever...why is it even on your device?
01:50:43[Saint]I dunno about you, but I put music I actually like on my DAPs.
01:51:16[Saint]If I have to consider rating it to order it by preference I don't think it should be on my device to begin with.
01:51:26[Saint]But...that's me.
01:51:27RBNBI know it will be difficult to get an idea across that started years ago...and which I continued also because, after 7500 files, it would be a neverending work to start from scratch again
01:52:17[Saint]Oh. I understand. This is more about maintaining a collection and knowing something is broken than it is about having a real functional benefit.
01:52:34RBNBYou got a good point, that's why my ratings start at 6 - which is, for me, a track that is just solid enough so I listen to it at least occasionally. This goes up to 10 which is attributed to my absolute favourites
01:53:08RBNBThe procedure of sorting out bad tracks, for me, happens even before it ends on the device.
01:53:13[Saint]isn't that just another way of having a 5 star system by another name?
01:54:08[Saint]I would think though, personally, that it would be a lot faster and more robust to do this by playlists though personally.
01:54:34RBNBWell, I guess if you dissect the idea you will definitely find many ways how it could be improved. I could start these ratings at 1 star, just considering that "1 star" does not mean "bad", of course
01:54:46[Saint]Massive bonus being that you can include full directories and subdirectories into a playlist instead of addressing each file on a per-file basis.
01:55:07[Saint]Which must have taken you an unreasonably long time.
01:56:13RBNBYes it did. Actually, I'm adding folders of around 200 tracks each month, and whenever I listen to a track, I attribute a rating - because to give a rating, I need to listen to it anyway
01:56:55RBNBI really dig the idea how playlists might work (assuming that the process works easier, or at least, not even more tedious than the ratings)...but I would basically have to do it per-file again and start all 7500 tracks from scratch
01:57:05[Saint]I suppose that's a difference between yourself and myself, I put media on my device I _know_ I like as I had reviewed it prior.
01:57:49 Quit krabador (Quit: Leaving)
01:57:57RBNBWell, so do I, it's just that I just try to differentiate between my absolute favorites and tracks I just "like"
01:58:05[Saint]I'm generally comfortable with full albums though and I just use skip if there's every a track that pops up that I'm not into.
01:58:55[Saint]The difference with doing this with playlists is it makes it slightly easier to do in bulk.
01:59:10RBNBBased on the fact that there's only electronic/club music on the device, there's no albums anywhere since the genre is characterized more by single tracks than full albums
01:59:17[Saint]If there's a full album you know you like all of it, you can just impor tthe entire directory into a playlist.
02:00:10[Saint] your media isn't sorted by */artist/album/disc/ ?
02:00:47[Saint]Most of us do it that way so that the file browser ends up being just as useful as the database for most purposes.
02:00:59[Saint]Honestly, most of us don't even use the database.
02:01:48RBNBI know, but for the music I use on this device (with only single tracks because of the genre), that's the core of the problem for the playlist approach. It's sorted like this: 1) Files added in Sep 2015 2) Files added in Oct 2015 3)...
02:02:11[Saint]I honestly can't think of a single contributor or volunteer offhand who doesn't use playlists and a 'sane' direcotry structure with the file browser as opposed to the database.
02:02:28RBNBSorting the 7500 tracks by artists would lead to about 5000 different artists
02:02:28[Saint]Most of us find the 8~10MB the DB consumes to be more useful as audiobuffer.
02:02:35[Saint]But...different strokes, I guess.
02:03:02[Saint]RBNB: sure, but there's things that will automate that. It's not like you need to do it by hand.
02:03:46RBNBChanging my folder structure would just cause a truckload of work on another end...because sorting 200 tracks each months by artists folders would have me create at least 180 new folders because it's just single tracks
02:03:48fs-bluebot_Gerrit review #200 at : Fundamentally rewrite much of the audio DSP. by Michael Sevakis
02:03:49[Saint]makes it a lot easier to find a particular track when you're not using Rockbox, too.
02:04:07CtcpIgnored 6 channel CTCP requests in 5 days and 0 hours at the last flood
02:04:07*[Saint] slaps fs-bluebot_
02:04:07fs-bluebot_[Saint]: ouch!
02:04:17[Saint] like that? You little bitch...
02:04:44*[Saint] slaps fs-bluebot_ harder
02:04:44fs-bluebot_[Saint]: ouch!
02:04:49[Saint]Damn straight.
02:05:42[Saint]I really wish I could be more helpful than "I don't know why ratings are being applied and you're pretty much the only person in existence I am aware of that uses the DB in this fashion".
02:06:03[Saint]*aren't being applied, I mean.
02:06:47[Saint]You could be running up against some form of internal limit I suppose.
02:07:03[Saint]It really isn't intended that people have several thousand files under a single directory.
02:07:23[Saint](from discussion I presume this is your structure)
02:07:26RBNBI'm open to new suggestions, even if that means that several years of building a rating database were in vain...but the way I see it, for my listening habits with a music genre not typically being characterised by full albums, using playlists would cause similarly tedious work
02:08:02[Saint]How is your media organized, filesystem-wise?
02:08:19[Saint]Is it, as I think, several thousand files sitting under a directory?
02:08:24RBNBAt the moment the structure is like this: one subfolder with maybe 1000 tracks, and 31 subfolders with around 200 each, with one new subfolder being added per month.
02:08:45RBNBthere will never be several thousand files in one single directory
02:09:40RBNBSo the overall sorting structure is basically by date added.
02:10:23RBNBThat's the only thing the tracks added have in common, because every month, there will be tracks with different rating, different artists etc.
02:11:39 Quit ej0rge_ (Remote host closed the connection)
02:11:39[Saint]has the structure of this changed significantly or at all between when the db was exported and attempting to import it?
02:12:12RBNBNo, I have copied the entire structure, as it was on the other device, over to this new one. That's what makes this error seem so damn strange
02:12:46[Saint]We use relative paths so it /should/ be safe even if it has, but I suppose it is possible this could fail.
02:13:07[Saint]You still have that build I made for you from ages back from before the filesystem rework?
02:13:15[Saint](I presume you're the same guy)
02:14:00[Saint]Apologies if you're not, my memory is...prescription drug clouded.
02:14:28RBNBYour memory is 100% accurate
02:15:07RBNBI still have that build, and I tried both the current build from RB utility as well as that build (with Load to RAM) for the import process - both with the same result
02:15:43RBNBThe database would always end up with 2741 unrated tracks although only 110 should be unrated
02:18:14[Saint]I don't want to screw up your media but I'm really curious as to whether export/import works across the same build and device - we know it fails for you if you export to a different build/device, but it would be interesting to know if it worked as intended in both current and legacy builds.
02:20:01[Saint]But if the directory structure hasn't changed at all it really shouldn't be the (very minor) changes to the way relative paths are handled that I was thinking it might be.
02:21:24 Quit Rower (Ping timeout: 250 seconds)
02:21:44 Join Rower [0] (
02:22:11[Saint]This is very curious. But I haven't the foggiest notion where to look to debug this.
02:23:22[Saint]It would give a nice starting point if the failure was identical when trying to export/import the DB without changing the build or device though. But that risks further screwing up your media.
02:32:03 Quit ZincAlloy (Quit: Leaving.)
02:49:49 Quit idonob (Ping timeout: 265 seconds)
03:07:54 Join idonob [0] (
03:18:34 Join fs-bluebot [0] (
03:20:30 Quit bluebrother (Ping timeout: 250 seconds)
03:21:14 Quit fs-bluebot_ (Ping timeout: 265 seconds)
03:22:38 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
03:38:08***Saving seen data "./dancer.seen"
04:07:46 Quit prof_wolfff (Ping timeout: 258 seconds)
04:20:52 Join prof_wolfff [0] (
05:38:09***Saving seen data "./dancer.seen"
06:55:28 Quit TheSeven (Ping timeout: 245 seconds)
06:56:17 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:58:38 Quit cc___ (Quit: WeeChat 1.6)
07:06:02 Quit alexweissman (Remote host closed the connection)
07:08:46 Join alexweissman [0] (
07:38:10***Saving seen data "./dancer.seen"
08:00:55 Quit saratoga_ (Ping timeout: 260 seconds)
08:17:20 Join ender` [0] (
08:30:37SlasheriIf all metadata is included in the file created during database export, then it's likely that there is a difference in file paths/names between the metadata export and the actual filesystem so the import procedure can't find the matching database entry while reading the metadata file
08:32:58SlasheriIt should be also possible to transfer the database as a binary DB from device to device as long as the filesystem layout and DB version remains the same. It doesn't matter if there is endianess mismatch between devices, database system does endianess correction on the fly if needed
08:38:35 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
08:38:35 Quit pixelma (Quit: .)
08:41:13 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
08:41:15 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
08:43:08 Quit alexweissman (Remote host closed the connection)
08:49:37RBNBSorry [Saint], I fell asleep haha
08:51:38RBNBGotta quit for today, but it seems like some of you guys have some good ideas as far as that database problem is concerned. Please post in my forum topic,51510.0.html
08:52:01RBNBI'll then try what you suggest for finding the problem/debugging
08:54:50 Part RBNB
09:38:12***Saving seen data "./dancer.seen"
09:40:53 Join petur [0] (~petur@rockbox/developer/petur)
09:53:44 Join alexweissman [0] (
09:58:28 Quit alexweissman (Ping timeout: 260 seconds)
10:00:53 Join elensil [0] (~edhelas@2001:1c02:1903:d800:b5de:2c1:417f:77ea)
10:24:13 Join MrZeus [0] (~MrZeus@
10:37:55 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:9990:ddd4:fe55:1c2f)
10:45:25 Quit MrZeus (Quit: Leaving)
10:45:39 Join MrZeus [0] (~MrZeus@
10:47:45 Quit MrZeus (Client Quit)
10:52:10 Join MrZeus [0] (~MrZeus@
11:10:57 Join einhirn [0] (
11:38:13***Saving seen data "./dancer.seen"
11:46:44 Join krabador [0] (~krabador@unaffiliated/krabador)
12:12:36 Quit ZincAlloy (Quit: Leaving.)
13:38:15***Saving seen data "./dancer.seen"
14:07:07 Join rogeliodh [0] (~Thunderbi@
14:20:13 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
14:41:41 Quit krabador (Quit: Leaving)
14:43:42 Quit APLU (Ping timeout: 258 seconds)
14:55:28 Join APLU [0] (
15:10:00 Join alexweissman [0] (
15:14:48 Quit alexweissman (Remote host closed the connection)
15:33:27 Join amayer [0] (
15:38:19***Saving seen data "./dancer.seen"
15:46:54 Join alexweissman [0] (
16:08:55 Quit krnlyng (Ping timeout: 260 seconds)
16:20:37 Join krnlyng [0] (
16:36:15 Join krabador [0] (~krabador@unaffiliated/krabador)
16:42:14duo8is it possible to use a rockbox dap as a dac?
16:42:26duo8possible to implement such a feature i mean
16:44:48elensilduo8: pamaury was working on a USB-audio implementation in Rockbox
16:45:35duo8is it hardware dependant?
16:50:17elensilping pamaury
16:57:32pamauryduo8: usb drivers may need some adaptation, but currently is has problems and I haven't worked on it recently
16:58:51pamauryalso some usb controllers don't have the necessary hardware features to support it
17:03:35duo8is the usb controller a separate part or is it in the soc?
17:10:34pamauryusually it's in the soc
17:35:50 Quit petur (Quit: Connection reset by beer)
17:38:22***Saving seen data "./dancer.seen"
17:38:28 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:285f:16de:fdb4:fb6a)
17:51:32 Quit passstab (Ping timeout: 258 seconds)
18:04:29 Quit krabador (Quit: Leaving)
18:06:56duo8can the jz4760 do it?
18:07:10duo8since the fiios can i assume it's possible
18:08:34 Quit elensil (Quit: Leaving.)
18:10:11pamaurythe datasheets says yes
18:22:31 Join petur [0] (~petur@rockbox/developer/petur)
18:30:32 Join girafe [0] (
18:32:55 Quit rogeliodh (Ping timeout: 256 seconds)
18:38:15 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:55:27 Join robertd1 [0] (~root@
19:02:44 Join cc___ [0] (~ac@2001:910:113f:1:6a05:caff:fe1c:1627)
19:25:24 Join krabador [0] (~krabador@unaffiliated/krabador)
19:26:05 Join lebellium [0] (
19:29:04pamauryrobertd1: hi
19:29:23robertd1hi pamaury!
19:31:09pamauryI more or less finished disassembling appproxy and it confirms what I thought. I think sysmgrd launches the applications using appproxy and it has the capability to stop them by sending a SIGTERM signal. Disassembling sysmgrd would be a difficult task so tomorrow I'll try something on my device by replacing the main app by another program to see how sysmgrd handles usb and if it's safe
19:38:24***Saving seen data "./dancer.seen"
19:39:28robertd1pamaury, is that the same handler that removes the lock files? How are you dissasembling sysmgrd?
19:40:05pamaurywhich lock file? Using IDA, but sysmgrd is a huge program, it seems to do a lot of things
19:45:28robertd1pamaury it seems that during an upgrade the nvp app unlocks temporaly the filesystem. now I dont remember which app verifies the destination and country of the device then it loads
19:49:35 Join paulk-collins [0] (
19:49:36pamaurythe upgrade uses a different system, it runs from initfs
19:50:24 Join passstab [0] (
20:28:10 Join TheLemonMan [0] (~root@unaffiliated/thelemonman)
20:40:35 Quit pamaury (Ping timeout: 260 seconds)
20:53:08 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
21:21:54 Quit robertd1 (Ping timeout: 256 seconds)
21:38:27***Saving seen data "./dancer.seen"
21:38:30 Join rogeliodh [0] (~Thunderbi@
21:52:47 Join robertd1 [0] (~root@
21:55:10 Join Sansa-Clip [0] (55ce9d6f@gateway/web/freenode/ip.
21:55:55Sansa-Clipany new on sheduled FM recordings?...
21:56:52Sansa-ClipI just found that Sansa Clip in mono mode performs cleaner recording that CCrane Witness+
22:24:14 Quit petur (Quit: Leaving)
22:25:24 Join Link8 [0] (
22:25:52 Quit Link8 (Remote host closed the connection)
22:40:03 Quit tchan (Ping timeout: 260 seconds)
22:42:09 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
22:44:10 Join girafe2 [0] (
22:47:31 Quit girafe (Ping timeout: 260 seconds)
22:54:40 Quit atsampson (Ping timeout: 258 seconds)
22:55:33 Join atsampson [0] (
23:03:59 Quit Sansa-Clip (Quit: Page closed)
23:07:36 Quit girafe2 (Read error: Connection reset by peer)
23:14:12 Quit TheSeven (Ping timeout: 245 seconds)
23:27:28 Quit lebellium (Quit: ChatZilla 0.9.92 [Firefox 49.0.1/20160922113459])
23:30:42 Quit paulk-collins (Quit: Leaving)
23:31:54 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
23:38:31***Saving seen data "./dancer.seen"

Previous day | Next day