[Plugin] ePub3-itizer - epub3 output plugin for Sigil

[KevinH]

You can use regular expressions in Search and Replace to only remove charset inside link tags. Sigil supports this. I am sure someone here can help with creating the right regular expression.

You could also checkout just the plugin.py file from master from my github site and use it to replace the one you have on your computer and all should just work.

If you need help with either, just let us know.

[Julie34]

Yes, I could do a regex (I manage a bit, not much) but for such a small search and replace it is quick (there is rarely 70 chapters with the normal search/replace. I mentioned it for over peoples in my case in the purpose they notice this point.

Thank you really for your kindness.

[LauraB7]

I am brand new to the forums and hoping someone can help. I want to convert some old files to EPUB 3 but can't seem to get the plugin to work at all.

A screenshot of the error message is attached.

I am sure this is something simple that I can't see. Still, I apologize if this is a basic question.

[AlanHK]

Originally Posted by LauraB7

I am sure this is something simple that I can't see.

Have you installed Python? This, like other plugins, requires it.

[Doitsu]

Originally Posted by LauraB7

I am brand new to the forums and hoping someone can help. I want to convert some old files to EPUB 3 but can't seem to get the plugin to work at all.

macOS users will need to install the ActiveState TCL 8.6.6.8606 distribution.
(This package is required to display the plugin GUI.)

[DiapDealer]

There's a problem with the stock tcl implementation on Macs. The advice to install the ActiveState version of TCL that @Doitsu mentioned has been in Sigil's release notes for several versions.

Attention Mac OS X Users
Per the recommendation of www.python.org (see https://www.python.org/download/mac/tcltk/), due to bugs and the age of Apple's internal Tcl library, you should also download and install ActiveState's ActiveTcl Community Edition. To get the latest bug fixes, Sigil-0.9.8's embedded Python has been linked with the very latest version: Tcl 8.6.4.1

We may not redistribute this Package with Sigil due to ActiveState's binary non-redistribution policy.

So please, if you have not already done so download and install this version of ActiveState's Active Tcl Community Edition, until we get a chance to do our own build of Tcl 8.6.4.1 from source for future releases. ActiveState is a primary contributor to www.python.org and a respected and trusted source for python and tcl/tk related binaries on both Windows and Mac OS X.

See: http://www.activestate.com/activetcl/downloads
Version: 8.6.4.1

[LauraB7]

Thanks everyone. As I suspected, it was a simple error owing to me not reading the Github notes thoroughly. I appreciate your patience with me!

[DiapDealer]

Originally Posted by LauraB7

Thanks everyone. As I suspected, it was a simple error owing to me not reading the Github notes thoroughly. I appreciate your patience with me!

No problem. The note is easy to miss. I added a sticky message to the Sigil plugin subforum to hopefully get it a little more exposure.

[AlanHK]

I converted an epub2 to epub3, and it works fine in most situations.
However, ePubcheck complained that there were some named entities --   -- instead of numeric, and Azardi reader choked on those too.

Not hard to fix manually, but could the plugin convert such characters?

(Personally I much prefer named entities, don't know why epub3 is against them. Maybe ePub4 will be an all binary format.)

[DiapDealer]

Originally Posted by AlanHK

(Personally I much prefer named entities, don't know why epub3 is against them. Maybe ePub4 will be an all binary format.)

It's not epub3 that is "against" named entities, it's HTML5. Also note that this plugin does convert all named html entities to their numeric counterparts. if emsp is sticking around, it's likely because you have that named entity entered explicitly in the list of entities to preserve in your prefs. In that scenario, the numeric entity will be changed back to named whenever you make edits and/or save the new epub3. If you're going to be working a lot with epub3, you'll want to switch your Preserve Entities list to numeric. They're valid in both flavors of epub.

[KevinH]

Yes exactly. All named entities are converted to their numeric equivalents by the plugin. If you then load it back into Sigil without fixing your preserved entities first, Sigil will happily convert it back since effectively you asked it to.

[AlanHK]

Originally Posted by DiapDealer

It's not epub3 that is "against" named entities, it's HTML5.

Well, same sentiment applies.
Can't see any benefit at all in forcing numeric over mnemonics.

Originally Posted by DiapDealer

because you have that named entity entered explicitly in the list of entities to preserve i.

Yes. Mea culpa. I didn't know that "preserve" also meant "change back".

[DiapDealer]

Originally Posted by AlanHK

Well, same sentiment applies.
Can't see any benefit at all in forcing numeric over mnemonics.

It's just spec. Get used to not seeing much benefit to many of the ongoing changes to it.

[KevinH]

FWIW. There are DOS and attacks based on named entities and recursion that are not possible with numeric entries.

Thus one good reason for the change in the spec to limit potential attack avenues.

[AlanHK]

Originally Posted by KevinH

FWIW. There are DOS and attacks based on named entities and recursion that are not possible with numeric entries.

Has one of these possible attacks ever occurred in reality?
In an ebook?

Anyway, not blaming anyone here. Just bitching about a change for the marginal benefit of automatic processing at the cost to human editors; a change that makes manual editing a bit harder and more prone to error. As more entities are numeric, it becomes a stew of digits and impossible to make sense of by eye.