Attn plugin devs: porting to python 3

[JimmXinu]

Here's a rather more serious problem--Calibre 4.99.1 Portable doesn't recognize my Kindle Voyage on Win10.

I removed all plugins to make sure it was a clean Calibre.

Log attached.

[theducks]

Originally Posted by JimmXinu

(Tangentially off topic?)

So while working on this, I'm wondering: How common are pre-3.48 Calibre users now?

I'm thinking I might make the min Calibre version for all my plugins 3.48 and start removing some old cruft and the PyQt4 code...

There are quite a few folk that use XP systems. I am willing to guess $$ is a good reason behind sticking with what they have.
It in not just a matter of replacing the CPU box, really old WORKING, peripherals have no support on newer OS.
Then, there are places that only sell FULL SYSTEMS, when they may get away with a good 'referbed' Dell or HP business class from Walmart or the like.

[thiago.eec]

Originally Posted by thiago.eec

Just to be sure: is the metadata download fully functional on 4.99.1? I was trying to port a plugin for it, but I'm not sure it is working, since even the bundled plugins are failing.

Can anyone confirm this? Is metadata download working?

[jhowell]

Originally Posted by thiago.eec

Can anyone confirm this? Is metadata download working?

I tried it with two built in sources, Amazon and Google. It did not work for me either. It just shows:

Found 0 results
Downloading was aborted

Spoiler:

calibre, version 4.99.1
ERROR: No matches found: <p>Failed to find any books that match your search. Try making the search <b>less specific</b>. For example, use only the author's last name and a single distinctive word from the title.<p>To see the full log, click "Show details".

Running identify query with parameters:
{'title': 'The Legend of Eli Monpress', 'authors': ['Rachel Aaron'], 'identifiers': {'isbn': '0316193577', 'amazon': 'B0058ECNXU', 'barnesnoble': 'w/legend-of-eli-monpress-rachel-aaron/1100737227', 'ff': 'a/rachel-aaron/legend-of-eli-monpress', 'goodreads': '11754293', 'google': 'wZj7tgAACAAJ', 'odid': '0013215251@ax/ocls&598347@hcplc.overdrive.com', 'odpid': 'B0058ECNXU@ak/store', 'odrid': '598347@brooklyn.overdrive.com&598347@fairfax.over drive.com&598347@freelibrary.overdrive.com&apftzr9 @3m/bpl'}, 'timeout': 30}
Using plugins: Google (1, 0, 0), Amazon.com (1, 2, 11)
The log from individual plugins is below

****************************** Google (1, 0, 0) ******************************
Found 0 results
Downloading was aborted

************************************************** ******************************

****************************** Amazon.com (1, 2, 11) ******************************
Found 0 results
Downloading was aborted

************************************************** ******************************
The identify phase took 0.40 seconds
The longest time (-1.000000) was taken by: Amazon.com
Merging results from different sources
We have 0 merged results, merging took: 0.00 seconds

[thiago.eec]

Originally Posted by jhowell

I tried it with two built in sources, Amazon and Google. It did not work for me either. It just shows:

Found 0 results
Downloading was aborted

Spoiler:

calibre, version 4.99.1
ERROR: No matches found: <p>Failed to find any books that match your search. Try making the search <b>less specific</b>. For example, use only the author's last name and a single distinctive word from the title.<p>To see the full log, click "Show details".

Running identify query with parameters:
{'title': 'The Legend of Eli Monpress', 'authors': ['Rachel Aaron'], 'identifiers': {'isbn': '0316193577', 'amazon': 'B0058ECNXU', 'barnesnoble': 'w/legend-of-eli-monpress-rachel-aaron/1100737227', 'ff': 'a/rachel-aaron/legend-of-eli-monpress', 'goodreads': '11754293', 'google': 'wZj7tgAACAAJ', 'odid': '0013215251@ax/ocls&598347@hcplc.overdrive.com', 'odpid': 'B0058ECNXU@ak/store', 'odrid': '598347@brooklyn.overdrive.com&598347@fairfax.over drive.com&598347@freelibrary.overdrive.com&apftzr9 @3m/bpl'}, 'timeout': 30}
Using plugins: Google (1, 0, 0), Amazon.com (1, 2, 11)
The log from individual plugins is below

****************************** Google (1, 0, 0) ******************************
Found 0 results
Downloading was aborted

************************************************** ******************************

****************************** Amazon.com (1, 2, 11) ******************************
Found 0 results
Downloading was aborted

************************************************** ******************************
The identify phase took 0.40 seconds
The longest time (-1.000000) was taken by: Amazon.com
Merging results from different sources
We have 0 merged results, merging took: 0.00 seconds

Thanks, @jhowell.
I'll have to wait, then.

[kovidgoyal]

Originally Posted by JimmXinu

(Tangentially off topic?)

So while working on this, I'm wondering: How common are pre-3.48 Calibre users now?

I'm thinking I might make the min Calibre version for all my plugins 3.48 and start removing some old cruft and the PyQt4 code...

Usage stats are available here: https://calibre-ebook.com/dynamic/calibre-usage

[kovidgoyal]

Uploaded 4.99.2 with fixes for metadata download and windows device detetion.

[thiago.eec]

Originally Posted by kovidgoyal

Uploaded 4.99.2 with fixes for metadata download and windows device detetion.

Great! Metadata download is working again.

[JimmXinu]

Originally Posted by kovidgoyal

Uploaded 4.99.2 with fixes for metadata download and windows device detetion.

Can confirm Kindle Voyage working now. Thanks!

[DaltonST]

Users are going to upgrade to the Python 3 version of Calibre just like they always have. They will upgrade it, then start the new Calibre 4.99+ that uses Python 3. Therein lies an important "FYI".

This FYI can be easily demonstrated by Win10 .bat commands in this sequence:

cd "C:\Program Files\Calibre2\" # Python 2
calibre-customize -b "any_plugin_here"
calibre
REM [all is good so far...]
cd "C:\Program Files (x86)\Calibre2\" # Python 3
calibre
REM [Calibre fails with nonsense, such as a standard Calibre library component is missing. False error.]

Why? Because calibre-customize was Python 2, so when the Python 3 was started without first updating the plugin with calibre-customize for Python 3, Calibre fails the very first time it runs.

Execute Py3 Calibre again, and it works fine.

So, the FYI is that users should not be shocked if Calibre 4.99+ immediately fails the very first time they run it after upgrading, and need only run it again.

Just something to be aware of "in real life".

Added: Obviously, when a normal user adds or updates a plugin, they do so through Preferences > Plugins, which then adds/updates the specified plugin just like 'calibre-customize -b "any_plugin_here" ' would do via CLI. The method is different, but the result is the same.




DaltonST

[eschwartz]

And do pray tell why calibre-customize, a program which turns a directory into a .zip file and moves that file into the config directory, would break calibre?

Why would plugin .zip files originally created with python2, cause python3 calibre to not work?

You're really going to need to explain this one a lot better if you expect us to help.

[DiapDealer]

I know I'm confused as hell.

The plugins I've updated don't seem to care whether they're zipped/installed/run by Python 2 or 3 (or any combination thereof). Sort of figured that was the whole point of the endeavor.

[eschwartz]

What if you use neither python2 nor python3, but instead use the Info-Zip "zip" utility...

[JSWolf]

If you have some plugins that have not yet been upgraded, will running Calibre 4.99+ cause a problem with these plugins? Will Calibre work? So far, only one plugin has been updated.

[JimmXinu]

Here's another minor Win10 issue:

If I've copied something to the clipboard, when I quit there's an error:

Spoiler:

calibre 4.99.2  embedded-python: True is64bit: False
Windows-10-10.0.18362-SP0 Windows ('32bit', 'WindowsPE')
32bit process running on 64bit windows
('Windows', '10', '10.0.18362')
Python 3.8.0
Windows: ('10', '10.0.18362', 'SP0', '')
Interface language: None
Successfully initialized third party plugins: ColumnSum (2, 1, 0) && EpubMerge (2, 7, 1) && EpubSplit (2, 8, 2) && FanFicFare (3, 14, 2) && SmartEject (2, 2, 1) && View Manager (1, 5, 4)
Turning on automatic hidpi scaling
devicePixelRatio: 2.0
logicalDpi: 96.0 x 96.0
physicalDpi: 141.35652173913041 x 141.4020618556701
Using calibre Qt style: True
[0.00] Starting up...
[0.02] Showing splash screen...
[0.42] splash screen shown
[0.42] Initializing db...
[0.42] db initialized
[0.42] Constructing main UI...
EpubMerge: DEBUG: 2019-12-19 14:01:48,864: calibre_plugins.epubmerge.epubmerge_plugin(156): macmenuhack file_path:C:\Users\retief\Desktop\nook\CalibrePorts\4.99-2-py3\Calibre Settings\plugins\fanficfare_macmenuhack.txt
FFF: DEBUG: 2019-12-19 14:01:48,869: calibre_plugins.fanficfare_plugin.fff_plugin(209): Plugin FanFicFare macmenuhack file_path:C:\Users\retief\Desktop\nook\CalibrePorts\4.99-2-py3\Calibre Settings\plugins\fanficfare_macmenuhack.txt
[1.64] main UI initialized...
[1.64] Hiding splash screen
[1.92] splash screen hidden
[1.92] Started up in 1.92 seconds with 23 books
Traceback (most recent call last):
  File "__init__.py", line 997, in flush_clipboard
  File "__init__.py", line 359, in __init__
TypeError: a bytes-like object is required, not 'str'

Also, is this meant to clear the clipboard on quit? Because that's really irritating when I'm trying to copy an error message...