sigil 7.4 and maverick crashing crashing crashing

[KevinH]

Hi,
It is official ... user_none has added the new preserve entities patch to Sigil's github trunk. I still need to find and fix tidy to stop it from making svg single image into empty pairs since that freaks out kindlegen. If anyone else has any minor changes they need to related to tidy, let me know and I will try to track those tidy issues down at the same time I am tracking my own.

KevinH

[JSWolf]

Is there any way to fix Sigil so there is an option to leave the structure of the ePub alone without changing it?

[KevinH]

Hi,

You would have to be much more specific. Exactly what structure are you talking about? Please provide concrete examples of something Sigil changes when tidy is off, that you think it should not. Please note that Tidy itself makes many changes to make the code at least be technically correct ... no improperly nested tags, no missing end tags, etc. Those changes must be made so that wysiwyg editing has some chance to be successful.

Originally Posted by JSWolf

Is there any way to fix Sigil so there is an option to leave the structure of the ePub alone without changing it?

KevinH

[DiapDealer]

Originally Posted by KevinH

Hi,

You would have to be much more specific. Exactly what structure are you talking about? Please provide concrete examples of something Sigil changes when tidy is off, that you think it should not. Please note that Tidy itself makes many changes to make the code at least be technically correct ... no improperly nested tags, no missing end tags, etc. Those changes must be made so that wysiwyg editing has some chance to be successful.



KevinH

I think he's talking about the structure of the zip archive itself. Sigil silently rearranges all archives into a homogeneous folder hierarchy (regardless of Tidy settings). Has always done so. From what I've seen ... way too much of what Sigil does is predicated on that specific hierarchy being present to be able to easily "fix" it. That would require an overhaul IMO.

[JSWolf]

Originally Posted by KevinH

Hi,

You would have to be much more specific. Exactly what structure are you talking about? Please provide concrete examples of something Sigil changes when tidy is off, that you think it should not. Please note that Tidy itself makes many changes to make the code at least be technically correct ... no improperly nested tags, no missing end tags, etc. Those changes must be made so that wysiwyg editing has some chance to be successful.



KevinH

What Sigil does is it moves the files around to it's own idea of where they belong.

OEBPS/Styles
OEBPS/Images
OEBPS/Fonts
OEBPS/Text

So even if you have it structured how you want it, Sigil will change that to what it wants. I want an option to allow the structure to remain as is.

[Doitsu]

Originally Posted by KevinH

If anyone else has any minor changes they need to related to tidy, let me know and I will try to track those tidy issues down at the same time I am tracking my own.

I don't know if Sigil uses the XML version of Tidy when generating/formatting the .ncx file, but I've noticed that it'll delete manually inserted <pageList> nodes if the user regenerates the NCX TOC via Generate Table of Contents...
(You can test this with my pageList test file from the KindleUnpack thread.)

I understand that changing the NCX code so that it'll preserve existing <pageList> nodes isn't really worth the effort, since hardly anybody is using pageLists anyway, but it would be nice to have Sigil at least display a deletion warning if the user selects Generate Table of Contents... from the menu and the .ncx file contains <pageList> nodes (or other non-TOC nodes).

[eschwartz]

Originally Posted by JSWolf

What Sigil does is it moves the files around to it's own idea of where they belong.

OEBPS/Styles
OEBPS/Images
OEBPS/Fonts
OEBPS/Text

So even if you have it structured how you want it, Sigil will change that to what it wants. I want an option to allow the structure to remain as is.

Use calibre. Don't expect Sigil to turn into calibre ebook-edit -- there'd be no point in having it.

[KevinH]

Why?

I can understand how not touching the xhtml makes it easier to edit in CodeView and helps the person creating the ebook. But what does the layout of the epub zip structure have to do with it? It will not be seen and does not impact your ability to find and edit files in Sigil, since the gui handles that.

You need to ask yourself: "Is this really truly important in the big scheme of things?"

If so why don't you simply create a repackager program in python that moves things to the format you want, updates the opf, and the relative links. It is not that hard to do.

Since I am already volunteering my time in a number of projects, I was only looking to handle serious bugs inside of Sigil and things that impact getting real work done, and not things that could be handled in post processing.

Sorry,

KevinH

Originally Posted by JSWolf

What Sigil does is it moves the files around to it's own idea of where they belong.

OEBPS/Styles
OEBPS/Images
OEBPS/Fonts
OEBPS/Text

So even if you have it structured how you want it, Sigil will change that to what it wants. I want an option to allow the structure to remain as is.

[KevinH]

Hi Doitsu,

My guess is, it assumes that when it says "regenerate" it, you already know it will throw out everything there and build a new one by parsing for titles in each xhtml page, and the contents of h1, h2, etc, tags.

But not everyone thinks of "regeneration" in the same way. So, I agree, if the ncx exists, it should warn you that everything in the ncx will be lost upon regeneration and ask you if you want to continue.

That is something I can look creating a patch for when I get some free time.

Take care,

KevinH

Originally Posted by Doitsu

I don't know if Sigil uses the XML version of Tidy when generating/formatting the .ncx file, but I've noticed that it'll delete manually inserted <pageList> nodes if the user regenerates the NCX TOC via Generate Table of Contents...
(You can test this with my pageList test file from the KindleUnpack thread.)

I understand that changing the NCX code so that it'll preserve existing <pageList> nodes isn't really worth the effort, since hardly anybody is using pageLists anyway, but it would be nice to have Sigil at least display a deletion warning if the user selects Generate Table of Contents... from the menu and the .ncx file contains <pageList> nodes (or other non-TOC nodes).

[Doitsu]

Originally Posted by KevinH

But not everyone thinks of "regeneration" in the same way. So, I agree, if the ncx exists, it should warn you that everything in the ncx will be lost upon regeneration and ask you if you want to continue.

Thanks for looking into this!

Ideally, Sigil should only display the warning if the .ncx contains content that Sigil cannot automatically generate. E.g. if a Calibre generated .ncx file contains TOC links to non-heading tags or manually added pagelist nodes.

Otherwise Sigil users who don't use pagelists, but like re-generating the TOC as they add new content to a book might get upset about Sigil warning them that generating the TOC will overwrite the existing .ncx file.
I.e., if you display the warning indiscriminately, some users will most likely request a patch that suppresses this warning.

As I've said, this is a nice-to-have feature that very few users will benefit from and should only be implemented if it can be done with minimal code changes and without alienating the majority of Sigil users.

[theducks]

Originally Posted by KevinH

Hi,

You would have to be much more specific. Exactly what structure are you talking about? Please provide concrete examples of something Sigil changes when tidy is off, that you think it should not. Please note that Tidy itself makes many changes to make the code at least be technically correct ... no improperly nested tags, no missing end tags, etc. Those changes must be made so that wysiwyg editing has some chance to be successful.



KevinH

DiapDealer has it.

Sigil has a single OEBS folder structure (Text, Images, Styles)

You can NOT import multiple HTML safely with any overlapping path filenames because Sigil flattens the import, then shoves those into the above structure. All filenames must be unique to survive a Sigil import.

[eschwartz]

I think it makes sense regardless to not show this warning all the time, and allow people to ignore it.

calibre, for instance, allows on pretty much any warning popup to check a box "Don't show this message again".

At the very least, it could be shown only the first time in each session.

[mrmikel]

Unlike everyone else, it seems, I would be in favor of a warning that I am going to knock out an existing file so I get a chance to save myself from myself.

[Tex2002ans]

Originally Posted by KevinH

Since I am already volunteering my time in a number of projects, I was only looking to handle serious bugs inside of Sigil and things that impact getting real work done, and not things that could be handled in post processing.

Keep up the good fight.

Don't let any Negative Nancies suck your soul!

The entities patch sounds fantastic.

I personally only go from a big list of entities -> unicode char, and I set up a large group in "Saved Searches" to handle all of the ones I come across. The oddities I have not included deserve closer inspection (might be an OCR mistake, or typo, or something else).

[eschwartz]

Originally Posted by mrmikel

Unlike everyone else, it seems, I would be in favor of a warning that I am going to knock out an existing file so I get a chance to save myself from myself.

Wrong, I am in favor as well.

Although it should be able to be suppressed.