[Alpha] Kobo-UNCaGED - Calibre wireless connection

[Carmelocotonto]

Me too. The same problem. I will try copy 'log'

I have got Calibre 4.23
All the rest, the same

Originally Posted by MGlitch

This isn't working for me.

I removed the previous uncaged directory entirely, replacing it with the one in 0.5.1 I'm running NM 0.4.1

Initial run did install nickledbus

Upon subsequent runs the message "running precautionary library rescan" is displayed briefly, the browser then closes and the message Kobo UNCaGED Finished!" is displayed with a 'continue' button. Pressing the button dismisses the message and nothing further happens.

I have wifi on and have confirmed the Kobo can connect, likewise for my Mac running Calibre, they are on the same network. I told calibre to start a wireless device connection. I will note that intentionally having Calibre not running a wireless connection results in the same behavior.

I enabled debug, but I can't find any output of this after a run of KU.

just a recap of the versions of things

Calibre 5.72
NM 0.4.1
KU 0.5.1
Kobo firmware 4.25.15875

[sherman]

Ok, let me try a complete clean install on my Kobo to see if I can replicate.

EDIT: Unable to replicate on my Glo running FW 4.25.15875

I completely uninstalled KU, along with NickelDBus, downloaded the v0.5.1 release archive from GitHub and installed it. Let KU install NickelDBus and reboot automatically. It launched and connected to calibre fine.

[Carmelocotonto]

Originally Posted by sherman

Ok, let me try a complete clean install on my Kobo to see if I can replicate.

EDIT: Unable to replicate on my Glo running FW 4.25.15875

I completely uninstalled KU, along with NickelDBus, downloaded the v0.5.1 release archive from GitHub and installed it. Let KU install NickelDBus and reboot automatically. It launched and connected to calibre fine.

Well, here is a part of the log file created with "logread -F" over telnet

It is a part of it. If you want all the log, I have it.

partOFlog.txt


I think the important part is:

Dec 16 21:04:22 UNCaGED: Starting Kobo UNCaGED
Dec 16 21:04:22 KoboUNCaGED[1463]: 2020/12/16 21:04:22 Started Kobo-UNCaGED
Dec 16 21:04:22 KoboUNCaGED[1463]: 2020/12/16 21:04:22 Reading options
Dec 16 21:04:22 KoboUNCaGED[1463]: 2020/12/16 21:04:22 Creating KU object
Dec 16 21:04:22 KoboUNCaGED[1463]: 2020/12/16 21:04:22 Getting Kobo Info
Dec 16 21:04:22 nickel: (NickelDBus) web browser 'modal:http://127.0.0.1:8181/ ' (modal=1, url='http://127.0.0.1:8181/', css='') (NickelMenu/src/action_cc.cc:410)

Dec 16 21:04:26 nickel: ( 316.539 @ 0x30cadf0 / performance.warning) static void QObjectUtil::repolishChildren(QWidget*, Qt::FindChildOptions, int) "N3Dialog" repolish took 58

Dec 16 21:04:26 KoboUNCaGED[1463]: 2020/12/16 21:04:26 Getting Device Info
Dec 16 21:04:26 KoboUNCaGED[1463]: 2020/12/16 21:04:26 Reading Metadata
Dec 16 21:04:26 KoboUNCaGED[1463]: 2020/12/16 21:04:26 Reading metadata.calibre
Dec 16 21:04:27 KoboUNCaGED[1463]: 2020/12/16 21:04:27 New: failed to read metadata file: readMDfile: error reading metadata.calibre JSON: ReadJSON Decode: json: cannot unmarshal string into Go struct field CalibreBookMeta.application_id of type int

Dec 16 21:04:27 KoboUNCaGED[1463]: 2020/12/16 21:04:27 New: failed to read metadata file: readMDfile: error reading metadata.calibre JSON: ReadJSON Decode: json: cannot unmarshal string into Go struct field CalibreBookMeta.application_id of type int

[NiLuJe]

Nope, that's enough, thanks .

Unless you want to dig into your metadata.calibre file yourself and tell us which of the application_id entries in there are a string, and what they're set to .

@shermp: I, err, obviously have never seen this as anything other than an int myself ^^.

[sherman]

Originally Posted by NiLuJe

Nope, that's enough, thanks .

Unless you want to dig into your metadata.calibre file yourself and tell us which of the application_id entries in there are a string, and what they're set to .

@shermp: I, err, obviously have never seen this as anything other than an int myself ^^.

Neither have I.

It's easy enough to fix, but I'm getting a bit confused as why we're striking this now...

I'm wondering if some really old metadata files are rearing their head.

[sherman]

@Carmelocotonto can you give the attached test version a go please?

It should hopefully solve the 'application_id' issue.

[NiLuJe]

@sherman: At a *very* quick glance (read: ripgrep and no context ) to the code, it may default to '__calibre_dummy__' in some instances.

[sherman]

Originally Posted by NiLuJe

@sherman: At a *very* quick glance (read: ripgrep and no context ) to the code, it may default to '__calibre_dummy__' in some instances.

Siigh... Sometimes I hate dynamically typed languages!

Anyway, it's not a field KU actively uses, so made the struct field type 'interface{}' in UNCaGED as a quick fix.

[MGlitch]

Originally Posted by sherman

@Carmelocotonto can you give the attached test version a go please?

It should hopefully solve the 'application_id' issue.

This did get some progress for me. I now can connect, it displayed connected and gives an option to disconnect.

After a bit of a wait (haven't tried sending anything to my forma from calibre during this brief time of maybe 20 seconds) it then displays this message

Start: packet reading failed : readDecodeCalibrePayload : connectionclosed : readTCP : connection timed out : readTCP *my kobo's internal IP*:44764->*my macs internal IP:9090: i/o timeout

Then it runs the precautionary library rescan and gives the "Finished" message.

[sherman]

Originally Posted by MGlitch

This did get some progress for me. I now can connect, it displayed connected and gives an option to disconnect.

After a bit of a wait (haven't tried sending anything to my forma from calibre during this brief time of maybe 20 seconds) it then displays this message

Start: packet reading failed : readDecodeCalibrePayload : connectionclosed : readTCP : connection timed out : readTCP *my kobo's internal IP*:44764->*my macs internal IP:9090: i/o timeout

Then it runs the precautionary library rescan and gives the "Finished" message.

If you try again, can you try and see if Calibre is doing something during that time period? Check the bottom status bar.

That error usually occurs because for some reason, Calibre hasn't been sending the 'noop' messages that it usually sends every few seconds.

Also, if you can provide the log with debugging enabled, it might help point to where things have stopped. Be aware that the debug log will probably contain the list of books on device, so you may want to scrub those from the log. The debug log logs (in truncated form) very message that is sent and received.

[MGlitch]

Originally Posted by sherman

If you try again, can you try and see if Calibre is doing something during that time period? Check the bottom status bar.

I've had some success, but also failure. I tried again, noting the jobs and such in the bottom status bar. No jobs were running, and Calibre displayed connected to Kobo Forma. I got the time out again.

Tried again, but this time fairly quickly sent a book to the device from Calibre. This seems to have solved the issue. I think I was under the mistaken impression that the new GUI meant i could do everything from my Kobo, which while it would be nice is very much not really needed. Books sent do load up properly. However if I leave it doing -nothing- the error from my last post appears though books sent still are on the Kobo.

Also, if you can provide the log with debugging enabled, it might help point to where things have stopped. Be aware that the debug log will probably contain the list of books on device, so you may want to scrub those from the log. The debug log logs (in truncated form) very message that is sent and received.

Alas I do not seem to be able to do this. I do not have shell access, and despite having devmode on and debug I'm only able to generate a sys log from the dev settings which appears in the .kobo directory. This log contains absolute gibberish compared to the log Carmelocotonto presented. No trace of the English langue can be found in the file i have.

[sherman]

Originally Posted by MGlitch

Tried again, but this time fairly quickly sent a book to the device from Calibre. This seems to have solved the issue. I think I was under the mistaken impression that the new GUI meant i could do everything from my Kobo, which while it would be nice is very much not really needed. Books sent do load up properly. However if I leave it doing -nothing- the error from my last post appears though books sent still are on the Kobo.

It should not be timing out sitting on the connected screen. I should probably check with the latest version of Calibre to see if anything has changed in that regard.

Originally Posted by MGlitch

Alas I do not seem to be able to do this. I do not have shell access, and despite having devmode on and debug I'm only able to generate a sys log from the dev settings which appears in the .kobo directory. This log contains absolute gibberish compared to the log Carmelocotonto presented. No trace of the English langue can be found in the file i have.

Yeah, the syslog file is the one, but there is a lot of extra noise in it... You can search for 'UNCaGED' in a text editor to find the relevant lines.

I'm seriously considering logging to a standalone file instead of the syslog, to make things easier to find.

[sherman]

Originally Posted by sherman

It should not be timing out sitting on the connected screen. I should probably check with the latest version of Calibre to see if anything has changed in that regard.

Seems fine with a fresh installed Calibre 5.7.2. I'm going to need some sort of logs to try and determine what your issue might be.

[MGlitch]

Originally Posted by sherman

It should not be timing out sitting on the connected screen. I should probably check with the latest version of Calibre to see if anything has changed in that regard.




Yeah, the syslog file is the one, but there is a lot of extra noise in it... You can search for 'UNCaGED' in a text editor to find the relevant lines.

I'm seriously considering logging to a standalone file instead of the syslog, to make things easier to find.

For the syslog when I say no part of it appears in English, I meant it's literally just a garble of symbol characters, not technical jargon. I did try searching for uncaged just to double check and nothing was found.

The below is a random short snippet of a line from syslog, this kind of text with different characters is the entirety of the log as read by textedit. I did try looking at the log on a PC and got different characters in notepad and notepad+ but the same general result of it all being absolute gibberish.

"ø2∏∂)wéïçiœ\X"

[sherman]

Originally Posted by MGlitch

For the syslog when I say no part of it appears in English, I meant it's literally just a garble of symbol characters, not technical jargon. I did try searching for uncaged just to double check and nothing was found.

The below is a random short snippet of a line from syslog, this kind of text with different characters is the entirety of the log as read by textedit. I did try looking at the log on a PC and got different characters in notepad and notepad+ but the same general result of it all being absolute gibberish.

"ø2∏∂)wéïçiœ\X"

Ooof, that doesn't sound good!

Maybe others may have an idea of what's going on here? I certainly don't. I've never seen a corrupted syslog like that.