[GUI Plugin] Quality Check

[DNSB]

Originally Posted by RotAnal

DNSB,
I tried to install YOUR modified zip file and everything has been fine. Thanks so much!
Of course I have made some stupid error somewhere.
Thanks to all!

You're welcome. OTOH, the comment here is that I know enough about programming to be dangerous.

[RotAnal]

the only annoying thing now is that every time I try to install the QualityCheck plugin, Calibre eventually gives an error "same file", but the plugin appears among the other on the Calibre row. But at the next restart it disappears. How to fix it without reinstalling Calibre?

[DaltonST]

Read the bottom of my answering post from yesterday.

[davidfor]

Originally Posted by RotAnal

the only annoying thing now is that every time I try to install the QualityCheck plugin, Calibre eventually gives an error "same file", but the plugin appears among the other on the Calibre row. But at the next restart it disappears. How to fix it without reinstalling Calibre?

I frequently reinstall plugins for various reasons and have never seen a "same file" message. Can you post the full details of the error?

Or, are you trying to install the copy of the plugin that is already in the plugins directory? That might give a "same file" error. If so, either do the zipping of the file elsewhere, or don't bother doing the install. If you update the file in place correctly, it will just work when you restart calibre.

[BetterRed]

Originally Posted by davidfor

Or, are you trying to install the copy of the plugin that is already in the plugins directory? That might give a "same file" error.

Yep that's exactly what it does.

I always extract plugins to a desktop folder, make changes, zip it up and install via Prefs->Plugins->Load plugin from file.

BR

Spoiler:

calibre, version 4.11.2
ERROR: Unhandled exception: <b>Error</b>:`C:\_AppData\Calibre\Config_64bit\plugins\Barnes & Noble.zip` and `C:\_AppData\Calibre\Config_64bit\plugins\Barnes & Noble.zip` are the same file

calibre 4.11.2 [64bit]  embedded-python: True is64bit: True
Windows-10-10.0.18362-SP0 Windows ('64bit', 'WindowsPE')
('Windows', '10', '10.0.18362')
Python 2.7.16+
Windows: ('10', '10.0.18362', 'SP0', u'Multiprocessor Free')
Interface language: None
Successfully initialized third party plugins: CalibreSpy (1, 0, 78) && Job Spy (1, 0, 174) && Reading List (1, 6, 7) && Goodreads (1, 4, 0) && Extract ISBN (1, 4, 4) && EpubSplit (2, 9, 0) && GetFileName (0, 0, 13) && Find Duplicates (1, 6, 3) && View Manager (1, 3, 2) && Barnes & Noble (1, 2, 15) && Favourites Menu (1, 0, 5) && Resize Cover (1, 0, 3) && Quality Check (1, 9, 11) && Count Pages (1, 9, 0) && Drop Search Results (1, 0, 10) && Modify ePub (1, 4, 0)
Traceback (most recent call last):
  File "site-packages\calibre\gui2\preferences\plugins.py", line 319, in add_plugin
  File "site-packages\calibre\customize\ui.py", line 481, in add_plugin
  File "shutil.py", line 83, in copyfile
Error: `C:\_AppData\Calibre\Config_64bit\plugins\Barnes & Noble.zip` and `C:\_AppData\Calibre\Config_64bit\plugins\Barnes & Noble.zip` are the same file

[RotAnal]

Hi,
I got the last issue because I erroneously thought I had to put the Quality Check zip file to the plugin dir before loading it through Calibre. Now that I left it in the desktop, everything went fine. Thanks again!
DaltonST: I loaded the file which DNSB gave us, so no issue in rezipping it.

[kunkel321]

Hi KwikiDude. A feature request: Add the ability to swap Title <-> Author right from the main library window. Like with a hotkey, probably.

Here is the post where BetterRed suggested posting here:
https://www.mobileread.com/forums/sh...41#post3989841

Thanks for considering this.

[warrimoo]

Hi,

I could be missing a pre-existing option somewhere but would it be possible to scan the library for ebooks that contain links that refer outside the current book? Ideally as a requested scan within the quality checker, but also as a selectable action on the 'control the adding of books' dialog.

I recently encountered an ebook with a link to a now defunct Yahoo based page/image ( https://ad.yieldmanager.com/st?ad_ty...section=280303 ) that would probably have been used for advertising/usage tracking. Every time I viewed the book or a htm file within the book it opened a new browser window to report the page could not be found. Samples of this type of linking can be found in this epub at ( https://epdf.pub/the-complete-nero-wolfe.html )

Aside from the pain of the random page openings and needing to edit the files to remove the links, it seems to me that this could be a security weakness if people aren't aware of the possibility of links.

Of course, some books will be published with planned links to reference sources so the behaviour after discovery might need to be left to the user. But maybe an option to disable the links that are found could be useful or to tag the book in the metadata for later review.

Thoughts?

Thanks.

[JSWolf]

Originally Posted by warrimoo

Hi,

I could be missing a pre-existing option somewhere but would it be possible to scan the library for ebooks that contain links that refer outside the current book? Ideally as a requested scan within the quality checker, but also as a selectable action on the 'control the adding of books' dialog.

I recently encountered an ebook with a link to a now defunct Yahoo based page/image ( https://ad.yieldmanager.com/st?ad_ty...section=280303 ) that would probably have been used for advertising/usage tracking. Every time I viewed the book or a htm file within the book it opened a new browser window to report the page could not be found. Samples of this type of linking can be found in this epub at ( https://epdf.pub/the-complete-nero-wolfe.html )

Aside from the pain of the random page openings and needing to edit the files to remove the links, it seems to me that this could be a security weakness if people aren't aware of the possibility of links.

Of course, some books will be published with planned links to reference sources so the behaviour after discovery might need to be left to the user. But maybe an option to disable the links that are found could be useful or to tag the book in the metadata for later review.

Thoughts?

Thanks.

There is a way. Use the Quality Check plugin to search for <a href="http and you'll find what you want.

[davidfor]

Originally Posted by JSWolf

There is a way. Use the Quality Check plugin to search for <a href="http and you'll find what you want.

That will only work if the first attribute of the tag is the href. Just searching for the href="http is probably enough. Though if the search is using regex, that would be better.

But, that probably isn't it. If the URLs are being opened automatically, it isn't via links. It's probably some Javascript. You probably want to search for that and then use the Modify ePub plugin to remove it. Of course, without actually seeing a sample, we can't be sure what is actually happening.

[warrimoo]

Originally Posted by davidfor

That will only work if the first attribute of the tag is the href. Just searching for the href="http is probably enough. Though if the search is using regex, that would be better.

But, that probably isn't it. If the URLs are being opened automatically, it isn't via links. It's probably some Javascript. You probably want to search for that and then use the Modify ePub plugin to remove it. Of course, without actually seeing a sample, we can't be sure what is actually happening.

Thanks, and Yes, in this particular case the script looks like this:


<noscript><!-- BEGIN STANDARD TAG - 728 x 90 - Lycos - Angelfire Fallthrough - DO NOT MODIFY -->
<iframe class="calibre5" frameborder="0" height="90" marginheight="0" marginwidth="0" scrolling="no" src="http://ad.yieldmanager.com/st?ad_type=iframe&amp;ad_size=728x90&amp;section=2 80303" width="728"></iframe> <!-- END TAG --></noscript>

The preview in the Calibre editor fires this script from the <body> as well. So it would be really good to have a way to generically detect and deactivate this sort of code without directly opening the editor. I guess a regexp check for <noscript> is probably a first pass for the more literate, but adding functionality to the plugins might be helpful for everyone. The payload in this particular case is mild.

Thanks again.

[davidfor]

Originally Posted by warrimoo

Thanks, and Yes, in this particular case the script looks like this:


<noscript><!-- BEGIN STANDARD TAG - 728 x 90 - Lycos - Angelfire Fallthrough - DO NOT MODIFY -->
<iframe class="calibre5" frameborder="0" height="90" marginheight="0" marginwidth="0" scrolling="no" src="http://ad.yieldmanager.com/st?ad_type=iframe&amp;ad_size=728x90&amp;section=2 80303" width="728"></iframe> <!-- END TAG --></noscript>

The preview in the Calibre editor fires this script from the <body> as well. So it would be really good to have a way to generically detect and deactivate this sort of code without directly opening the editor. I guess a regexp check for <noscript> is probably a first pass for the more literate, but adding functionality to the plugins might be helpful for everyone. The payload in this particular case is mild.

Yes, it should be easy to remove with a regex in the editor. And it shouldn't be to hard to do in something like the Modify ePub plugin. The problem is going to be whether anyone is interested in adding the function. It doesn't feel like a very common function. You should ask in the plugin's thread and see.

[warrimoo]

Originally Posted by davidfor

Yes, it should be easy to remove with a regex in the editor. And it shouldn't be to hard to do in something like the Modify ePub plugin. The problem is going to be whether anyone is interested in adding the function. It doesn't feel like a very common function. You should ask in the plugin's thread and see.

Yes, a regex like <script.*?</script> with dot all on seemed to find most entries in the editor and similar for noscript tags. For some reason a few of the script> tag entries weren't matched and were still present when I checked later.

Maybe I'm over-worrying about this as a vector for attack and should stop thinking of needing virus prevention at the point of adding a book. I can just run a 'scan' occasionally.

[JSWolf]

Originally Posted by warrimoo

Yes, a regex like <script.*?</script> with dot all on seemed to find most entries in the editor and similar for noscript tags. For some reason a few of the script> tag entries weren't matched and were still present when I checked later.

Maybe I'm over-worrying about this as a vector for attack and should stop thinking of needing virus prevention at the point of adding a book. I can just run a 'scan' occasionally.

You could run the Modify ePub plugin. It allows you to remove Javascript.

[davidfor]

Originally Posted by warrimoo

Yes, a regex like <script.*?</script> with dot all on seemed to find most entries in the editor and similar for noscript tags. For some reason a few of the script> tag entries weren't matched and were still present when I checked later.

Maybe I'm over-worrying about this as a vector for attack and should stop thinking of needing virus prevention at the point of adding a book. I can just run a 'scan' occasionally.

For your example, you probably need:

<noscript.*?<noscript>

But, a more general would be:

<((?:java|no)?script).*?<\/\1>

That should get them all. But, it won't handle any self-closed tags. Or any mismatched tags.