epubcheck plugin for Sigil

[Doitsu]

exaltedwombat found a bug in version 0.3 of the epubcheck plugin that causes Sigil to write an ePub with two mimetype entries in the ePub file list. (For more information see this thread.)
This bug has been fixed in the latest epubcheck plugin version (0.3.5), which has been attached to the first post in this thread.

If you saved an ePub file immediately after running the old epubcheck plugin, you might have created an ePub file with two mimetype file entries.
You can tell whether ePubs that you've edited are affected by this bug, by opening them with an archive manager, e.g. 7Zip. If you see two mimetype files, your ePub file has been saved with an invalid zip file list and you'll need to re-open and re-save the ePub file(s) with Sigil or Calibre Editor.

I created a batch file that'll help you track down these files.

Prerequisites: Download and install 7Zip (both the GUI version and the standalone console version).

1. Extract 7za.exe from 7z1514-extra.7z to a temporary folder.

2. Create a text file with the follwing contents in the same folder:

FOR /R %%f IN ("*.epub") DO  (
    echo %%f >> mimetype.log
	7za l "%%f" | findstr -N "mimetype" | find /c ":" >> mimetype.log
)

3. Rename the text file to mimetype.cmd.

4. Copy both mimetype.cmd and 7za.exe to your ePub folder and double-click mimetype.cmd. This will recursively search that folder as well as all sub-folders for ePubs and write the results to mimetype.log. (If you don't want to search subfolders change FOR /R to FOR).

5. Open mimetype.log with a text editor and look for ePub file names followed by a 2, which indicates ePub files with two mimetype file list entries.

[DiapDealer]

I realized I never got back to this after promising to take a look.

I don't see anything in the update checker portion (or the rest of the plugin) that could cause the plugin to be removed from the list of installed plugins.

I did notice, however, the exception when on Windows to ignore any errors when attempting to delete the temp directory.

# delete temp folder
ignore_errors = (sys.platform == 'win32')
shutil.rmtree(temp_dir, ignore_errors)

The reason errors are being raised on Windows is because you're cd'ing into temp directory in the beginning of the plugin.

# create mimetype file
os.chdir(temp_dir)
mimetype = open("mimetype", "w")
mimetype.write("application/epub+zip")
mimetype.close()

Windows won't let you delete the working directory (which is what 'temp_dir' became when you did the os.chdir call). Either change to a different directory before attempting to delete it, or never cd into it in the first place. That way you won't be leaving temp directories behind on Windows (ignoring the error does just that).

Something like:

# create mimetype file
mimetype = open(os.path.join(temp_dir, "mimetype"), "w")
mimetype.write("application/epub+zip")
mimetype.close()

...should work without actually cd'ing into the directory. You could still ignore errors on Windows if you like, but I guarantee there would be a lot less errors to ignore (and a lot less tmp folders left behind).

[JSWolf]

Originally Posted by Doitsu

Prerequisites: Download and install 7Zip (both the GUI version and the standalone console version).

Why not make a ZIP file of the batch file and all the bits needed? Unless you have 7-zip, you cannot get the standalone version because you then need 7-zip to get at it. And you also have to install 7-zip. This just makes it a PIA. The Kobo patches come with all the utilties needed to do the job and that just makes it much easier.

[Doitsu]

@DiapDealer: Thank you very much for your thorough code review!!!

                        

I've implemented all your suggestions and attached a new version.

Originally Posted by JSWolf

Why not make a ZIP file of the batch file and all the bits needed?

1. There are 32bit and 64bit versions of 7za.exe and some users simply won't run re-packaged executable files at all, unless they can validate the integrity of the files. Unfortunately, the developer didn't provide checksum files.

2. Usually, developers insist on their files being redistributed in their original archives.

[eschwartz]

2) Nu-uh.

7-Zip is open-source.

[rubeus]

Open source does not mean you can do what you want. It just says: You may have a look at the source.

[eschwartz]

Originally Posted by rubeus

Open source does not mean you can do what you want. It just says: You may have a look at the source.

Methinks you could use a read-through of the definition of Open-Source Software and related Wikipdia article.

tl;dr Eyeballing the source code is only one of several rights inherent in the standardized definition of Open-Source.

[rubeus]

SAP is Open Source. You can study it, but you cant distribute it due its license. In this case the definition of Open Source in Wikipedia(en) is bullshit.

[Blossom]

Doitsu has requested that his thread stay on topic so please take the open source discussion to the appropriate forum.

[crutledge]

I am totally unfamiliar with java.
Would you please provide me with a URL to Java 1.6 for Win 7 64bit.

Thank you.

[rubeus]

http://www.oracle.com/technetwork/ja...s-2133154.html

This is the Server Version. It does not require a server operating system, it perfectly runs on any client. The advantage: it does not come with any browser plugin not opening any security issues. Disadvantage: it just copies the file to your disk - you have to set environment variables and file associations for yourself.

[Doitsu]

Originally Posted by crutledge

I am totally unfamiliar with java.
Would you please provide me with a URL to Java 1.6 for Win 7 64bit.

The download link for the current 64bit Windows Java version is here.

After installing Java, you'll need to open a command prompt window and enter the following command to check the availability of Java:

java -version

If you don't get a version information message, you'll need to manually add the Java path to the Path environment variable following the instructions in the first post.

If you don't get a message about disabled browser extensions during the Java installation, you'll need to manually disable the Java browser extensions.

[eschwartz]

Easiest is to install this here: https://chocolatey.org/packages/javaruntime

Will download and install the latest version of java, auto-decline the bundled malware , and set the appropriate environment variables.
Chocolatey is a general must-have for Windows, really.

[wrCisco]

Hello everyone, I was trying the Epubcheck plugin but I got this error:

Running ePubCheck ... please wait
Traceback (most recent call last):
File "/Applications/Sigil.app/Contents/plugin_launchers/python/launcher.py", line 135, in launch
self.exitcode = target_script.run(container)
File "/Users/francesco/Library/Application Support/sigil-ebook/sigil/plugins/EpubCheck/plugin.py", line 43, in run
epc_errors = result[1].decode(os_encoding).split('\n')
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 136: ordinal not in range(128)
Error: 'ascii' codec can't decode byte 0xc3 in position 136: ordinal not in range(128)

I'm using Sigil 0.9.5 in MacOSX 10.11 and the bundled python interpreter.
If I change the value of os_encoding from locale.getpreferredencoding() (which seems to be US-ASCII unless I start the interpreter from the bash, in which case it is utf-8) to 'utf-8' all goes well, so it's not a big deal to me, but a more general solution would be probably better.

Thanks!

[Doitsu]

Thanks for the bug report!

Originally Posted by wrCisco

If I change the value of os_encoding from locale.getpreferredencoding() (which seems to be US-ASCII unless I start the interpreter from the bash, in which case it is utf-8) to 'utf-8' all goes well, so it's not a big deal to me, but a more general solution would be probably better.

Unfortunately, I don't have a Mac and os_encoding = locale.getpreferredencoding() works for both ًWindows and Linux machines. BTW, the line is required because Windows machines don't use utf-8 as the default locale. (My machine uses cp1252.)
I'll fix this when the IDPF releases the next ePubCheck version.