How to use the search bar ";xxx" commands to launch an app?

[knc1]

OK, I have all three of those models, will see if I can find why the differences.

Please don't hold your breath on an answer, there are a lot of things that I need to do before that testing.

[nasser]

Thanks, knc1!
No hurry.
Do it at your convenience.

[nasser]

{PW3, G090G1, 5.7.4 factory} {KT3, G000KA, 5.8.0 factory}

Tried copying /mnt/us/extensions/kterm/bin/kterm.sh to /usr/bin and then modifying the line accordingly in /usr/share/webkit-1.0/pillow/debug_cmds.json.

Still doesn't work.

Somehow feel the PW1 was the ideal model for hacking. Compared to earlier and later models:

• backlight
• higher resolution
• active content worked
• Amazon loopholes were still open (like the one being discussed in this thread!)
• Many of the extensions/apps work properly on this resolution only

Oh Well.. The price of progress. A later model like the PW3 has:

• Faster Processor (1GHz instead of 800 MHz, I believe)
• 512 MB RAM instead of 256
• 4GB storage instead of 2
• and of course, 300 ppi instead of 212 or 167

Can't argue with that, either!

[knc1]

The 5.7.4 version of the firmware was released **before** we started re-using search bar commands. So it isn't likely to be an Amazon reaction to what we where doing.

The two that are reported not working are i.MX6 devices.
The one that is reported as working is an i.MX50 series device.

It might be some change made, either accidentally or on purpose with the change to the newer SoC device.

Try this:

ssh into the Kindle
dmesg -c
try the new search bar command
dmesg >/mnt/us/dmesg.txt

Then see if the kernel complained about anything.

Note: these i.MX6 models are most likely running a newer version of Busybox ash.
It could be that the newer version is stumbling over something in the script.

[nasser]

Originally Posted by knc1

Try this:

ssh into the Kindle
dmesg -c
try the new search bar command
dmesg >/mnt/us/dmesg.txt

Then see if the kernel complained about anything.

Note: these i.MX6 models are most likely running a newer version of Busybox ash.
It could be that the newer version is stumbling over something in the script.

dmesg -c gave a big report
then, i tried the new search bar command (;kterm)
now dmesg > mnt/us/dmesg.txt only gave a blank file

[knc1]

Originally Posted by nasser

dmesg -c gave a big report
then, i tried the new search bar command (;kterm)
now dmesg > mnt/us/dmesg.txt only gave a blank file

The "-c" dumped and cleared the message ring buffer.

So if there had been any messages generated by the command, you would have seen it the second time.

Maybe testing command length?
try setting it up for ;kt
we know two character commands work (as in: ;dm).

- - - -

OH, the dmesg command dumps the KERNEL's ring buffer -
The ;dm dumps an edited version of the (d-bus) message buffer.
There are a lot of "message" sources in the device.

[nasser]

Originally Posted by knc1

Maybe testing command length?
try setting it up for ;kt
we know two character commands work (as in: ;dm).

Modified /usr/share/webkit-1.0/debug_cmds.json -
- changed ";kterm" to ";kt"

Same result. Doesn't work. dmesg doesn't show any messages.

Running /usr/bin/dm.sh dumped 4 files into /mnt/us/documents/, 3 of them huge..

[knc1]

Originally Posted by nasser

Modified /usr/share/webkit-1.0/debug_cmds.json -
- changed ";kterm" to ";kt"

Same result. Doesn't work. dmesg doesn't show any messages.

Running /usr/bin/dm.sh dumped 4 files into /mnt/us/documents/, 3 of them huge..

That seems reasonable.
The "factory" firmware probably has the maximum amount of messages possible (to aid in debugging).

[nasser]

Just thought I'd reboot this thread, since a similar thread got started recently! : https://www.mobileread.com/forums/sh...d.php?t=289562

A quick recap/refresher:
I was unable to make new(user-defined) "search bar" commands work, even after a FW restart, in my jailbroken PW3(G090G1, FW 5.7.4, factory image) and KT3(G000KA, FW 5.8.0, factory image), though I could do it on my PW1(B02416, FW 5.3.6, customer image).

The question remained unresolved... though knc1 promised to check on his kindles... https://www.mobileread.com/forums/sh...4&postcount=16

I feel it may be related to the fact that I had retained the factory images and not updated to a customer image.

[knc1]

Originally Posted by nasser

Just thought I'd reboot this thread, since a similar thread got started recently! : https://www.mobileread.com/forums/sh...d.php?t=289562

A quick recap/refresher:
I was unable to make new(user-defined) "search bar" commands work, even after a FW restart, in my jailbroken PW3(G090G1, FW 5.7.4, factory image) and KT3(G000KA, FW 5.8.0, factory image), though I could do it on my PW1(B02416, FW 5.3.6, customer image).

The question remained unresolved... though knc1 promised to check on his kindles... https://www.mobileread.com/forums/sh...4&postcount=16

I feel it may be related to the fact that I had retained the factory images and not updated to a customer image.

Sorry, I thought I had answered that one.

The new "dispatch.sh" script has some serious problems, it can not be extended the way it was intended.
It does not pass arguments.
It does not return error codes.

Your stuck with the one reserved name in the top of visible USB storage:
RUNME.sh
with that capitalization.
;log runme


The backup copy of the installed dispatcher script is:
mkk/dispatch.sh

Changing / fixing that one does nothing to the behavior, it is a copy which you can't see over USB that is actually running.
And we don't have a handy 'update-scripts.sh' command include in the hotfix package.
(That is as far as I got with the question, and I was sure I had posted my non-results.)

I would now suggest working with coplate -
you might be able to make a 'fixed' dispatcher script for inclusion in one of his special 'hotfix' packages.

[coplate]

Hi all, I'm not sure what the dispatch script is, but I just tested adding noew commands on my PW3.

I added this to /usr/share/webkit-1.0/pillow/debug_cmds.json

[root@kindle local]# cat /usr/share/webkit-1.0/pillow/debug_cmds.json
{
    ";coplate" : "/var/local/flash.sh",
    ";dm" : "/usr/bin/dm.sh",
    ";dmcc" : "/usr/bin/dmcc.sh",

[root@kindle local]# cat /var/local/flash.sh
#!/bin/sh
sleep 5
eips 0 0 -f

A 5 second timeout is outside it's wait period, but it did draw a '0' on the screen after 5 seconds.

170821:233951 Xorg: I GestureEngine:perfScenario::Sending button 1 up 1503358791.023752
170821:233951 cvm[4693]: I GoButtonTask:Information::Debug command entered
170821:233951 SCRIPTEXECUTOR[7669]: I script-Executor:main:Ready for execution:Command=/var/local/flash.sh
170821:233956 cvm[4693]: W lipc:perf:prop=executeDebugCommand, source=com.lab126.system, ms=5027:lipc property access took too long

It originally logged that it was running, but didn't actually, I gave the script full permisions., (777) and it ran. I didn't even need a reboot to change to flash, though the ';coplate' was in there on the last reboot, with a copy of the dump entry.

Nasser, if you are still trying, check the execute permissions on the script?

[nasser]

Thanks, Coplate!
Will try it out.

[knc1]

Originally Posted by coplate

Hi all, I'm not sure what the dispatch script is, but I just tested adding noew commands on my PW3.

I added this to /usr/share/webkit-1.0/pillow/debug_cmds.json

[root@kindle local]# cat /usr/share/webkit-1.0/pillow/debug_cmds.json
{
    ";coplate" : "/var/local/flash.sh",
    ";dm" : "/usr/bin/dm.sh",
    ";dmcc" : "/usr/bin/dmcc.sh",

[root@kindle local]# cat /var/local/flash.sh
#!/bin/sh
sleep 5
eips 0 0 -f

A 5 second timeout is outside it's wait period, but it did draw a '0' on the screen after 5 seconds.

170821:233951 Xorg: I GestureEngine:perfScenario::Sending button 1 up 1503358791.023752
170821:233951 cvm[4693]: I GoButtonTask:Information::Debug command entered
170821:233951 SCRIPTEXECUTOR[7669]: I script-Executor:main:Ready for execution:Command=/var/local/flash.sh
170821:233956 cvm[4693]: W lipc:perf:prop=executeDebugCommand, source=com.lab126.system, ms=5027:lipc property access took too long

It originally logged that it was running, but didn't actually, I gave the script full permisions., (777) and it ran. I didn't even need a reboot to change to flash, though the ';coplate' was in there on the last reboot, with a copy of the dump entry.

Nasser, if you are still trying, check the execute permissions on the script?

It is the script called by the
;log
searchbar command after the 'hotfix' is installed.

There is a backup copy of it in:
(Kindle/mkk/dispatch.sh.
(I know, not a very imaginative name.)

[nasser]

Coplate,

I duplicated your entry to the debug_cmds.json file, added the flash.sh file to /var/local, checked that it was executable(chmod 777), restarted my PW3 and entered ";coplate" into the search bar.

It didn't work. Only got the standard "not found" message.

As I've mentioned, my PW3 is still on the factory image. Don't know if this makes a difference.

On my PW1 I use a :kt entry to point to /mnt/us/extensions/kterm/bin/kterm.sh", which works fine. The same entry doesn't work on both my PW3 and KT3.

Originally Posted by coplate

Hi all, I'm not sure what the dispatch script is, but I just tested adding noew commands on my PW3.

I added this to /usr/share/webkit-1.0/pillow/debug_cmds.json

[root@kindle local]# cat /usr/share/webkit-1.0/pillow/debug_cmds.json
{
    ";coplate" : "/var/local/flash.sh",
    ";dm" : "/usr/bin/dm.sh",
    ";dmcc" : "/usr/bin/dmcc.sh",

[root@kindle local]# cat /var/local/flash.sh
#!/bin/sh
sleep 5
eips 0 0 -f

A 5 second timeout is outside it's wait period, but it did draw a '0' on the screen after 5 seconds.

170821:233951 Xorg: I GestureEngine:perfScenario::Sending button 1 up 1503358791.023752
170821:233951 cvm[4693]: I GoButtonTask:Information::Debug command entered
170821:233951 SCRIPTEXECUTOR[7669]: I script-Executor:main:Ready for execution:Command=/var/local/flash.sh
170821:233956 cvm[4693]: W lipc:perf:prop=executeDebugCommand, source=com.lab126.system, ms=5027:lipc property access took too long

It originally logged that it was running, but didn't actually, I gave the script full permisions., (777) and it ran. I didn't even need a reboot to change to flash, though the ';coplate' was in there on the last reboot, with a copy of the dump entry.

Nasser, if you are still trying, check the execute permissions on the script?

[coplate]

I'll downgrade to 5.7.4 and try it there later this week.