MobileRead Forums

MobileRead Forums (https://www.mobileread.com/forums/index.php)
-   Sigil (https://www.mobileread.com/forums/forumdisplay.php?f=203)
-   -   New Sigil Release: Sigil-0.8.901 (https://www.mobileread.com/forums/showthread.php?t=266008)

KevinH 10-04-2015 05:13 PM

New Sigil Release: Sigil-0.8.901
 
Hi All,

The second (and hopefully final!) in our Sigil pre-release series, Sigil-0.8.901 is now available. This release is primarily a bug-fix release, although a couple of new features have slipped in. This release is much more stable than its predecessor all thanks to your testing and bug reports. If no major showstoppers are found, this pre-release will form the official Sigil 0.9.0 release and we will then start work towards Sigil 1.0 and reasonable epub3 support.

As far as new features go, the tool box for plugins that use the embedded Python 3.4 interpreter, has been expanded. The embedded site-packages now include: six, html5lib, lxml, sigil_bs4 (Sigil's version of Beautiful Soup4), Google's Gumbo parser via a sigil_bs4_to_gumbo_adapter, PIL (Pillow image library), regex, cssutils, cssselect, chardet, and our own ctypes interface to the hunspell spell checker. This along with the new epub3 versioned plugin interfaces (see https://www.mobileread.com/forums/sho...8&postcount=95) should provide a quite powerful toolbox for plugin developers.


Complete details of what is fixed and how to get is is available from our wordpress site:

http://sigil-ebook.com

To download binaries for Mac OS X and Windows, please see our github releases page:

https://github.com/Sigil-Ebook/Sigil...es/tag/0.8.901

Please note, beginning with the Sigil 0.8.900 release, the process to build from source has changed greatly. Please check new instructions for building on Mac OS X, and Linux available on our github site. See the README.md and docs/ directory for complete instructions.

As always, only download Sigil from its official home github site and remember to run a sha 256 checksum on your downloads and compare them against those posted in the release on github to make sure you have downloaded the exact same binary that we have uploaded.

When the the next full release build is made, I will be signing the Mac OS X builds with my Mac OS X Developer account (assuming I can figure out how to do that!). Leave it up to Apple to charge volunteer developers just to build and release GPL OpenSource projects! They are indeed clueless.

Again, for fastest resolution, please post any bug reports or feedback on this release in this thread in the Sigil forum and we will try our best to address any serious issues in quick follow-up release if needed.

Thank you for using Sigil and helping to make it better!

KevinH and DiapDealer

rubeus 10-04-2015 05:51 PM

:thumbsup: :thanks:

Arios 10-04-2015 07:43 PM

Wonderful!

All seem to work perfectly and faster than before: many thanks.

theducks 10-05-2015 02:37 PM

W10 X64

:eek: Using a Sigil split marker (loaded by S&R) files, results in a scrambled (splits at the end) book IIRC this was a OLD issue that was fixed

S&R status (how many, not found) missing. The lS&R box wobbles like it was trying ;), but nada showed

:blink: Crash
unfortunately, I can't remember the precise edit function I was performing, only that I had been doing a lot of successful (scan) cleanup before that happened

DiapDealer 10-05-2015 02:59 PM

I have no earthly idea what you mean by a "wobbling" S&R box! ;)

dynabook 10-05-2015 03:58 PM

I have an MR download "Bleak House" from HarryT that uses divs instead of p for paragraph classes. Any span within the div gets a lot of whitespace put around it by Prettyprint. This adds spaces into the text. Look for italic spans to see the effect.
--MH

theducks 10-05-2015 04:13 PM

Quote:

Originally Posted by DiapDealer (Post 3182813)
I have no earthly idea what you mean by a "wobbling" S&R box! ;)

The box changes size and right back. The box jiggles in height just a bit

DiapDealer 10-05-2015 04:29 PM

Quote:

Originally Posted by dynabook (Post 3182854)
I have an MR download "Bleak House" from HarryT that uses divs instead of p for paragraph classes. Any span within the div gets a lot of whitespace put around it by Prettyprint. This adds spaces into the text. Look for italic spans to see the effect.
--MH

Downloadable from here. Especially easy to see what you're referring to in the "Bleak_House_split_007.html" file (cast of characters).

Regardless of the crazy whitespace added in code view, Pretty-Print results in extra space being rendered after "Matthew Bagnet:" In the third entry ("Matthew Bagnet:*The owner of a music shop;"). That extra spacing isn't there when pretty print is turned off.

DiapDealer 10-05-2015 04:39 PM

Quote:

Originally Posted by theducks (Post 3182867)
The box changes size and right back. The box jiggles in height just a bit

I think we're going to need some specific examples in order to get a good handle on what you're describing concerning splitting and S&R.

-Some attached html
-The exact search/replace expression you're using to insert split markers
-what you're seeing after you split, and what you think you should be seeing after you split.

Keep in mind that you may not be able to use saved expressions/clips that are based on the way the old Sigil pretty-printed code (without tweaking them first). The Gumbo Pretty-print does not produce the exact same thing that the old Sigil Pretty-Print did.

I'm still not quite certain what you're describing with regard to the S&R functionality.
Quote:

S&R status (how many, not found) missing.

gipsy 10-05-2015 05:16 PM

Quote:

Originally Posted by DiapDealer (Post 3182882)
I'm still not quite certain what you're describing with regard to the S&R functionality.

For example... When you search for something and it returns 0 results... It doesn't return a "Nothing found" in the Search-Replace box.

I don't remember when. But I had the same and with Matches. It displayed nothing and not "Matches found:xx"
I think when i haven't any of the Option cheked.

KevinH 10-05-2015 05:50 PM

Hi,

I really and truly need an example ebook and the exact sequence of steps to have any idea what you are talkign about here? I really haven't a clue. So please attach a small epub example and please include a detailed sequence of steps to recreate the issue.

None of the4 S&R code was touched in the large change since Sigil 0.8.7, so it should not be broken in any way but perhaps the problem is caused by something outside of the S&R code? I really need an example and specific steps to track it down.

Thanks,

KevinH

theducks 10-05-2015 06:29 PM

The Search and Replace part works fine.
in Sigil.8.x there is a status line as Gipsy indicated. That is missing.

When it SHOULD appear, the S&R box does some sort of resize/refresh jiggle.


Splitting (the same way I do with 0.8.7)

I insert the standard: <hr class="sigil_split_marker" />
Manually, or by S&R code into the various HTML files

later, I press <F6> to split the whole lot


BTW When this happened before (Sigil2?) , valloric could not find it because he only had 1 file being split. I was splitting 4+ (each 240K limit break point of older devices ). All the splits ended up AFTER the last original pre-split file

DiapDealer 10-05-2015 11:31 PM

Quote:

Originally Posted by theducks (Post 3182929)
The Search and Replace part works fine.
in Sigil.8.x there is a status line as Gipsy indicated. That is missing.

When it SHOULD appear, the S&R box does some sort of resize/refresh jiggle.

I see what you're talking about here. Some statuses are showing (usually the searches/replaces/counts with results), some are not. Looking at the code... I have no idea why yet. The status updates ARE being sent. Just not shown.

This source file hasn't been touched since before the 0.8.7 release.

theducks 10-05-2015 11:42 PM

Quote:

Originally Posted by DiapDealer (Post 3183033)
I see what you're talking about here. Some statuses are showing (usually the searches/replaces/counts with results), some are not. Looking at the code... I have no idea why yet. The status updates ARE being sent. Just not shown.

This source file hasn't been touched since before the 0.8.7 release.

a latest QT snafu?

Turtle91 10-06-2015 12:07 AM

Quote:

Originally Posted by theducks (Post 3183037)
a latest QT snafu?

I noticed something today about "jiggling boxes" that may help track it down... hopefully it's not a red herring.

Sigil 8.7:
I extended my desktop so that one monitor (primary) had the preview pane all to itself, the other monitor had Sigil. When I tried to change a S&R option ("This Page" to "All HTML pages") the box just jiggled and did NOT drop down to show the list of options. When I returned the preview pane to it's rightful place (docked on the right side of the window) and went back to displaying only on the secondary monitor...then the problem went away.

So, this very well may be a QT5 issue that doesn't have anything to do with 8.901

Cheers,

KevinH 10-06-2015 12:29 AM

Hi,

I think this problem is my fault. The damn tr() function to actually translate text is actually what is returning blanks instead of the message it is supposed to show.

I think this is because somehow I have generated a base.ts that seems to be seems to be missing all file info and line numbers. I compared it to the base.ts that was in the Sigil 0.8.X branch and it properly had line numbers and did not display this empty message nonsense.

So I have taken a chance and rebuild cleanly base.ts so it now has all of the file info and line numbers in it and will upload it to transifex for translating. Hopefully I have not lost anything by doing that but that was the only way to fix the issue and get the messages back for me.

I will check out things tomorrow to make sure I have not messed up any other language translations doing it. I have my fingers crossed.

But my what a strange problem!

KevinH

theducks 10-06-2015 01:06 AM

Quote:

Originally Posted by Turtle91 (Post 3183044)
I noticed something today about "jiggling boxes" that may help track it down... hopefully it's not a red herring.

Sigil 8.7:
I extended my desktop so that one monitor (primary) had the preview pane all to itself, the other monitor had Sigil. When I tried to change a S&R option ("This Page" to "All HTML pages") the box just jiggled and did NOT drop down to show the list of options. When I returned the preview pane to it's rightful place (docked on the right side of the window) and went back to displaying only on the secondary monitor...then the problem went away.

So, this very well may be a QT5 issue that doesn't have anything to do with 8.901

Cheers,

Preview is on my Upper Monitor, Sigil Main is on the Lower :thumbsup:

theducks 10-06-2015 01:08 AM

Quote:

Originally Posted by KevinH (Post 3183056)
Hi,

I think this problem is my fault. The damn tr() function to actually translate text is actually what is returning blanks instead of the message it is supposed to show.

I think this is because somehow I have generated a base.ts that seems to be seems to be missing all file info and line numbers. I compared it to the base.ts that was in the Sigil 0.8.X branch and it properly had line numbers and did not display this empty message nonsense.

So I have taken a chance and rebuild cleanly base.ts so it now has all of the file info and line numbers in it and will upload it to transifex for translating. Hopefully I have not lost anything by doing that but that was the only way to fix the issue and get the messages back for me.

I will check out things tomorrow to make sure I have not messed up any other language translations doing it. I have my fingers crossed.

But my what a strange problem!

KevinH

Hey! If it was easy, we would all be writing code :o

gipsy 10-06-2015 03:55 AM

Select some text in code view - click to Superscript. Sigil crush.
I try to duplicate it without result.

KevinH 10-06-2015 08:31 AM

Is this with Windows? If so, it may be clipboard related. I have not been able to duplicate this subscript issue on Linux or Mac. If you run into it again please let us know.

KevinH

KevinH 10-06-2015 10:56 AM

Hi All,

Given the very few bug reports, I think we have achieved some sort of stability here. We will wait a few weeks in case anything else shows up and then declare the pre-release series done and release a very slightly updated version (hopefully with working messages in Find/Replace), plus a number of other minor changes) as the official Sigil 0.9.0.

Thanks all for your continued use and bug reports.

KevinH

gipsy 10-06-2015 11:27 AM

Yes Kevin, in Windows 10

gipsy 10-06-2015 05:42 PM

In the previously releases when you save and you have something like this...
Code:

      <p class="subtitle">      ΠΕΡΙΘΩΡΙΑΚΑ
</p>

it changes to
Code:

      <p class="subtitle">ΠΕΡΙΘΩΡΙΑΚΑ</p>
Now it doesn't work, I tried it with the two Cleaning types (at Windows machine).

EDIT: Another example
Code:

        <div class="footnote">
          <p>[1]      Some text.
 [<a href="#id001" id="ftn.id001">↑</a>]</p>
        </div>


KevinH 10-06-2015 06:10 PM

Hi,
The change you want makes no sense as it removes whitespace that may in fact be part of the paragraph both before and after the exact text (up to one character of each.

Quote:

Code:

      <p class="subtitle">      ΠΕΡΙΘΩΡΙΑΚΑ
</p>

it changes to
Code:

      <p class="subtitle">ΠΕΡΙΘΩΡΙΑΚΑ</p>

Now it doesn't work, I tried it with the two Cleaning types (at Windows machine).

Yes, pretty print has changed a lot, It will no longer trim whitespace from inside of paragraph tags as it should not since whitespace can be important both before and after any text (at least up to one character's worth).

Quote:

EDIT: Another example
Code:

        <div class="footnote">
          <p>[1]      Some text.
 [<a href="#id001" id="ftn.id001">↑</a>]</p>
        </div>


And how would you expect this to turn out? Is there at least one char of white space after the footnote marker and before the "Some text." and after the text?

I am not sure either of these are bugs. You seem to be asking for the old Tidy approach of a deep clean. I am sorry but those days are pretty much gone. If you want me to modify prettyprint to reduce multiple spaces into one space inside of the p tag, that may be doable, but stripping it all away is simply incorrect.

KevinH

gipsy 10-06-2015 06:21 PM

The old days the last example turn to...

Code:

        <div class="footnote">
          <p>[1](space)Some text.[<a href="#id001" id="ftn.id001">↑</a>]</p>
        </div>

It's not a problem if they stay like this. The reader displays them right.
Thanks KevinH

KevinH 10-06-2015 06:31 PM

Hi,

So do you think prettyprint should ignore the newline and completely remove it as opposed to replacing it with a single space? See the extra space below:


Code:

        <div class="footnote">
          <p>[1](space)Some text.(space)[<a href="#id001" id="ftn.id001">↑</a>]</p>
        </div>

Otherwise, you can see things like:
Code:

<p>
This
is
a
line
of
text.
</p>

become:

Code:

<p>Thisisalineoftext.</p>
instead of:

[code]
<p>This is a line of text.</p>
[/code.

So dropping newlines and not replacing them with spaces can lead to nonsense just as easily as it can lead to good things. That is why I try to have prettyprint leave whitespace inside of tags alone as much as possible.

KevinH

gipsy 10-06-2015 06:57 PM

My mistake. The pretty in 0.8.7 change it like your code, with a space after "Some text".
And I don't want to completly remove the newline :)

But the
Code:

<p>
This
is
a
line
of
text.
</p>

in 0.8.901 stays as is that is a little confusing in code view, instead of 0.8.7 that changes to
Code:

<p>This is a line of text.</p>
edit: As for
Code:

<p>This<br/>is<br/>a<br/>line<br/>of<br/>text.</p>
if you save some poetry, and you want to see it in code view. In 0.8.901 with pretty and google, stays as is, without line change after the break tag

KevinH 10-06-2015 07:18 PM

Hi,
Okay, I can live with changing newlines to spaces and replacing a sequence of multiple spaces with a single space when text is inside a tag like p. I will try to play around with gumbo prettyprint to do that hopefully without breaking anything else.

KevinH

gipsy 10-06-2015 07:25 PM

Thanks KevinH,
leave it as is now and focus to other things for the next release :p
As I say and before, it's not that important. :thumbsup:

pete6055 10-06-2015 07:37 PM

Hello:

Very nice release guys!

I do, however, have a question:
Will the next non-beta release rewrite the tags:
Quote:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
"http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title></title>
</head>
at the top of html documents? Sigil 8.6 did so, depending on how clean source was set, but the new gumbo-parcer and gumbo-pretty print seem to handle this differently.

I sometimes start with html from the web, and the initial declaration is sometimes different, so short of copying the snippet to before the <body> tag, is there a more efficient way (perhaps a plugin)?

Thanks….

KevinH 10-06-2015 08:22 PM

Hi,

As far as I can tell, gumbo will happily parse all types of doctypes and return then to their original condition including html5 doctypes !DOCTYPE html.

If you find a case where it does not do that, please report it and I will try to track down where things are getting messed up.

Take care,

Kevin

Quote:

Originally Posted by pete6055 (Post 3183538)
Hello:

Very nice release guys!

I do, however, have a question:
Will the next non-beta release rewrite the tags:

at the top of html documents? Sigil 8.6 did so, depending on how clean source was set, but the new gumbo-parcer and gumbo-pretty print seem to handle this differently.

I sometimes start with html from the web, and the initial declaration is sometimes different, so short of copying the snippet to before the <body> tag, is there a more efficient way (perhaps a plugin)?

Thanks….


eggheadbooks1 10-06-2015 11:22 PM

I installed this latest version today and the ePub I made it in will not display properly in any of my ePub testing apps (ADE, Aldiko), nor my Kobo; it's as if the CSS cannot be read. Yet it displays properly in Sigil, and I can convert the file to Kindle using Kindlegen and the ebook displays as coded.

All my ePubs built using earlier versions display properly in my ePub apps. I have compared the header code and Content.opf code between the old ebooks and this new one and cannot find anything amiss.

Any ideas?

eggheadbooks1 10-06-2015 11:37 PM

Quote:

Originally Posted by eggheadbooks1 (Post 3183636)
I installed this latest version today and the ePub I made it in will not display properly in any of my ePub testing apps (ADE, Aldiko), nor my Kobo; it's as if the CSS cannot be read. Yet it displays properly in Sigil, and I can convert the file to Kindle using Kindlegen and the ebook displays as coded.

Well isn't this the way ... no sooner do I post this message than I found the problem. Odd that Kindlegen converted properly...

theducks 10-06-2015 11:38 PM

Quote:

Originally Posted by eggheadbooks1 (Post 3183636)
I installed this latest version today and the ePub I made it in will not display properly in any of my ePub testing apps (ADE, Aldiko), nor my Kobo; it's as if the CSS cannot be read. Yet it displays properly in Sigil, and I can convert the file to Kindle using Kindlegen and the ebook displays as coded.

All my ePubs built using earlier versions display properly in my ePub apps. I have compared the header code and Content.opf code between the old ebooks and this new one and cannot find anything amiss.

Any ideas?

Check the CSS for Errors (ADE will barf on the tiniest malformed rule that others deal with)
Right-click on the CSS in the Book browser: Validate with W3C (the red ones)

DiapDealer 10-07-2015 12:00 AM

Yep. Looks fine using WebKit, but not with ADE/RMSDK = CSS syntax error. Nine times out of ten. The tenth time = an embedded font that's incompatible with ADE/RMSDK. ;)

eggheadbooks1 10-07-2015 01:08 AM

Quote:

Originally Posted by theducks (Post 3183648)
Check the CSS for Errors (ADE will barf on the tiniest malformed rule that others deal with)
Right-click on the CSS in the Book browser: Validate with W3C (the red ones)

I did find it earlier, but the CSS was indeed the problem. A missing closing bracket. (I had added a new style by copying and pasting and missed the closing bracket.) I was too busy checking all the other code, trying to find the problem, that I missed the typo in the CSS.

Thanks for the tip to validate with W3C. Will try that next time I run into a problem.

eschwartz 10-07-2015 01:55 AM

Quote:

Originally Posted by KevinH (Post 3183494)
Hi,
The change you want makes no sense as it removes whitespace that may in fact be part of the paragraph both before and after the exact text (up to one character of each.



Now it doesn't work, I tried it with the two Cleaning types (at Windows machine).

Yes, pretty print has changed a lot, It will no longer trim whitespace from inside of paragraph tags as it should not since whitespace can be important both before and after any text (at least up to one character's worth).



And how would you expect this to turn out? Is there at least one char of white space after the footnote marker and before the "Some text." and after the text?

I am not sure either of these are bugs. You seem to be asking for the old Tidy approach of a deep clean. I am sorry but those days are pretty much gone. If you want me to modify prettyprint to reduce multiple spaces into one space inside of the p tag, that may be doable, but stripping it all away is simply incorrect.

KevinH

Also spaces from the beginning or end of the paragraph. ;)

FWIW, calibre preserves whitespace entirely, on the rationale that when using <pre>, whitespace matters.
IIRC, only whitespace outside a block-level tag gets reformatted when pretty-printing.


Sigil may or may not wish to follow the same approach. :)

...

Seems like an unwise (and unlikely) thing to use, IMHO, but I'd hate to be responsible for Sigil breaking a book because of this. :D

KevinH 10-07-2015 12:22 PM

Hi eschwartz,

Yes, I agree. That is why we moved away from Tidy (it cleaned too much and did not fully support all html5 tags, svg tags, and math tags). Regular gumbo serialization, changes no whitespace whatsoever. gumbo prettyprinting will now condense multiple whitespace on any allowed tag (and that is quite a specific list to prevent problems) and will trim leading and trailing whitespace from inside p tags only. All other places, people will just have to live with it or edit it by hand.

I am thinking of creating a "Clean-and-Sanitize" plugin using the html5lib's parser and sanitize code and a few things like that just to give people the option of "heavy cleaning" if they end up starting with heaping piles of crap html code (ie. read that they have unpacked an entire mobi7 book to one huge html file with lots of out of date html 3.2 pieces flying around.

Thanks!

KevinH

DiapDealer 10-07-2015 12:48 PM

I'm think that in the future (not immediate mind you ;) ), we may want to add a new pretty-print plugin type. One that can be used in the"traditional" PluginRunner dialog setting (for debugging purposes), but can also be "assigned" via the clean source settings using a combobox populated with installed pretty-print plugins.

I've been mulling the idea around in my head for a bit. And it'll probably mull a little longer. The bottom line is: Sigil's not overly concerned with making your code look "pretty" any more. It's more concerned with not breaking anything. We are, however, dedicated to providing a framework that offers users a way to manipulate their code in any way they see fit--through a robust plugin system.

My two cents anyway.

Leonatus 10-07-2015 03:37 PM

Windows 10, 32-bit.
Sigil crashes repeatedly
a) at uploading a cover image (in "Tools"->add cover, not at uploading other Images in the respective Folder of the book browser)),
b) when trying to create TOC via Tools->TOC->create Toc.

No Problem when I add a cover/TOC in a different program and open the epub afterwards in Sigil.

Am I the only concerned Person?


All times are GMT -4. The time now is 08:14 PM.

Powered by: vBulletin
Copyright ©2000 - 3.8.5, Jelsoft Enterprises Ltd.
MobileRead.com is a privately owned, operated and funded community.