Tolerance and parsing mistakes - Page 5

roger64 · 01-30-2016, 04:36 AM

Hi

I keep my promise to DiapDealer.

After I introduced -unwillingly- a mistake in one EPUB, I tried to open it with Sigil to perform further tasks. I ask Sigil not to perform any previous control on my ebooks, because some of them are in EPUB3 format and I had been advised to stay away from such a foolish enterprise.

Sigil refused to open it. It's seemed like a trivial mistake, a punctuation sign ; which should never have been there.This same EPUB is opened without problem with the Calibre Editor.

Please find enclosed the warning message from Sigil, a screenshot of the mistake (with a red arrow) and the EPUB. The EPUB is really a work in progress and was not intended to be seen but I understand it may be useful for analysis.

DiapDealer · 01-30-2016, 07:38 AM

When it comes to epub3, I think you're simply going to want to wait on the next version of Sigil to do much of anything. It will be more robust in its abilities to recognize and edit epub3s. Your error indeed stems from the fact that you've instructed Sigil NOT to fix a parsing error it could very easily fix if you allowed it to.

But you're not entirely wrong to turn off the "Mend" feature (for editing Epub3s you've edited with calibre's editor) in Sigil 0.9.2, either. calibre's editor doesn't use (and will often remove) doctypes. Sigil 0.9.2 will assume a document without a doctype is xhtml 4.01. That will definitely break a valid epub3 that doesn't contain doctypes if you allow it (Sigil) to Mend code.

Rest assured that Sigil 0.9.3 will be able to handle this type of situation (IFyou turn the "Mend" feature on and allow it to do its thing). 0.9.3 uses the opf package version to handle doctype decisions now. It won't break epub3s coming from calibre's editor with no doctypes. But you WILL have to turn on "Mend", at that time, if you want Sigil to be more tolerant of simple parsing errors. It can't fix errors you instruct it not to fix. Mend is what makes Sigil more tolerant of simp!e parsing errors. Turning it off is instructing Sigil to be intolerant..

I hope that explains things. The only "real" danger in leaving Mend Code turned on in Sigil 0.9.2 is when opening Epub3s with no doctype. And that's fixed in 0.9.3. You will also be able do a lot more editing of epub3s with 0.9.3. 0.9.2 is very limited in what it can do with epub3s Admittedly so.

Thanks for the sample. We'll check to see if Sigil's xml/opf parsing routines can accomodate a syntax error of that nature. Just understand that Sigil is going to be less forgiving of syntax errors in an opf file than it it is of markup in an html file. The OPF is, after all, the beating heart of an epub. A lot depends on it being right.

Doitsu · 01-30-2016, 08:41 AM

@roger64: While Sigil 0.9.2 requires you to manually add ePub3 properties to the .opf file, you won't have to do this in Sigil 0.9.3, which has a menu option that will automatically add mandatory ePub properties such as properties="svg".

roger64 · 01-30-2016, 08:45 PM

@DiapDealer, Doitsu

Thanks for your explanations.

eschwartz · 01-30-2016, 09:44 PM

Diap -- can you please change that error message to include the recommendation that turning on Mend Code will allow the EPUB to be opened?

It might prevent people from making this mistake again.

KevinH · 01-30-2016, 10:00 PM

Hi,
Mend won't actually fix a not well formed content.opf. The content.opf parser, like most pure xml parsers, has never been very forgiving. I will look to see if upon error on import we try to pass the content.opf through an xml repair parser of its own as Mend is gumbo/xhtml specific. We have closed the tree to changes until the official Sigil-0.9.3 release hopefully quite soon now. So any changes will have to wait until the next release.

01-30-2016, 07:38 AM	#62
DiapDealer Grand Sorcerer Posts: 28,891 Karma: 207182180 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	When it comes to epub3, I think you're simply going to want to wait on the next version of Sigil to do much of anything. It will be more robust in its abilities to recognize and edit epub3s. Your error indeed stems from the fact that you've instructed Sigil NOT to fix a parsing error it could very easily fix if you allowed it to. But you're not entirely wrong to turn off the "Mend" feature (for editing Epub3s you've edited with calibre's editor) in Sigil 0.9.2, either. calibre's editor doesn't use (and will often remove) doctypes. Sigil 0.9.2 will assume a document without a doctype is xhtml 4.01. That will definitely break a valid epub3 that doesn't contain doctypes if you allow it (Sigil) to Mend code. Rest assured that Sigil 0.9.3 will be able to handle this type of situation (IFyou turn the "Mend" feature on and allow it to do its thing). 0.9.3 uses the opf package version to handle doctype decisions now. It won't break epub3s coming from calibre's editor with no doctypes. But you WILL have to turn on "Mend", at that time, if you want Sigil to be more tolerant of simple parsing errors. It can't fix errors you instruct it not to fix. Mend is what makes Sigil more tolerant of simp!e parsing errors. Turning it off is instructing Sigil to be intolerant.. I hope that explains things. The only "real" danger in leaving Mend Code turned on in Sigil 0.9.2 is when opening Epub3s with no doctype. And that's fixed in 0.9.3. You will also be able do a lot more editing of epub3s with 0.9.3. 0.9.2 is very limited in what it can do with epub3s Admittedly so. Thanks for the sample. We'll check to see if Sigil's xml/opf parsing routines can accomodate a syntax error of that nature. Just understand that Sigil is going to be less forgiving of syntax errors in an opf file than it it is of markup in an html file. The OPF is, after all, the beating heart of an epub. A lot depends on it being right. Last edited by DiapDealer; 01-30-2016 at 08:05 AM.

01-30-2016, 08:41 AM	#63
Doitsu Grand Sorcerer Posts: 5,763 Karma: 24088559 Join Date: Dec 2010 Device: Kindle PW2	@roger64: While Sigil 0.9.2 requires you to manually add ePub3 properties to the .opf file, you won't have to do this in Sigil 0.9.3, which has a menu option that will automatically add mandatory ePub properties such as properties="svg". Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Understanding two reported mistakes	roger64	Editor	11	02-19-2014 03:47 AM
Manage gramm. mistakes?	Dalirose	Onyx Boox	1	01-01-2014 01:22 PM
5 Mistakes Of New Fiction Writers	VydorScope	Writers' Corner	67	07-08-2012 08:36 AM
Noob install mistakes	Shiryas	Calibre	2	04-05-2010 02:30 AM
Mistakes in The Hobbit	AlexBell	News	58	05-12-2009 09:26 PM

01-30-2016, 08:45 PM	#64
roger64 Wizard Posts: 2,625 Karma: 3120635 Join Date: Jan 2009 Device: Kindle PW3 (wifi)	@DiapDealer, Doitsu Thanks for your explanations.

01-30-2016, 09:44 PM	#65
eschwartz Ex-Helpdesk Junkie Posts: 19,421 Karma: 85400180 Join Date: Nov 2012 Location: The Beaten Path, USA, Roundworld, This Side of Infinity Device: Kindle Touch fw5.3.7 (Wifi only)	Diap -- can you please change that error message to include the recommendation that turning on Mend Code will allow the EPUB to be opened? It might prevent people from making this mistake again.

01-30-2016, 10:00 PM	#66
KevinH Sigil Developer Posts: 9,072 Karma: 6361556 Join Date: Nov 2009 Device: many	Hi, Mend won't actually fix a not well formed content.opf. The content.opf parser, like most pure xml parsers, has never been very forgiving. I will look to see if upon error on import we try to pass the content.opf through an xml repair parser of its own as Mend is gumbo/xhtml specific. We have closed the tree to changes until the official Sigil-0.9.3 release hopefully quite soon now. So any changes will have to wait until the next release.

Advert

Advert