Mobileread
Sigil-0.9.991 (1.0 Pre-release)
#1  DiapDealer 11-09-2019, 11:08 PM
Announcing Sigil-0.9.991 (pre Sigil-1.0) Alpha Release

----------------------------------------------------------------------------------

Before Sigil 1.0 can be released, we felt it was very important to remove the restriction that all epubs must be in Sigil standard form which included one OEBPS folder, specific names for the OPF, and NCX, only one folder for each type of resource (Text, Images, etc.).

Therefore, over the last two months, we have had to tear out much of the internals of Sigil to teach it to not standardize epubs on initial load, not to rename the OPF, NCX, not to move files to our specific folders, etc. Then we had to teach Sigil how to properly work with files in any structure or locations, how to build the correct links between them. This included extending the Sigil plugin interface appropriately.

To accomplish this over 4000 lines of code were added and over 2000 were deleted, touching almost every file in Sigil. All of this work has been very disruptive with the potential for lots of breakage and new bugs being introduced.

Special thanks must be given to both BeckyEbook and Doitsu for regularly building and testing the changes as they were being made and reporting back any breakage or issues they found and providing feedback on ideas. Without their hard volunteer effort, Sigil would be in a much much worse condition.

The work on Sigil-1.0 has now stabilized, and it is now ready for additional volunteers and testers to try and break it so that any final nits or issues can be tracked down and fixed before the upcoming Sigil 1.0 release.

The main changes include the following:
  1. Upon initial load, the epub structure that exists will be loaded and kept in its "as-is" state, moving and renaming nothing
  2. The BookBrowser will still show the Text, Images, Styles, Fonts, Audio, Video, and Misc "folders" but these are now "virtual folders" that simply organize the existing resources by type but do not restrict their actual location within the epub.
  3. A new Preferences setting has been added to the Appearance Preferences that will determine if the book folder relative path will be shown in the BookBrowser or its shorter unique name (filename) will be used instead. In either case, the BookBrowser tooltips will always show the file's true book folder relative path. Duplicate file names (in different subfolders) will always show the longer path to be able to differentiate between them regardless of the new Appearance Preference.
  4. A Tools menu item has been added to standardize any epub already loaded to Sigil's old structure
  5. A Tools menu item has been added to allow a user to design a custom "empty" epub layout
  6. A "Move" context menu item has been added to BookBrowser to enable moving of resources to where you want them to be.
  7. A "Rename" context menu item has been added to BookBrowser to enable renaming of the NCX and OPF resources.
  8. The python plugin interface has been extended to support all of this as documented in the new Sigil_Plugin_Framework_rev12.epub and online at MobileRead in our Plugin Development thread.

Additional bug fixes include:

Please note, the design goal of this project was to make minimal changes for users and ebook developers that still want to use Sigil as they did. The new Preference setting defaults to showing only the filename as before, and the default empty epub for creating new epubs is the same as before. The interface and functionality should stay almost exactly the same. Loading any epub previously saved by Sigil will still work exactly as before. Only editing new epubs that are in non-standard form will necessitate a use of the new Tool menu (see #4 above) to standardize it.

Call For Feedback, Testing, Plugin Developers, etc...

----------------------------------------------------------------------------------

So we really need your help to test Sigil to make sure we have not broken anything and to provide feedback on the new features.

This release will also serve to allow Plugin Developers to test their plugins and make any changes needed to make sure their plugins will still work as expected.

So please take a minute to download Sigil-0.9.991 and test it as you would use earlier versions of Sigil, reporting back on MobileRead or via our GitHub site any bugs you find with instructions on how to recreate them. This is especially true of some of Sigil's more advanced features such as HTML TOC, Index generation, Merging and Splitting of Files, Reports, movement of resources, and any other feature of Sigil you would normally use on a regular basis. Though we're not going to ignore any reported issues, we're asking that you try to limit your reports to broken functionality or incorrect behavior. Cosmetic issues and minor inconveniences (with easy workarounds) can be dealt with later.

Note: Sigil-0.9.991 should not be considered a "production-ready" release. It is being made available to make sure our upcoming Sigil 1.0 release is as bug free and stable as possible.

Sigil-0.9.991 (1.0 pre-release) page and downloads.
Reply 

#2  un_pogaz 11-10-2019, 03:32 AM
Yeah, well done on the job.

Now bug report: First ePub, first bug :
I have applied a manual restructuring "OEBPS/Text/Scheer,K.-H.&Darlton,Clark-[Perry Rhodan-001]Operation Astree (1961)_split_006.html" is named in the OPF "../Scheer%2CK.-H.%26amp%3BDarlton%2CClark-%5BPerry%20Rhodan-001%5DOperation%20Astree%281961%29_split_006.html" (old index before restructuring)
Consequently, it is deleted.
I suspect a special character because the files in [A-z] only are correctly restructured.

Also, I think that an option to automatically apply a structure (either Sigil's or the custom one) at the opening of any ePub, would be very useful (no ask user, old working mode).

Another good thing would be to be able to save several structures and be able to apply them easily (a "Structures" sub-menu).

PS: I don't know if you planned to add it to this version, but fortunately Sigil warns us when he encounters a none (or poorly) indexed files.

EDIT: the files with special character are not even renamed, under any circumstances!
EDIT2 : It's even more vicious, if I save the ePub even without renaming anything, the second opening will fail.
Reply 

#3  BetterRed 11-10-2019, 05:29 AM
Any known reason why I shouldn't try it in the Windows 10 Pro Sandbox ?

BR
Reply 

#4  Doitsu 11-10-2019, 05:32 AM
Quote un_pogaz
Now bug report: First ePub, first bug :
I have applied a manual restructuring "OEBPS/Text/Scheer,K.-H.&Darlton,Clark-[Perry Rhodan-001]Operation Astree (1961)_split_006.html" is named in the OPF "../Scheer%2CK.-H.%26amp%3BDarlton%2CClark-%5BPerry%20Rhodan-001%5DOperation%20Astree%281961%29_split_006.html" (old index before restructuring)
I created a non-standard epub with the exact same file name and wasn't able to reproduce your bug.

Did you run EPUBCheck before your test?

Please post the following marked as spoilers:

1. The contents of the original .opf file.
2. The contents of the original .ncx file.
3. The contents of the new .opf file.
4. The contents of the new .ncx file.
[epub] non_standard.epub (5.2 KB, 26 views)
Reply 

#5  DiapDealer 11-10-2019, 05:34 AM
My god. The ridiculousness of some filenames out there that need to be accommodated never ceases to amaze me. We'll look into it, of course, but... my god.

Also, I'm unclear what you mean by:

Quote
PS: I don't know if you planned to add it to this version, but fortunately Sigil warns us when he encounters a none (or poorly) indexed files.
Reply 

#6  Doitsu 11-10-2019, 05:59 AM
Quote DiapDealer
Also, I'm unclear what you mean by [...]
AFAIK, French speakers often refer to the TOC as the index.

I might have been able to narrow the issue down. If the original NCX TOC contains invalid hrefs, the standardized EPUB will also contain invalid hrefs.

To test this, download the test epub that I attached to my previous post and do the following:

1. Open the epub.
2. Manually change the following line in the ncx file:

from:

Code
<content src="txt/Scheer%2CK.-H.%26Darlton%2CClark-%5BPerry%20Rhodan-001%5DOperation%20Astree%20%281961%29_split_006.xhtml"/>
to:
Code
<content src="test/Scheer%2CK.-H.%26Darlton%2CClark-%5BPerry%20Rhodan-001%5DOperation%20Astree%20%281961%29_split_006.xhtml"/>
3. Select Restructure Epub to Sigil Norm.

The above entry will be changed to:

Code
<content src="../test/Scheer%2cK.-H.%26Darlton%2cClark-%5bPerry%20Rhodan-001%5dOperation%20Astree%20%281961%29_split_006.xhtml"/>
Ideally, Sigil should make sure that all NCX file hrefs are valid and refuse to run Restructure Epub to Sigil Norm, if invalid hrefs were found.

D.
Reply 

#7  DiapDealer 11-10-2019, 06:25 AM
Quote BetterRed
Any known reason why I shouldn't try it in the Windows 10 Pro Sandbox ?

BR
None that I know of. Though I've absolutely no experience with W10's sandbox, so there's that to consider.
Reply 

#8  Notjohn 11-10-2019, 07:38 AM
> French speakers often refer to the TOC as the index

Thank you, thank you! I am regularly puzzled by seeing that usage on the KDP forums, since an index, in American usage, always appears at the other end of the book (and has little practical benefit in an ebook).
Reply 

#9  najgori 11-10-2019, 07:43 AM
Quote DiapDealer
My god. The ridiculousness of some filenames out there that need to be accommodated never ceases to amaze me.
Pluto & Verso publishing in order to protect their books from piracy change all filenames inside of epub to something like chapter01JohnSmithjoh43324nsmithgmailcom.xhtml where John Smith is buyers name and after that is buyers email address. They also add exlibris with buyers name written in crappy font with limited characters support.

books are really good, though.
Reply 

#10  DiapDealer 11-10-2019, 08:59 AM
Quote najgori
Pluto & Verso publishing in order to protect their books from piracy change all filenames inside of epub to something like chapter01JohnSmithjoh43324nsmithgmailcom.xhtml where John Smith is buyers name and after that is buyers email address. They also add exlibris with buyers name written in crappy font with limited characters support.

books are really good, though.
That's just silly. If there's DRM, such filenames are pointless. And if there's no DRM (regardless of how/why) the filenames are a mass rename away from being pointless. Either way: their efforts/money to thwart piracy would be better spent elsewhere.

If their books are drm-free, there's no court in the land that would prosecute based on filenames found in the wild, anyway. "Your honor: I misplaced my reader during that time and found it a week later at work. I have no idea who might have used it (or what they might have done with it) during that time." *shrug*
Reply 

  Next »  Last »  (1/28)
Today's Posts | Search this Thread | Login | Register