[GUI Plugin] Goodreads Sync - Page 13

MykelAlvis · 05-07-2011, 09:22 PM

Even more interesting stuff:
My results above were in KDE, admittedly not the primary desktop environment for Fedora. So I switched over to a GNOME Desktop and I got the following

Code:

[mykel@ramses ~]$ calibre-debug -e test.py 
gnome-open: /opt/calibre/lib/libz.so.1: no version information available (required by /lib64/libgio-2.0.so.0)
[mykel@ramses ~]$ Couldn't load XPCOM.

This produced no browser, as expected.

Code:

[mykel@ramses ~]$ calibre-debug -e test2.py 
[mykel@ramses ~]$

This produced a browser, as expected.

From the code above, we can expect that since it can't find libz in calibre, maybe it should find it in the filesystem?

Code:

mykel@ramses ~]$ sudo yum whatprovides /lib64/libz.so.1
Loaded plugins: langpacks, refresh-packagekit
Adding en_US to language list
zlib-1.2.5-2.fc14.x86_64 : The zlib compression and decompression library
Repo        : fedora
Matched from:
Filename    : /lib64/libz.so.1



zlib-1.2.5-2.fc14.x86_64 : The zlib compression and decompression library
Repo        : installed
Matched from:
Other       : Provides-match: /lib64/libz.so.1
[mykel@ramses ~]$ ll /lib64/libz.so.1
lrwxrwxrwx. 1 root root 13 Jan 24 02:31 /lib64/libz.so.1 -> libz.so.1.2.5
[mykel@ramses ~]$ ll /lib64/libz.so.1.2.5
-rwxr-xr-x. 1 root root 95872 Jun 17  2010 /lib64/libz.so.1.2.5
[mykel@ramses ~]$

As you can see, the package that provides libz is installed and the file is available, at least to the operating system. Why KDE didn't state the issue on the command line is a mystery to me, but so many things about KDE are.

I have some experience with Python, but essentially none with Calibre's custom-built system so I'm afraid I would need help debugging this any further.

DoctorOhh · 05-07-2011, 09:27 PM

Quote:

Originally Posted by MykelAlvis

I have some experience with Python, but essentially none with Calibre's custom-built system so I'm afraid I would need help debugging this any further.

Was this using plugin 1.4.10 which came out 15 minutes ago?

MykelAlvis · 05-07-2011, 09:27 PM

Quote:

Originally Posted by kiwidude

@MykelAlvis - give this version a go and let me know how you get on.

Smashing success! I will now proceed with tainting GoodReads with my selections!

However, based on the errors I was getting, should I report the anomalous behavior as a bug?

DoctorOhh · 05-07-2011, 09:29 PM

Quote:

Originally Posted by MykelAlvis

However, based on the errors I was getting, should I report the anomalous behavior as a bug?

No. It works now, correct?

MykelAlvis · 05-07-2011, 09:53 PM

Quote:

Originally Posted by dwanthny

No. It works now, correct?

Sorry. I was unclear.

The new version of the goodreads plugin solves my issue. @kiwidude is unarguably the bomb.

The remaining question isn't about the plugin. It's about the plugin environment. The webbrowser.open call should have worked but doesn't. More precisely, it does not run on a 64-bit Fedora box using a non-RPM distribution of Calibre. Beyond that, I'm uncertain.

The webbrowser.open behavior is likely a fringe issue, but just because a bug doesn't affect a number of people doesn't mean it shouldn't be reported. Along with that, it might not be a bug. It's possible that this is a known limitation within the plugin environment. Given that I don't work with the calibre code, it's probable that the reporting should fall to @kiwidude since he has far greater context to share than I do. Also, @kiwidude might not think it particularly worthy of reporting and if that's the case, I accept his evaluation. I was just offering to do the scut work of entering the issue if the situation warrants it, which I believe it does.

DoctorOhh · 05-07-2011, 10:05 PM

Quote:

Originally Posted by kiwidude

Change all webbrowser launching to use Calibre's wrapper for the default browser for better Linux support

Quote:

Originally Posted by MykelAlvis

The remaining question isn't about the plugin. It's about the plugin environment. The webbrowser.open call should have worked but doesn't. More precisely, it does not run on a 64-bit Fedora box using a non-RPM distribution of Calibre. Beyond that, I'm uncertain.

I understand better now. My impression was that since Kiwidude found a wrapper within calibre for dealing with this problem that it was more a question of documentation to ensure developers use the correct call for opening a web browser across platforms.

Quote:

Originally Posted by MykelAlvis

Given that I don't work with the calibre code, it's probable that the reporting should fall to @kiwidude since he has far greater context to share than I do.

I concur with you if Kiwidude thinks this needs to be corrected or better documented he would be in the best position to effect the needed change/clarification.

kiwidude · 05-08-2011, 05:00 AM

Several months ago when I did the rewrite of the Search the Internet plugin Kovid did suggest to me to change to using open_url(). However he didn't say why and I found that I had problems using certain url's with it. As I had a todo list a mile long (why is it that it still is that long now?) I didn't have time to figure out the solution so I took the "if it ain't broke don't fix it approach".

I figured out the answers to my issues to apply it to the other plugins at around 3am when you posted last night (using QUrl.fromEncoded() to support URLs which already have been encoded, and prefix with 'file:///' if opening local files) so for me the issue is done with. Kovid is presumably already aware of the issue given he doesn't use webbrowser anywhere in the Calibre codebase.

Thanks for your help in confirming the problem/workaround. There was a Linux user who had posted on the Search the Internet complaining about it stopping working for them after a Linux upgrade so I would guess this may fix things for them too.

nynaevelan · 05-08-2011, 10:58 AM

Hi Kiwidude:

First, the update seems to have fixed the error I was getting when adding new books that have not been linked previously in Gr. Second, a question for you, is there a way when using the plugin to tell which books have not been linked yet?? I am trying to figure out if any books are missing their link prior to running the download shelves feature.

Nyn

kiwidude · 05-08-2011, 11:58 AM

Quote:

Originally Posted by nynaevelan

Second, a question for you, is there a way when using the plugin to tell which books have not been linked yet?? I am trying to figure out if any books are missing their link prior to running the download shelves feature.

Nyn

See the first post in this thread for instructions on how to add a custom column. I have a goodreads column with a tick indicating it has an id.

You can also do a search like:
identifiers:goodreads:false

nynaevelan · 05-08-2011, 01:50 PM

Quote:

Originally Posted by kiwidude

See the first post in this thread for instructions on how to add a custom column. I have a goodreads column with a tick indicating it has an id.

You can also do a search like:
identifiers:goodreads:false

Thank you the column will work perfectly. Is the plugin the only thing in Calibre that will add the goodreads id??

kiwidude · 05-08-2011, 01:53 PM

No, the Goodreads metadata download plugin can add it now too. Either way though that means it is "linked", but it won't tell you whether a book is on a shelf or not if you were attempting to use the field in that way.

nynaevelan · 05-08-2011, 02:48 PM

Quote:

Originally Posted by kiwidude

No, the Goodreads metadata download plugin can add it now too. Either way though that means it is "linked", but it won't tell you whether a book is on a shelf or not if you were attempting to use the field in that way.

No, I do not need the shelf info, I already have a column for that but I was running into problems downloading my shelves because of unlinked books, so I needed a way to fix the links before going to the download shelves portion.

potestus · 05-09-2011, 07:28 AM

calibre, version 0.8.0
ERROR: Unhandled exception: <b>AttributeError</b>:'NoneType' object has no attribute 'strip'

Traceback (most recent call last):
File "calibre_plugins.goodreads_sync.action", line 220, in add_or_remove_to_shelf
File "calibre_plugins.goodreads_sync.action", line 346, in update_calibre_database_ids_for_selection
File "site-packages\calibre\library\database2.py", line 2759, in set_isbn
File "site-packages\calibre\library\database2.py", line 2710, in set_identifier
File "site-packages\calibre\library\database2.py", line 2705, in _clean_identifier
AttributeError: 'NoneType' object has no attribute 'strip'

kiwidude · 05-09-2011, 09:08 AM

Changes in this release:

Do not try to write null ISBN to database after performing a sync

@potestus - thanks for reporting this, hopefully this should fix your issue? Let me know if you see any further problems.

Ivy · 05-13-2011, 05:33 PM

I've never bothered with plugins, but when I saw this I had to have it.
Installed and I'm trying to figure out how to use it.

Thanks for your work!

05-07-2011, 09:22 PM	#181
MykelAlvis Member Posts: 12 Karma: 1748 Join Date: Aug 2009 Location: Birmingham, AL Device: Nook	Even more interesting stuff: My results above were in KDE, admittedly not the primary desktop environment for Fedora. So I switched over to a GNOME Desktop and I got the following Code: [mykel@ramses ~]$ calibre-debug -e test.py gnome-open: /opt/calibre/lib/libz.so.1: no version information available (required by /lib64/libgio-2.0.so.0) [mykel@ramses ~]$ Couldn't load XPCOM. This produced no browser, as expected. Code: [mykel@ramses ~]$ calibre-debug -e test2.py [mykel@ramses ~]$ This produced a browser, as expected. From the code above, we can expect that since it can't find libz in calibre, maybe it should find it in the filesystem? Code: mykel@ramses ~]$ sudo yum whatprovides /lib64/libz.so.1 Loaded plugins: langpacks, refresh-packagekit Adding en_US to language list zlib-1.2.5-2.fc14.x86_64 : The zlib compression and decompression library Repo : fedora Matched from: Filename : /lib64/libz.so.1 zlib-1.2.5-2.fc14.x86_64 : The zlib compression and decompression library Repo : installed Matched from: Other : Provides-match: /lib64/libz.so.1 [mykel@ramses ~]$ ll /lib64/libz.so.1 lrwxrwxrwx. 1 root root 13 Jan 24 02:31 /lib64/libz.so.1 -> libz.so.1.2.5 [mykel@ramses ~]$ ll /lib64/libz.so.1.2.5 -rwxr-xr-x. 1 root root 95872 Jun 17 2010 /lib64/libz.so.1.2.5 [mykel@ramses ~]$ As you can see, the package that provides libz is installed and the file is available, at least to the operating system. Why KDE didn't state the issue on the command line is a mystery to me, but so many things about KDE are. I have some experience with Python, but essentially none with Calibre's custom-built system so I'm afraid I would need help debugging this any further.

05-09-2011, 07:28 AM	#193
potestus Enthusiast Posts: 27 Karma: 10 Join Date: Jul 2010 Device: sony prs-600	I get this occasionally. No Attribute 'strip' calibre, version 0.8.0 ERROR: Unhandled exception: <b>AttributeError</b>:'NoneType' object has no attribute 'strip' Traceback (most recent call last): File "calibre_plugins.goodreads_sync.action", line 220, in add_or_remove_to_shelf File "calibre_plugins.goodreads_sync.action", line 346, in update_calibre_database_ids_for_selection File "site-packages\calibre\library\database2.py", line 2759, in set_isbn File "site-packages\calibre\library\database2.py", line 2710, in set_identifier File "site-packages\calibre\library\database2.py", line 2705, in _clean_identifier AttributeError: 'NoneType' object has no attribute 'strip'

05-09-2011, 09:08 AM	#194
kiwidude Calibre Plugins Developer Posts: 4,636 Karma: 2162064 Join Date: Oct 2010 Location: Australia Device: Kindle Oasis	v1.4.11 Released Changes in this release: Do not try to write null ISBN to database after performing a sync @potestus - thanks for reporting this, hopefully this should fix your issue? Let me know if you see any further problems.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[GUI Plugin] Clipboard Search	kiwidude	Plugins	29	04-02-2024 10:05 PM
[GUI Plugin] Search the Internet	kiwidude	Plugins	433	04-01-2024 05:48 PM
[GUI Plugin] Open With	kiwidude	Plugins	403	04-01-2024 08:39 AM
[GUI Plugin] Quick Preferences	kiwidude	Plugins	62	03-16-2024 11:47 PM
[GUI Plugin] Book Sync Deprecated	kiwidude	Plugins	111	06-07-2011 07:47 PM

05-08-2011, 05:00 AM	#187
kiwidude Calibre Plugins Developer Posts: 4,636 Karma: 2162064 Join Date: Oct 2010 Location: Australia Device: Kindle Oasis	Several months ago when I did the rewrite of the Search the Internet plugin Kovid did suggest to me to change to using open_url(). However he didn't say why and I found that I had problems using certain url's with it. As I had a todo list a mile long (why is it that it still is that long now?) I didn't have time to figure out the solution so I took the "if it ain't broke don't fix it approach". I figured out the answers to my issues to apply it to the other plugins at around 3am when you posted last night (using QUrl.fromEncoded() to support URLs which already have been encoded, and prefix with 'file:///' if opening local files) so for me the issue is done with. Kovid is presumably already aware of the issue given he doesn't use webbrowser anywhere in the Calibre codebase. Thanks for your help in confirming the problem/workaround. There was a Linux user who had posted on the Search the Internet complaining about it stopping working for them after a Linux upgrade so I would guess this may fix things for them too.

05-08-2011, 10:58 AM	#188
nynaevelan eBook Junkie Posts: 1,526 Karma: 1464018 Join Date: May 2010 Location: USA Device: Kindle Fire 2020, Kindle PW2	Hi Kiwidude: First, the update seems to have fixed the error I was getting when adding new books that have not been linked previously in Gr. Second, a question for you, is there a way when using the plugin to tell which books have not been linked yet?? I am trying to figure out if any books are missing their link prior to running the download shelves feature. Nyn

05-08-2011, 01:53 PM	#191
kiwidude Calibre Plugins Developer Posts: 4,636 Karma: 2162064 Join Date: Oct 2010 Location: Australia Device: Kindle Oasis	No, the Goodreads metadata download plugin can add it now too. Either way though that means it is "linked", but it won't tell you whether a book is on a shelf or not if you were attempting to use the field in that way.

05-13-2011, 05:33 PM	#195
Ivy Sometimes I purr. Posts: 252 Karma: 5979384 Join Date: Dec 2010 Location: Texas, USA Device: Kindle	I've never bothered with plugins, but when I saw this I had to have it. Installed and I'm trying to figure out how to use it. Thanks for your work!