Okay, Hijack v152 is released, at http://empeg-hijack.sourceforge.net/

This version has:

-- Slightly longer timeout for two-drive booting. If not long enough, let me know and I'll increase it some more.

-- The .U and .N flags in [ir_translate] are no longer supported left of the '=' equal sign (were broken anyway).

-- Symbolic button names are now used for [ir_translate]. Hijack does string substitution when echoing back translations to the serial port, and also accepts button "names" as well has hex codes in the translations in config.ini. This means you can now do:

Bottom.LT=Player.N,Bottom.LU,null ;; Tuner active, switch to player

Bottom.LM=Radio.N,Bottom.LU,null ;; Player active, switch to Tuner 

instead of:

06.LT=b9461E.N,06.LU,ffffff ;; Tuner active, switch to player

06.LM=b9461C.N,06.LU,ffffff ;; Player active, switch to Tuner

The button names are, for the most part, exactly as spelled on the remote(s), except that the numbered buttons are *spelled* ("one" instead of "1"), and a few common synonyms have been included.

For a full and current list, look for "button_names" in the patchfile. Here are the "names" known in v152 (all names are case-insensitive):

Initial, null, One, Two, Three, Source, Four, Five, Six, Tuner, Seven, Eight, Nine, SelectMode, Cancel, Zero, Search, Sound, PrevTrack, Prev, Track-, NextTrack, Next, Track+, Menu, Ok, VolUp, Vol-, VolDown, Vol+, Info, Detail, Visual, Visual+, Play, Pause, Top, Bottom, Left, Right, Knob, KnobLeft, KnobRight, AM, FM, Direct, Star, *, Radio, Auxiliary, Tape, Player, CD, CDMDCH, DNPP

Hijack v153 will be released in a minute or so, with a small fix for the button name lookup's.. so that "CDMDCH" and others work again.

If anyone out there has a bunch of [ir_translate] data already defined in their config.ini file, it will still work with the latest Hijack versions (no worries).

And if you want to convert it to the new "easy to read" format, all you have to do is capture the "ir_translate: " lines from the serial port at boot time, and then paste that data back into config.ini.

Cheers

Hey Mark,
thats a Way Cool implementation.

How much extra valuable kernel memory [if any] does symbolic buttons liek this take up compared to the hex codes implementation?

Quite a bit right now, perhaps around 1KB or so.

But I'm going to rewrite some other features to just leach from the same code/data (already done some of that), and there are some planned features that will also sit on top of it, so not too bad overall.

Building the strings into the kernel is a lot cheaper than providing run-once code to read/parse macro names from config.ini, so I am happier with this form than with other implementations.

But we still must be vigilant.. I don't like the fact that the compressed kernel is now over 500KB in size -- we only have 640KB of flash for kernels.. (some of which I might steal for a MUCH larger Hijack savearea someday).

Cheers

mlord, thanks for your constant attetion to this project and the requests of others. its good to see people like you actively listening as well as creating. one thing i was wondering, if i could. i have a wireless keyboard, is it possible to change your button codes display function to show all the signals that the ir receiver picks up? and i could use my keyboard with the ir translations?

Doesn't the keyboard use IRDA (which is a different device/interface than the consumer IR)?

i wouldnt know anything about it more than the fact that the keyboard works.

:P

it was just a thought i was throwing up.

This was discussed back in November(?). Basically, the keyboard current will be ignored at a lower-level than Hijack, so Hijack cannot currently capture/display codes from it.

The IR driver (empeg_input.c) doesn't "recognize" the IRDA format, so it just tosses the bits that it sees.

You can edit/instrument that driver to show incoming data, and from that you could add IR-keyboard support to the existing IR driver. This is something I have on the back-burner, for the unlikely day that I actually obtain an IR-keyboard that is frequency-compatible with my Empeg.

cheers

Okay, I've just put out Hijack v154.

This version removes the built-in "PopUp" button menu that so many of us love and curse, and instead allows/requires users to define their own version of it.

Details are on the IR translation page at http://empeg-hijack.sourceforge.net/ir_translate.html

Basically, if you want the "old" functionality back, just add this to your config.ini file on the player:

[ir_translate]

PopUp0=Info,Mark,Repeat,SelMode,Shuffle,Source,Tuner,Visuals

But you can now replace any of those buttons with your own choice(s), or add/delete to your heart's content. I would suggest, though, that no more than eight buttons be placed on PopUp0, since only 3-bits are used to keep track of the most recently-pressed button in flash.

Other PopUps can be configured as well, and assigned to the button(s) of your choice, as in:

[ir_translate]

PopUp0=Info,Mark,Repeat,SelMode,Shuffle,Source,Tuner,PopUp1

PopUp1=Artist,Album,Genre,Year,PopUp0

Bottom.LU=PopUp1



[hijack]

PopUp1=Tweak

Hopefully I haven't broken much with all of this.. or at least it all works for me!

Cheers

Nice! Now is there anyway to make a menu load something on the empeg? That would be extreemly cool. Such as:
[hijack]
PopUp1=Tweak ; rename User1 menu to "Empetis" ins

Where User1 would run /usr/emptris&

-Greg

Now THAT is some sweet functionality. Let me try this sucker out...

Yes, there is. The author of Emptris (bmihulka) could port it to use Hijack's display commands and bind to the hijack menu... Not sure on Mark's plans to launch arbitrary user programs though...

I forget, does ir_translate allow you to re-define "chord" keypresses like the early hijack versions? Like I'd like to use front-panel chords for my popup menus...

Also, is there any way you might be able to allow the user to configure the delay between when a press is considered "short" versus "long"? It seems that it's one second or so. I'd like something a bit shorter, maybe 0.5s or something. Just another feature request to pile on...

Also, I didn't see any available symbol for the "source/power" button on the Rio Remote (0x20DF03) It showed up as a hex string instead of a symbol in the Hijack debug output.

That nested popups feature is SLICK! I love it.

So, can Popup definitions have .M and .T qualifiers so they show differently depending on if you're in MP3 or Tuner mode? That'd make my source button easier (I don't like cycling through Aux since i don't use it, so I'd just have the Tuner PopUp menu have "MP3" and the MP3 PopUp have "Tuner.")

No chords that I know of..

Hello Mark,

I did install V156 and the PopUP0 is not working. I added the ir_translate code in the config.ini with the emplode! Running 2.0B7 on a MK2.

Here my config.ini

[Options]
Name=CrazyMelki's Empeg
[User Info]
[email protected]
Phone=+41796284116
Name=Melkon Torosyan
[Network]
Gateway=10.0.1.5
Netmask=255.255.255.0
IPAddress=10.0.1.10
DHCP=0
[sense]
muteaction=0
mute=-1
[display]
visual_names=1
caching=1
[hush]
hushpercent=10
[volumecap]
enabled=1
[kenwood]
disabled=1
[ramp]
disabled=1
[controls]
stalk_side=left
[tuner]
region=europe
[aux]
source=Only the best
title=MP3 Player
artist=PJBOX100
[power]
off_timeout=120000
[synchronise]
seq_num=75
[wendy]
flag_amount=0
[custom]
wendy=
[searches]
amount=0
[startup]
reservecache=12
[menu]
quit=1
sort_playlist=1
[ir_translate]
PopUp0=Info,Mark,Repeat,SelMode,Shuffle,Source,Tuner,Visuals

do i miss somemthing.

Thanks for your great work once again.

bye

Okay, so now go into the Hijack menus, to the "Knob Press Redefinition", and select "PopUp0" from the menu there.

It should work fine after that.

-- I'll add the missing "Source" button definition in v157 (later today?).

-- The 1-second "Long Press" timeout is not configurable, and I'd rather it stay that way -- it has to synchronize with how the player works, and I think the player uses something around 1-second as well. Also, the code size would increase by a few hundred bytes if I made it configurable rather than leaving it as a small constant.

- The Popup menus ignore most/all modifiers, but the translation you use to invoke them can use any modifiers you like, as in:

   Bottom.LT=popup1

   Bottom.LM=popup2

Cheers

The 1-second "Long Press" timeout is not configurable, and I'd rather it stay that way -- it has to synchronize with how the player works

In what way? Would things just not work if the timeout was changed? Hmm... If you don't want to make it configurable, can you maybe point me at the constant I need to change to experiment with it? I can see why you might want everyone to have it the way it is, but if it's possible to hack a smaller delay, I'd like to do that...

The Popup menus ignore most/all modifiers, but the translation you use to invoke them can use any modifiers you like, as in:

Bottom.LT=popup1
Bottom.LM=popup2


Ah, yes, but the problem is I wanted to have the PopUp0 (the one bound to the knob) act differently depending on whether it's in tuner or aux. But from your comments on your ir_translate page, it sounded like making the PopUp menus act like the other translations (with modifiers and such) was counterproductive or prohibitively difficult or something. Not a problem.

Oh, what does renaming the popup menus actually do besides change the symbol you use to refer to it in the ir_trans section? I thought it would change the actual "name" of the menu but now it always says "Select Action:" I figured it would say "Tweak:" for the tweak menu, for instance, but on mine (v154) it doesn't.

If I recall correctly, the PopUp1=Tweak, etc lines only change what gets displayed in the PopUp menu, and not what you use in the ir_trans section.

Also, I would reccomend setting it to 'Tweak...' to indicate it's a menu, and not an action.. Makes it easier to navigate through the menus.

That's mostly it.

Changing the menu label, changes what it is called on any (other) Popup menus that reference it. Changing PopUp0 also affects the name shown in the "Knob Press Redefinition" choices. The [ir_translate] section is processed BEFORE the [hijack] section of config.ini, specifically to PREVENT the new name from confusing any translations you might have (it really does replace the name globally..).

Cheers

Am I right in the understanding that it requires a power-reboot (or through hijack) to reload the config.ini file? I noticed when resync'ing through emplode, that changes to the config.ini didn't take all the time.

Sorry Mark, but it does not! Also the NextSrc is not working. I had until today almost all your great HiJacks installed. And the "Knob Press" was working very well!

Thanks

Hmm... Gotcha. I just thought the "select action" could be replaced by whatever that redefined name is (Tweak for instance).

Also, I've noticed that if I try to use the knob press while a message box (from the player) is up, it doesn't bring up the pop-up menu, it instead cycles through the conventional volume/balance/fader settings. This prevents me from tapping the knob a whole slew of times to tweak the order a bunch of times (to queue up from the same artist or whatever.) In order to do it, I have to wait for each "Found Matching Artist" message box to disappear first. Any way this can be averted? Definitely falls under the "minor inconvenience" category though. Solve world hunger first.

Okay, I had a typo in my original posting on getting PopUp0 to work, in that "Visuals" should be spelled "Visual", as shown on the remote.

So, for the "old knob PopUp behaviour", use these lines in config.ini:

   [ir_translate]

   PopUp0=Info,Mark,Repeat,SelMode,Shuffle,Source,Tuner,Visual

-ml

The Popup likely is not working because you probably copied my typo.. "Visual" is spelled "Visual", not "Visuals" on the remote.. fix that in your config.ini file and try again.

NextSrc works fine. What makes you suggest otherwise (in other words, GIVE ME SOME FRIGGIN DETAILS, otherwise I cannot fix it).

Cheers

Mmm.. I'm not going to touch that one. You could always use any button other than the KNOB for your tweak menu..

Cheers

IR translations can be updated by just using Emplode to edit config.ini and restart the player. But things like the kftpd/khttpd parameters are only updated on reboots.

The current versions of hijack also have an issue whereby if you rename the popup's, then the new names "stick" even when restarting the player, which will cause the ir_translate to fail since the names are now different. Thus reboots are required in that case as well.. fixed in the next release.

-ml

The 1-second "Long Press" timeout is not configurable, and I'd rather it stay that way -- things are getting hard enough to keep working as is. Kinda makes me appreciate just how amazing Hijack really is, that it even works at all (considering what it is doing to a "blackbox" player).

-ml

Okay... Maybe I wasn't clear... I wasn't asking you to go change it for Hijack, just to point me to where I might try to change it myself in the kernel source, like the name of a constant or something.

And to your other point, we've all been telling you how amazing Hijack is since v1... Where the hell have you been? You sound like an artist who just finally stepped back to see the whole painting and realized "hey, this thing friggin' rules!"

Mark,

I found the problem. in your example is written:

[ir_translate]
PopUp0=Info,Mark,Repeat,SelMode,Shuffle,Source,Tuner,Visual

But I did now only:
[ir_translate]
PopUp0=Info,Visual

And this is working! I think there is a little mistake on your documentation.

Because of the NextSrc. I did this only test here on my desk. But I haven't here a Tuner...I think this is the problem. Because you are writing that Player -> Tuner -> Aux are the steps.

bye

The problem with the Popup0 example was that I said "Visuals" instead of "Visual".

The "NextSrc" works for me, on a standalone player with no tuner.

-ml

Sorry Mark,

But I have two (Mk2 and Mk2a) Players with 2.0B7 and V156. But If I choose form the "Knob Press Redefiniton" "NextSrc" there is no action on my quick press. But if I change to all other definitions it works.

Also if I have it on the PopUp0 Menu, I can see this name but after pressing the knob nothing happens. It goes back to the normal player screen.

I think that the Device have to go to the AUX mode, if the Tuner is not present. Or not?

Let me try it tomorrow in my car.

I really estimate your work! And the only think I try to do, is helping you to develop a great work.

bye

Okay, so you mean "NextSrc" on the knob-press menu (it also exists in [ir-translate]..).. Well, that works just fine for me, on both a Mk2 and Mk2a, with and without a Tuner.

Too bad for you.

And the PopUp0 requires that you add a line to your config.ini, and the example I originally posted said "Visuals" on that line, which is an error, it should have said "Visual". Fix that, and it also works fine.

-ml

Yes I know

I changed also the mistake in the config.ini. If there is a mistake like "Visuals" the PopUp0 is not working.

But anyway. I have thousand of of more features because of Hijack. So there is no problem, if only one is not working.....

Thanks

Two questions:

1) If knob press is redefined to activate "Details" and the Empeg is left displaying details should a redefine of knob press to something else still work. I got into this trap inadvertantly and when I tried to redefine knob press as a popup it wouldn't! Not a big deal, but I had to go back to details and back out to a regular display mode before I could use the knob press linked popup.

2) Would it be possible to create a symbolic name for cycling through the auto volume adjust modes that could be included in a popup menu as is available in the regular Hijack menu with "VolAdj+"?

Playlist browsing is GREAT!!!

Lynn

1) Yes, other knob press definitions should still work when Details are displayed.. or at least that's what happened when I tried it just now.

2) I'll add a "VolAdj" buttoncode as part of v157 tonight.

Cheers

Okay, v157 is out. New this time around:

-- new syntax for playlist browsing:

http://your.empeg/drive0/fids/101?.html

-- two drive systems will still have troubles (not fixed yet).

-- .m3u (mp3 playlist) files are generated now, for individual tunes as well as albums. But not for nested playlists. Let me know if it works or not.

-- A "VolAdj" virtual button code has been added for ir_translate. Completely untested. Again, feedback will be useful.

-- The missing button name for "Source" has been added.

Have fun!

-ml

2) Thanks, who could ask for anything more (responsiveness wise at least).

1) Working now for me too, just had to wait for "Details" to time out.

Lynn

Ooops small bug (m3u for single tunes was broken) in v157 -- v158 now replaces it.

-ml

v158

m3u for single tunes is working.
m3u for playlists comes up with a 404 for each tune inside winamp...

The file info for a tune inside winamp looks like http://192.168.168.20/drive0/fids/13920
but when I play it by itself it looks like http://192.168.168.20/drive0/fids/3660 and works ok. Looks like a wrong fid.

Tom

Edit:.. ok.. easy solution.. the numbers generated in the playlist .m3u are in decimal, should be hexidecimal..

Edit: The times still appear to be 1 second too long.

The main problem is just a missing step in the conversion...he'll catch that one right quick. The one second thing could be a rounding problem. Certainly no biggie.

Thanks a lot for your time, Mark...it's coming along very nicely!

Mark is it possible PopUp menus got broken in v158? I flashed from v154 to v158 and my PopUp doesn't, well, pop up. Instead, I get this on the serial when I push the knob:

popup_activate(0fffff9): 00000000

The relvant lines in config.ini:

[hijack]

PopUp1=Tweak



[ir_translate]

PopUp0=Tweak,Info,Shuffle,Mark

Tweak=Artist,Album,Genre,Year,Swap,Cancel

your lines:
PopUp0=Tweak,Info,Shuffle,Mark
Tweak=Artist,Album,Genre,Year,Swap,Cancel

should be:
PopUp0=PopUp1,Info,Shuffle,Mark
PopUp1=Artist,Album,Genre,Year,Swap,Cancel

This must have changed then because it worked fine in v154. The line in the [hijack] section is supposed to substitute PopUp1 with Tweak

What Mark said here is that the ir_translate section is read before the hijack to prevent the renaming from affecting the translations..

Er um... Hmm. This worked in v154 and it doesn't in v158 so something changed... I'm sure Mark will be overjoyed to explain it to us.

In any event, changing it worked.

Already explained elsewhere.. it was a bug that it happened to work by accident in the early versions.. Note that if you actually tried it after a power-off in the early version, it would NOT work.

-ml

OK.

Yeah, couple've goofs in the sprintf() formatting..

v159 will be out shortly with fixes.

Cheers

-ml

Instead of "waiting for details to time out", just hit the knob again.. "Details" cancels "Details", so does the bottom button on the front panel (I think).

-ml

I have V158 of hijack on my mk2 Empeg, while testing the mp3 streaming to winamp [which works well] after I'd played a few songs (and I had closed winamp) -
Then I tried another song (and I hit the stop button in winamp after it had restarted itself) and within a few seconds of a song starting playing.

Then any song I tried to play from them on via the http interface would stop streaming at the same few seconds point where the song I stopped stopped playing.

Once I rebooted the empeg, problem went away again.
[variable not reinitialised properly or at all?].

Not sure if the problem is a bad error code returned from Winamp or what. The Winamp I'm using is 2.78

[Edit]
Try as I might I can't reproduce it now on V158 - and now I've upgraded to V160. I assume its related to some weirdness with Winamp as I had the Moodlogic plugin enabled at the time and it popped itself up when winamp did and when I closed the plugin it must have caused Winamp to spit back a bad error code to the hijack http server causing it to serve later mp3 files to that point only]. I've no idea what the problem is. I'm not worried by it as a reboot fixed it. However it could be a potential bug in the http server that should be fixed. If needed I will try and reproduce on a later version of the kernel to see if its still there].
[/Edit]

I just noticed on my system [running hijack V160 now], that the version reported on boot up from the Empeg software
[player?] is 2.0Beta3
[the screen that appears after waving Tux which has text at the top left hand side that looks like:
empeg-car
Version 2.0 Beta3
empegcar

And has the small empeg logo on the right hand side.


I also notice that Marks kernels are now named in this format:
vXXX.hijack.v200b8.mkY.zImage
with the orange text being b8 for presumably beta 8?
[b8 was b7 a while ago].


Is the Beta3 version number in the Player software or in the kernel as a hangover from Beta 3 or what?
I assume its the player software which if it is the case is not Something that Mark [or any kernel hacker] can do much about.

That text comes from the player software. Mine says "beta 7 (internal)" as it should. I think you should make sure you really do have the beta7 version of the player installed.

I am sure I upgraded my software from B3 to B7 on or during Christmas [about 5 weeks ago now].

I also upgraded my emplode at the same time - and it reports as Beta 7a.

I could swear on a stack of bibles that I upgraded my player as well as my wifes one to beta 7 - but mine shows beta3 for the player and my wifes (correctly) shows beta 7 (internal).

Anyway. I've [re]applied 2.00 Beta 7 [it now says beta7 Internal in the player version# too].
Then I put Mark Lords latest Hijack kernel on - all seems well.

And you know what - I now get that flashing dot in the solid line when playing a tune with Info enabled - which I recall everyone commenting on when beta7 came out, but I never saw on my player - now I know why [D'uh].

If you see "bet3", then that's the player software that you have installed. You should probably upgrade to beta7. The kernel is independent of all of this.

-ml

hello Mark,

Now if found where the bug is . I had enabled in the emplode under ->File->Configure Player->Advanced the checkbox "Disable Kennwood Remote Control". And with this settings the NextSrc is really not working on MK2 and MK2a.

Hopefully you trust in my something more ....

Ahh good one. I will modify the NEXTSRC implementation to use RIO codes instead of Kenwood codes.. it will work slightly differently as a result, though, be should be fine (the Kenwood remote has more functionality than the Rio for this purpose).

-ml