#260416 - 18/07/2005 21:35
Re: Bugs in V3 A10
[Re: SE_Sport_Driver]
|
peter
carpal tunnel
Registered: 13/07/2000
Posts: 3349
Loc: Cambridge, England
|
Quote: no room for private writable mapping
Your player is running out of memory. I never tested the alphas (or 2.01 for that matter) on players as big as 48Mb, and it may be that the policy of giving the entire difference in memory over to the music cache, is not ideal. (Perhaps the kernel needs more memory to keep its pagetables for such a big system, or perhaps things such as TCP/IP award themselves bigger buffer sizes if run on 48Mb players.)
See whether adding a "reservecache=20" line, or something similar, in config.ini makes any difference. This leaves 20 cache chunks (or 1280K) unallocated, so any greedy TCP/IP or pagetable that comes along, is welcome to use it.
Peter
|
|
Top
|
|
|
|
#260421 - 19/07/2005 08:13
Re: Bugs in V3 A10
[Re: SE_Sport_Driver]
|
Roger
carpal tunnel
Registered: 18/01/2000
Posts: 4554
Loc: London, UK
|
Quote: BTW, I love the progress % for the database build! Great feature!
Is that a v3-only feature? Wow. It's been so long since I added that: I thought it was in v2.x as well...
|
|
Top
|
|
|
|
#260422 - 19/07/2005 08:17
Re: Bugs in V3 A10
[Re: schofiel]
|
Roger
carpal tunnel
Registered: 18/01/2000
Posts: 4554
Loc: London, UK
|
(Note: this is a v3a8 bug, and I've not tried v3a10 yet, but since Dave was focussed on the tuner code, I don't think he'll have fixed it. If someone gets the chance to reproduce it before I install v3a10, please post your results here)
Using Tweak Order can confuse/crash the player:
If you have the player shuffled, and you decide you'd like more of the current artist, you hit the tweak button on the remote. If you decide you'd like lots more, you keep hitting it until it says "no match found".
At this point, if you hit track skip, it doesn't -- the screen changes to the next song (and the next song), but it never actually skips track.
You need to reboot the player to fix it.
If the position in the first track is past the end of the next track (IYSWIM), this action can also cause the player to reboot.
|
|
Top
|
|
|
|
#260424 - 19/07/2005 18:59
Re: Bugs in V3 A10
[Re: schofiel]
|
schofiel
carpal tunnel
Registered: 25/06/1999
Posts: 2929
Loc: Cambridge, UK
|
Guys, could we keep the discussions to a minimum and keep a concise list of bugs posted here. If you want to discuss the behaviour, problems, or whatever, start another thread. Don't just sit here spinning gas about observations: at some point, someone is going to have to pick through a bug list and if this thread is populated by 10,000 pointless observations then it's going to be a bit difficult, no?
So from now on: concise bug descriptions of confirmed bugs ONLY, with a description of the stimulus, and the symptoms, with an indication of repeatability. The areas of interest are focussed around the tuner functionality.
One of the few remaining Mk1 owners... #00015Stop CO2 Emissions Now! - calculate your waste! Methane is killing us Slowly
|
|
Top
|
|
|
|
#260425 - 19/07/2005 21:31
Re: Bugs in V3 A10
[Re: schofiel]
|
AndrewT
old hand
Registered: 16/02/2002
Posts: 826
Loc: Oxford, UK
|
1. Unable to perform station searches using the L & R buttons (remote & front panel).
2. Impossible to disable AF. The AF display symbol reflects the menu selection but AF is still always active.
3. AF changes frequency to 98.0MHz every few minutes. This appears to happen while tuned to BBCR1 (99.1MHz) and BBCR4 (93.9MHz), I haven't observed this behaviour on independent stations.
I mostly listen to radio in the middle of the south-central area shown here - see the light blue 99.1MHz patch. With BBCR1 for example, it always switches from 99.1MHz to 98.0MHz within a few minutes. Meanwhile, the AF debug display permanently displays the "WAIT" screen. Changing the preset station away and back again corrects it for anything up to 2 minutes, rinse and repeat.
I cannot 100% confirm this but I think that in areas of weak reception, the 98.0MHz switching tends not to happen.
Edited by AndrewT (20/07/2005 01:08)
|
|
Top
|
|
|
|
#260432 - 22/07/2005 00:54
Re: Bugs in V3 A10
[Re: schofiel]
|
AndrewT
old hand
Registered: 16/02/2002
Posts: 826
Loc: Oxford, UK
|
AF occasionally causes the empeg to jump to a different station.
Perhaps this is caused by interference from a local station or possibly this is because TA is always enabled (it is currently set to 'Off' on my player) regardless of the menu selection (just like AF) - I'm not sure? Local radio stations have a bad habit of leaving their TA beacon active both before and after the TA to sucker you into listening to their ad's.
I'm struggling to think of any useful data that I can provide here to backup my claims in this thread, please feel free to ask if there's anything worthwhile that I should add.
Edited by AndrewT (22/07/2005 00:59)
|
|
Top
|
|
|
|
#260433 - 22/07/2005 13:44
Re: Bugs in V3 A10
[Re: pim]
|
peter
carpal tunnel
Registered: 13/07/2000
Posts: 3349
Loc: Cambridge, England
|
Quote: Tunes are marked unintentionally, randomly.
This is a feature, not a bug, but it's clearly unpopular enough to remove. The markings are not random, they indicate when the track ended earlier than it indicated it would; often, this is the result of a corrupt MP3. However, if the marking is as overzealous as some claim, perhaps there are bugs in its determination of early ending. FITNR.
Quote: /drive0/var/database3 is read with 512 bytes per read() system call. A larger buffer would speed up the database load.
The actual problem was the dynamic data partition, not database3 (which is read with a single system call). FITNR.
Quote: Version 3 of the player uses utf-8 internally, but the used fonts lack japanese glyphs.
FITNR (for the larger font; it's not really possible to display Japanese in the smaller font).
Quote: Play count display cannot be suppressed.
FITNR.
Quote: Unable to perform station searches using the L & R buttons (remote & front panel).
FITNR.
Peter
|
|
Top
|
|
|
|
#260434 - 22/07/2005 14:10
Re: Bugs in V3 A10
[Re: peter]
|
Roger
carpal tunnel
Registered: 18/01/2000
Posts: 4554
Loc: London, UK
|
Quote: ...they indicate when the track ended earlier than it indicated it would; often, this is the result of a corrupt MP3...
I always suspected that -- more often -- this is the result of whichever version of emplode had a rounding error when calculating the length of the track. Theoretically, if you recalculate the length tag, it'd go away.
The problem is that the "marked" flag should be reserved for the user, or should differentiate between the user setting it, or the player setting it.
|
|
Top
|
|
|
|
#260435 - 22/07/2005 14:23
Re: Bugs in V3 A10
[Re: Roger]
|
peter
carpal tunnel
Registered: 13/07/2000
Posts: 3349
Loc: Cambridge, England
|
Quote: The problem is that the "marked" flag should be reserved for the user, or should differentiate between the user setting it, or the player setting it.
On the car-player, certainly. This was really another product's feature, which kind of leaked over.
Peter
|
|
Top
|
|
|
|
#260436 - 22/07/2005 15:41
Re: Bugs in V3 A10
[Re: Roger]
|
peter
carpal tunnel
Registered: 13/07/2000
Posts: 3349
Loc: Cambridge, England
|
Quote: you keep hitting it until it says "no match found". At this point, if you hit track skip, it doesn't
How sucky. FITNR.
Peter
|
|
Top
|
|
|
|
#260437 - 22/07/2005 15:42
Re: Bugs in V3 A10
[Re: peter]
|
pgrzelak
carpal tunnel
Registered: 15/08/2000
Posts: 4859
Loc: New Jersey, USA
|
WITNR?
Paul Grzelak 200GB with 48MB RAM, Illuminated Buttons and Digital Outputs
|
|
Top
|
|
|
|
#260440 - 22/07/2005 16:11
Re: Bugs in V3 A10
[Re: mlord]
|
peter
carpal tunnel
Registered: 13/07/2000
Posts: 3349
Loc: Cambridge, England
|
Quote: The "Previous Visual" virtual button code vanished a while back.
FITPR -- it was in alpha 10, but you're right that it wasn't in previous alphas. Dave did a (manual) v2 to v3 merge for the alpha10 release, and in the bits of code that haven't changed much in the interim (such as the input handling) the merge went in fairly smoothly.
Peter
|
|
Top
|
|
|
|
|
|