How to make Sigil live (in calibre)

[samy2]

Originally Posted by kovidgoyal

@samy2: While that scheme works well in your particular case, consider when you have two files, already named:

part000_1.html
part000_2.html

Now when you split part000_1.html you will get
part000_1.html
part000_3.html
part000_2.html

So that the names become out of order.

That is not the problem. I have should explain it in diffrent way.

Now when you split part000_1.html you will get:

part000_1.html
part000_1_1.html
part000_1_2.html
part000_1_3..html

So it is no disorder!
Something like this is better than a monster like " "part0000_1_1_1_1_1_1"

[Katsunami]

Originally Posted by samy2

That is not the problem. I have should explain it in diffrent way.
[...]
Something like this is better than a monster like " "part0000_1_1_1_1_1_1"

After batch renaming is implemented, it doesn't really matter how the splits are called. Then you just split, and in the end rename everything at once.

There are two things I've not seen an answer to, or I've missed it:

- Is it supposed to be correct that the GUI must be restarted if you add or change a shortkey? It doesn't work or show up if you don't restart.
- Is it possible to navigate using the TOC, somehow?

[jackie_w]

Originally Posted by jackie_w

BTW I did have occasional problems today with the Preview not always tracking to the current match. If I can reliably reproduce it in a test file I'll give more details.

I've been able to reproduce this. I will send you a PM with a link to the epub. (Perhaps you can let me know once you've downloaded it. It's a copyright file and I don't want DropBox closing my account for illegal filesharing. I've renamed it kovid.zip, so it just needs renaming back to kovid.epub)

Anyway these are the steps to reproduce:

1. Open epub and select first text file.
2. Find: atx1 (this is a css class in the book)
3. Cycle 'Down' through the 5 matches. On the first pass there are no problems, but if you cycle again there is a problem where match 1 (of 2) in chap8 (CHAIRMAN, FINANCE...) goes to the correct place in Code View but the Preview is incorrectly positioned at the top of chap8. Keep cycling and the same match has the same problem each time.
4. If the search direction is changed to 'Up' there is a similar problem except the 'problem match' is match 2 of 2 in chap8 (403rd MEETING...)

Hope you can shed some light and thanks for the fixes to last_modified and highlighting current file in File Browser

[kovidgoyal]

@jackie: Thanks, it's not actually a sync problem. What happens is that when you go back to the previously opened editor, the cursor position remains the same as it was before, which means no sync request is sent causing the preview to remain at the top of the file. I'll look into the best solution for that, that is not too inefficient. Feel free to delete the file.

@Katsunami: Shortcut changes no longer require a restart. Update to the last beta I posted. And as I mentioned previously (to arspr, I think), a TOC View is on my todo list.

EDIT: Fixed: https://github.com/kovidgoyal/calibr...808259a5388f7d

[Divingduck]

Kovid,

I have some unknown property names in CSS when I do a check for:

panose-1:

Spoiler:

@font-face {
  src: url(../Fonts/Arial-Unicode-MS.ttf);
  panose-1: 2 11 6 4 2 2 2 2 2 4;
  font-family: "Arial Unicode MS";
  }

Hyphenation: adobe-hyphenate, hyphens, -moz-hyphens,, -webkit-hyphens

Spoiler:

h1 {
  font-weight: bold;
  font-size: 1.5em;
  adobe-hyphenate: none;
  hyphens: none;
  -moz-hyphens: none;
  -webkit-hyphens: none;
  text-align: center;
  margin-top: 0;
  margin-bottom: 1.2em;
  margin-left: 0;
  margin-right: 0;
  text-indent: 0;
}

An other observation:
When I move the inspector window with the mouse the window flips back to the old position. I think, it will be a better choice to have no moving or undock for this window, as long as the window will lose the content by doing this (only a guess).

[arspr]

Originally Posted by Divingduck

Kovid,
I have some unknown property names in CSS when I do a check for:

I had also noticed it. But I think that implementing ALL the possible and personalized properties the whole world can have will be completely impossible. I suppose Kovid can only implement checks for the "standard" ones. A small price for so great tool.

Originally Posted by arspr

• I know that you are going to ignore entities like   which is probably the best thing to do. But nevertheless, why   generates a warning/error but, as example, — or ² don't? Irrelevant? Or the symptom of another possible hidden issue?

Kovid, this issue is still present. I attach a screenshot ("Ignored entities.jpg") with some example entities in the beginning of line 20. No warning about them, but I get the warning with  . Intended? In that case, why is there a distinction between   and other ones?

Originally Posted by kovidgoyal

1) Check Book now checks CSS as well, in stylesheets and inside <style> tags

Don't kill me, please

You forgot about styles inside tags... I attach a screenshot ("You forgot CSS inside tags....jpg") of a real book where I have deleted a ";" in the middle of it. (Line 12 between "center" and "background-color").



And then even another suggestion request (no real hurry about it). The next screenshot ("CSS Warning Texts.jpg") shows errors because of deleted ";" (lines 3 and 10) and a deleted "}" (line 19). I really, but really really, think that the error texts are completely UNfriendly. If you could somehow "translate" them to common world information, it would be really welcomed.


EDITED ----

And another possible issue. Please check "Before.jpg" and "After.jpg" (a Beautify action). I've deleted the ending </i> in line 15. When Beautify (or Fix HTML) is launched an extra pair of "<i> </i>" appears outside the paragraph (lines 17 - 21).

[kovidgoyal]

@Divingduck: Those are all non-standard css properties. You will be warned about any property not in the CSS 2.1 standard (plus a few from CSS 3). Extra CSS properties are usually harmless, so just ignore the warning, if you dont want to delete them.

@arspr: If you have modified text in an editor, all entities are automatically converted to unicode before checking, which is why you will not get message about them. Close and re-open the editor and you will see that.

I'm afraid the error messages come from the cssutils library, I really dont have the time/interest to make them more friendly. Patches are welcome.

You really should not be using style attributes, but remind me about it after the release and I will see if checking can be added for them.

The HTML 5 spec requires unclosed <i> tags to render the remaining text in italic, which is why you get an extra pair of <i> tags. It's a silly requirement, but it's in the spec, probably for legacy compatibility requirements. Try viewing the following html snippet in a modern browser,

<!DOCTYPE html>
<html>
    <head>
        <title>test</title>
        <meta http-equiv="content-type" content="text/html; charset=utf-8"/>
    </head>
    <body>
    <p>some <i>italic</p> after text  will also be italic 
    </body>
</html>

[Divingduck]

Originally Posted by kovidgoyal

@Divingduck: Those are all non-standard css properties. You will be warned about any property not in the CSS 2.1 standard (plus a few from CSS 3). Extra CSS properties are usually harmless, so just ignore the warning, if you dont want to delete them.

@Kovid: It is fine to me and I am happy that you have integrated this check at all.
I was reporting these points, because the font setting came out form Calibre by integrating a reduced font set (a mix of European and Japanese characters) and I guess there will come up questions about findings if users integrate fonts with Calibre and go then to the editor to make small adjustments.
Well, Hyphenation becomes more and more common in todays reader. This was why I brought it up. But you are right, it is as well a non-standard.

Only a quick thinking around: Maybe you can integrate some when in the future a switch to hide some of the findings like uncritical warning informations.

[kovidgoyal]

@Divingduck: Good point about panose-1, I have special cased it so that the checker does not report it as an unrecognized property.

Unfortunately, the css library calibre uses -- cssutils is not very good for this purpose. The Book Check tool actually just intercepts its log messages and converts them into error entries. Therefore, there is no easy way to ignore some types as errors, as the errors are not really identified in any concrete way. That is why the error messages are also fairly cryptic. Still, I think having the ability to detect errors in CSS is valuable, despite the sub-optimal implementation.

[Divingduck]

Originally Posted by kovidgoyal

Still, I think having the ability to detect errors in CSS is valuable, despite the sub-optimal implementation.

Yes, absolutly true.

[kovidgoyal]

FYI, A table of contents viewer and checking for style attributes will be in tomorrows release.

[jackie_w]

Is there a way to insert a new image? I can see how to delete or replace an existing image. Have I missed something?

[kovidgoyal]

File->New file

[jackie_w]

Ah, logical enough. I'd assumed that meant open a new epub!

[kovidgoyal]

I have changed the menu entry to read "New file (images/fonts/HTML/etc)"