A bug in Sigil?

[RbnJrg]

Hi Kevin and Diap;

In my epubs (3), sometimes I add a javascript to the nav.xhtml file and when I do that, the property "scripted" is not added in the .opf file. That is not a big problem because I can add that property by hand (and I do that), but the issue arises when Sigil open the epub because the property added by me, is erased; I have to do the thing every time that Sigil opens the ebook. Could you fix that in the next Sigil version?

Many thanks in advance.

[DNSB]

If you use the ePub3 Tools item in the Tools menu and click on Update Manifest Properties, does this correct the error(s) for you. I haven't tried it with a js linked to the nav document, but it works for standard html documents though I seem to recall at one point, reading that what the navigation document can contain is limited and javascript was not listed as a allowed item. Now if I could find that reference...

Edit: The reference basically stated that since the navigation document is intended to be machine readable, javascript should not be used. Instead make the navigation document hidden and create an dedicated for human use html ToC. So perhaps a nav.xhtml navigation document and a toc.xhtml human intended ToC.

[KevinH]

I will look into it. That property should not be erased if present on loading of any epub if it already exists in the opf manifest for any file.

[KevinH]

Yes the SetNavResource() in the OPFResouce is the issue as it just assigns the nav overwriting any other properties.

Also the Epub3 Update Manifest Properties Tool needs to be allowed to work on the nav.

I will try to get a fix for both done soon.

[RbnJrg]

Originally Posted by DNSB

If you use the ePub3 Tools item in the Tools menu and click on Update Manifest Properties, does this correct the error(s) for you. I haven't tried it with a js linked to the nav document, but it works for standard html documents though I seem to recall at one point, reading that what the navigation document can contain is limited and javascript was not listed as a allowed item. Now if I could find that reference...

Edit: The reference basically stated that since the navigation document is intended to be machine readable, javascript should not be used. Instead make the navigation document hidden and create an dedicated for human use html ToC. So perhaps a nav.xhtml navigation document and a toc.xhtml human intended ToC.

Yes, but the issue is that the property is erased when Sigil loads the epub. And Epubcheck allows the property "scripted" for nav.xhtml (in the .opf, property="nav scripted").

[DNSB]

Originally Posted by RbnJrg

Yes, but the issue is that the property is erased when Sigil loads the epub. And Epubcheck allows the property "scripted" for nav.xhtml (in the .opf, property="nav scripted").

I'm not disagreeing that allowing the scripted property for the nav document is not prohibited. What I got from the article I read and made a couple of notes from is that the authors did not recommend using javascript in the nav document since it could cause issues with the renderer's ability to parse the file. Their suggested workaround was a vanilla navigation document that was hidden and a ToC file that could be fancified without needing to worry about the renderer parsing it.

[RbnJrg]

Originally Posted by DNSB

I'm not disagreeing that allowing the scripted property for the nav document is not prohibited. What I got from the article I read and made a couple of notes from is that the authors did not recommend using javascript in the nav document since it could cause issues with the renderer's ability to parse the file. Their suggested workaround was a vanilla navigation document that was hidden and a ToC file that could be fancified without needing to worry about the renderer parsing it.

So far, I never had an issue for including a .js file in nav.xhtml. Sometime ago, I used both (in the same epub), the nav.xhtml an a xhtml toc, but as some ereaders had issues about not to show the nav.xhtml, I decided to do full visible the nav and not to include the toc. And as told you before, I didn't have any issue so far by including there a .js file (Readium and Webkit seem to handle that very well). Maybe this happens because the script doesn't change the structure of nav.xhtml; it only changes some .css styles.

[DNSB]

Originally Posted by RbnJrg

So far, I never had an issue for including a .js file in nav.xhtml. Sometime ago, I used both (in the same epub), the nav.xhtml an a xhtml toc, but as some ereaders had issues about not to show the nav.xhtml, I decided to do full visible the nav and not to include the toc. And as told you before, I didn't have any issue so far by including there a .js file (Readium and Webkit seem to handle that very well). Maybe this happens because the script doesn't change the structure of nav.xhtml; it only changes some .css styles.

That is quite possible. The sample shown in the item I read modified the structure of the human readable depending on the size of the screen so a large screen (iPad for instance) would show the pagelist while a smaller screen (Kobo Glo) would only show the ToC. I can't find the original article which was linked off Liz Castro's Pigs, Gourds and Wikis page quite a few years back though I did add to my 404 collection. Sheesh, it's only been 12 years...

[KevinH]

This should now be fixed in master.
Thank you for your bug report!

[RbnJrg]

Originally Posted by KevinH

This should now be fixed in master.
Thank you for your bug report!

Many thanks Kevin!