Please Help Test Sigil 2.7.5 Beta

[BetterRed]

No problems with cursory checks of Folder In and Out, Shady Characters and DOCX Import.

Nothing more to add to the child windows window mgt issue. Except, I suspect many users would move the non-modal child window to a second monitor if it was in the way. The handful of people I know who edit books for a living have multiple monitors on their desks.

BR

[KevinH]

Originally Posted by BetterRed

No problems with cursory checks of Folder In and Out, Shady Characters and DOCX Import.

Nothing more to add to the child windows window mgt issue. Except, I suspect many users would move the non-modal child window to a second monitor if it was in the way. The handful of people I know who edit books for a living have multiple monitors on their desks.

BR

Thanks for testing those plugins! I had the same thought about multiple monitors, or just offset overlapping the two windows and popping back and forth but on some laptops, both are hard to do at times. Leaving the solution up to the user and using add-ons makes sense and users can still do that.

But, fwiw, I think I have found a way to auto close stale minimized windows after changing books. I will try that first before tossing in the towel.

Thanks

Update:

With my most recent commit, old minimized non-modal dialogs are now properly closed upon loading a new book preventing the stale dialog problem. I have pushed this change to master.

So as it stands, people on Windows and Mac can easily use their own supplemental tools to add in titlebar and window resizing abilities if they so desire. Windows and Linux users (depending on window manager used) can use the Minimize if desired. It just will not work on MacOS without making the window being created as a top-level window (parent = nullptr) which I am not willing to do. People who do not like allowing dialogs to minize need not use that feature.

Hope this is enough.

[BetterRed]

I have found a way to auto close stale minimized windows after changing books.

I remember a time when OS/X would roll up a window if you double clicked the title bar, not sure if there was a stay on top.

And, I can't recall which one, but one of the GUI OS's from the 80s had roll up, stay on top, minimise to desktop etc - wasn't Lisa, might have been VisiOn or Gem.

BR

[KevinH]

Originally Posted by retiredbiker

I got the 2.7.5 appimage and tried a couple of epubs. Exercised the ODTImport,, Check Internal Links, Remove unused selectors, Epubcheck, Footnote manipulator, InsertImageSVG, Punctuation Smarten, Remove Inline Styles. Got no errors, but on a quick test some of the plugins had little or nothing to do; still they ran with no errors. Running on Pop_os.

Thanks for testing and reporting back.

[jugaor]

Hi.
Windows 11, tried the plugins:

Access-Aide
CSSValidator
DOCXImport
EnvVarEditor
Epub2LegacyMetaData
Epub2Text
FixHeadStyles
InsertImageSVG
KindleImport
QuickPrefsEdit
SubsetFonts

No issues so far.

[KevinH]

Originally Posted by jugaor

Hi.
Windows 11, tried the plugins:

Access-Aide
CSSValidator
DOCXImport
EnvVarEditor
Epub2LegacyMetaData
Epub2Text
FixHeadStyles
InsertImageSVG
KindleImport
QuickPrefsEdit
SubsetFonts

No issues so far.

Thank you. Glad to hear it.

[recook]

Hi,

running the 2.7.5 Beta3 Appimage on LinuxMint 22.3 (Ubuntu 24.04),
i tried to import into the open EPUB file a HTML page containing a big <script> tag in the Head.
Like before with 2.7.0, the very same error persists:

> The following file(s) were not loaded due to invalid content or not well formed XML:
>
> /.../1_.html (line 128: Expected ';', but got '='.)

The lines 127-129 in there read

				pushAppServerKey: '',
				url: {
					fullBase: 'https://www.alternatehistory.com/forum/',

There's not a '=' to be seen.

The builtin Ebook editor of Calibre on import of this file sometimes presents the same error, sometimes just imports it, so that i can remove the <script>.*</script>.

[KevinH]

Your script tag contains xhtml reserve characters such as '<', '&', or '>' that confuses the parser because it is not protected by a CDATA tag pair:

'<![CDATA[' and ']]>'

So edit that file before import and add that cdata tag pair inside that script tag or
alternatively you could edit that html file and extract that script and paste that script into its own file and just properly link to it.

Sigil will not allow you to import it until it is fixed mainly for security reasons.

If this is something you run into often, then the best solution is to use a Sigil plugin to add selected files after removing their broken script tags.

Sorry.

[recook]

Hi,

as indeed it does, to my chagrin, in the 2nd and last line of this excerpt:
<code>
<script>
XF.ready(() =>
{
XF.extendObject(true, XF.config, {
//
userId: 0,
enablePush: false,
pushAppServerKey: '',
url: {
fullBase: 'https://www.alternatehistory.com/forum/',
basePath: '/forum/',
css: '/forum/css.php?css=__SENTINEL__&s=14&l=1&d=1771704368',
</code>

Since i'm trying to import multiple (about 50 to 100) HTML files like this, i'd very much prefer a warning instead of an error, so that i then can automatically remove the `<script>.*</script>` from all of them, instead of painstakingly for every single file.
Since the Calibre ebook editor finally consented to the import, that's what i was allowed to do, in the end.

[KevinH]

Again, if you need custom import of non-well formed html, using a plugin is the way to go in Sigil. Importing not well formed code into an existing epub is generally not a good idea, especially those containing large javascripts directly in a script tag.

It is relatively trivial to create a plugin to strip out script tags in the head from html using regex. You could easily multi-select as many files to process at the same time and import them simultaneously. It would be a good first project for learning how to create plugins.

Happy to provide hints if you get stuck. Otherwise as you said, you can use calibre.