I think a file got corrupted using Sigil

[Gregg Bell]

I was working on an epub in Sigil. It's a novel and I was breaking it into chapters with the Split At Cursor button. Specifically I would do the first chapter in Book, then hit the Find feature (where I had "chapter" input). That would kick the Book into Code and I would then use the Find again to get to the next chapter.

Then I would use the Split at Cursor button to split off another chapter.

It worked perfectly for ten chapters, then suddenly bogged. It still worked but everything took like 30-45 seconds.

I shut Sigil down. I re-started the computer.

When the computer came back up and I brought up Sigil and opened the file it was even worse. First of all it took 45 seconds to open the file. Then whatever I did (no matter how slight) bogged it. I started watching the Task Manager and yeah it was Sigil tying up 100% of the CPU.

This window came up:

The application Keyboard Input Methods has closed unexpectedly.

I'm on Xubuntu 15.04 Vivid.

I copied the "details" and attached them.

I re-launched the Keyboard Input Methods. (I thought that might somehow be the problem because earlier I was using the F5 key for the 'next spelling error' check.)

When I tried the file again in Sigil though it was the same result--totally bogged.

I made a copy and tried it in Sigil--same thing.

Anyway, then I started thinking the file is corrupt.

I opened a different epub in Sigil and it worked perfectly.

IF it is a corrupt file I only lost a day's work so it's no big deal. BUT I wonder if I did something to corrupt the file (it started as an odt, I turned it into an html in LibreOffice and input the html into Sigil). Maybe by switching back and forth between Book and Code via the Find search. AND I'm wondering if yesterday's file (I checked and it worked fine) is going to suffer the same fate (corruption--again if it is corruption).

And is there anything I can do to somehow deal with this thing more effectively than just keeping my fingers crossed and starting over?

Sorry for the novel (it's just less than the novel I was working on!). Thanks.

[theducks]

Greg
I insert a sigil_split Marker at the places I want (ctrl-shift-enter (I made a clip)
(Using REGEX when possible)

then press F6 at the end

I only use the Single, split here (icon) for odd cases or when the conversion code is a horrible mashup

[alexa04]

Mine got corrupted too, thanks to this post I'd fixed mine.

[KevinH]

Hi,
The Keyboard Input Methods are system level not Sigil level. No where in your graphic was anything related to Sigil that I can see. It appears to be a linux system level failure.

That said, try renaming the broken .epub file to a .zip file and then unzip it. Then open any of the .xhtml files in the OEBPS/Text folder with any other text editor and look to see what was written to your files.

In addition, it would not hurt to let us know what version of Sigil you are using, who built it, etc. It also would not hurt to run a disk diagnostic on your hard drives to rule out disk issues at the linux system level.

Please let us know what you find out.

KevinH

ps, when working on any large file, it never hurts to use a File->Save-As to a new name (in my case I increment a revision number as part of the name). Then if your system has issues or Sigil has problems, or your power blinks, or ... you will never lose much if anything. Sigil only ever writes to temp files until you actually tell it to save.

Originally Posted by Gregg Bell

I was working on an epub in Sigil. It's a novel and I was breaking it into chapters with the Split At Cursor button. Specifically I would do the first chapter in Book, then hit the Find feature (where I had "chapter" input). That would kick the Book into Code and I would then use the Find again to get to the next chapter.

Then I would use the Split at Cursor button to split off another chapter.

It worked perfectly for ten chapters, then suddenly bogged. It still worked but everything took like 30-45 seconds.

I shut Sigil down. I re-started the computer.

When the computer came back up and I brought up Sigil and opened the file it was even worse. First of all it took 45 seconds to open the file. Then whatever I did (no matter how slight) bogged it. I started watching the Task Manager and yeah it was Sigil tying up 100% of the CPU.

This window came up:


I'm on Xubuntu 15.04 Vivid.

I copied the "details" and attached them.

I re-launched the Keyboard Input Methods. (I thought that might somehow be the problem because earlier I was using the F5 key for the 'next spelling error' check.)

When I tried the file again in Sigil though it was the same result--totally bogged.

I made a copy and tried it in Sigil--same thing.

Anyway, then I started thinking the file is corrupt.

I opened a different epub in Sigil and it worked perfectly.

IF it is a corrupt file I only lost a day's work so it's no big deal. BUT I wonder if I did something to corrupt the file (it started as an odt, I turned it into an html in LibreOffice and input the html into Sigil). Maybe by switching back and forth between Book and Code via the Find search. AND I'm wondering if yesterday's file (I checked and it worked fine) is going to suffer the same fate (corruption--again if it is corruption).

And is there anything I can do to somehow deal with this thing more effectively than just keeping my fingers crossed and starting over?

Sorry for the novel (it's just less than the novel I was working on!). Thanks.

[Gregg Bell]

Originally Posted by theducks

Greg
I insert a sigil_split Marker at the places I want (ctrl-shift-enter (I made a clip)
(Using REGEX when possible)

then press F6 at the end

I only use the Single, split here (icon) for odd cases or when the conversion code is a horrible mashup

Thanks ducks. Your way worked great. Now when I use your method is that going to prevent my file from being corrupted? And the follow-up question to that is, the way I had split the chapters, did I cause the file to become corrupted?

And an additional note. I have a work and a home computer. They both have Sigil 0.8.4 and Xubuntu 15.04. The corrupted file happened on the work computer, which is not as powerful as the home computer. (Sometimes the graphics is distorted a little, like in drop-down menus and sometimes on the screen in general, on the work computer.) see screenshot

I noticed that I was able to use your method of splitting the chapters on my home computer (the more powerful one) in Book, but when I tried it on my work computer, when I did the first split it did if fine but then when I hit "chapter" in the "Find/Replace" box it took me to the Code version. Whereas on my home computer when I did that it remained in Book.

I mean, it was okay. I just did it your way in Code and it was fine. I was just wondering why the same version of Sigil would work differently on different computers.

[Gregg Bell]

Originally Posted by alexa04

Mine got corrupted too, thanks to this post I'd fixed mine.

How did you fix yours?

I could see avoiding corrupting a new file, but from theducks post how did you fix it?

[Gregg Bell]

Originally Posted by KevinH

Hi,
The Keyboard Input Methods are system level not Sigil level. No where in your graphic was anything related to Sigil that I can see. It appears to be a linux system level failure.

That said, try renaming the broken .epub file to a .zip file and then unzip it. Then open any of the .xhtml files in the OEBPS/Text folder with any other text editor and look to see what was written to your files.

In addition, it would not hurt to let us know what version of Sigil you are using, who built it, etc. It also would not hurt to run a disk diagnostic on your hard drives to rule out disk issues at the linux system level.

Please let us know what you find out.

KevinH

ps, when working on any large file, it never hurts to use a File->Save-As to a new name (in my case I increment a revision number as part of the name). Then if your system has issues or Sigil has problems, or your power blinks, or ... you will never lose much if anything. Sigil only ever writes to temp files until you actually tell it to save.

Thanks Kevin.

Yes, I do that incremental version saving, but only day-by-day.

Anyway, I did the renaming to a zip and opened it up and the last file I split got gigantic (see screenshot 73). I opened that gigantic section 10 html file in Kate and there were tons of non-breaking spaces. It started here (075) and ended here (076).

I use this as the break point between chapters:

<p class="asterisks">*&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;* &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;*</p>

I don't know. There is a beginning and ending P tag in Kate so that seemed normal, but somehow the non-breaking spaces multiplied exponentially. It seems to me (and I am a non-techie) that all those non-breaking spaces overwhelmed the file and probably broke/corrupted it.

I checked one of the other html files in a text editor and it looked normal.

The Sigil version on the computer I used is 0.8.4 by John Schember and you(?) and some others.

I'm using an old Dell (Optiplex 170L) and running Xubuntu 15.04. If you can check the second to last post (to theducks) there's a screenshot of some distortion I've been having since I upgraded the computer to 15.04. But the computer has been (besides that inconvenience) functioning excellently.

And it looks like the hard drive is okay. (77)

the ducks suggested a way of spitting the chapters and it worked well.

Anyway, that's about all I can think of.

Thank you.

[theducks]

I am NOT using 8.901 (i REVERTED TO 8.7)
That scrambled the split order for me on w10

I used to have this problem back in Valoric's reign. It was only happening (reported) to me which was why it took so long

If I started with a book from Calibre that was split only because a section exceeded 260K, it happened when I pressed F6

file1
file2
file3

resulted in
file1
file2
file3
file1_split 001
file1_split 002
file1_split 003
file1_split 004
file1_split 005
file2_split 001
file2_split 002
file2_split 003
file3_split 001
file3_split 002
file3_split 003
file3_split 004
file3_split 005

Always Use a file (re)naming convention that allows for a (right click) alphanumeric sort.
Random applied splits will result in the next available Section### being assigned, which lessens your chance of easy recovery

[KevinH]

Are you in Bookview or Codeview when you initiate the split and where exactly is your split point (cursor)? I still have not been able to recreate it. Please provide a test case and specific sequence of events so I can track this down and get it fixed.

[KevinH]

Where exactly was your cursor when you started the last split? Was it by any chance inside one of the nbsp entities?

[Gregg Bell]

Originally Posted by theducks

I am NOT using 8.901 (i REVERTED TO 8.7)
That scrambled the split order for me on w10

I used to have this problem back in Valoric's reign. It was only happening (reported) to me which was why it took so long

If I started with a book from Calibre that was split only because a section exceeded 260K, it happened when I pressed F6

file1
file2
file3

resulted in
file1
file2
file3
file1_split 001
file1_split 002
file1_split 003
file1_split 004
file1_split 005
file2_split 001
file2_split 002
file2_split 003
file3_split 001
file3_split 002
file3_split 003
file3_split 004
file3_split 005

Always Use a file (re)naming convention that allows for a (right click) alphanumeric sort.
Random applied splits will result in the next available Section### being assigned, which lessens your chance of easy recovery

Thanks ducks but sorry this is pretty over my head.

[Gregg Bell]

Originally Posted by KevinH

Are you in Bookview or Codeview when you initiate the split and where exactly is your split point (cursor)? I still have not been able to recreate it. Please provide a test case and specific sequence of events so I can track this down and get it fixed.

I am in Bookview when I initiate the split. The cursor is to the left of Chapter One.

(cursor)Chapter One

What happens is I do the split. (It is successful.) Then I have "chapter" in the "Find" tool. When I click on Find, Sigil automatically switches to Codeview and takes me to "Chapter One." I then click Find again and it of course takes me to Chapter Two. Then I switch back to Bookview. Make sure the cursor is to the left of "Chapter Two" and repeat the process.

Every split worked properly until I got to xhtml10. Then it bogged.

Regarding your question in the other post about non-breaking spaces. The non-breaking spaces I added, those were all serving as scene breaks so "Chapter" would not be there at all. So it was impossible that a non-breaking space would get in the way.

What I thought was unusual was that on my home computer I could stay on Bookview while using the Find tool (in the process described above), whereas on my work computer (the one that I had the problem on) when I used the Find tool (in Bookview looking for "chapter") it kicked me into Codeview.

I need to work on this book, so what I was thinking was prudent was to do the splits in Codeview. (Although the User Manual suggests doing it in Bookview.)

Any suggestions for how to do the chapter splits without breaking another file? (I don't know if you saw theducks suggestion of doing shift-control-enter for the individual splits and then F6 to do the collectively.)

Thanks.

[KevinH]

Originally Posted by Gregg Bell

And it looks like the hard drive is okay. (77)
Thank you.

For the record ... , that entry, I think, is about the temperature of the hard drive. To truly test your hard drive you should run "fsck" with the proper command line options to see if anything needs to be fixed. I am sure Linux probably has a gui wrapper around that tool that will make it easier to use. It is worth checking out if you have the time. The testing can take a good bit of time depending on how big your drive is and how full it is. Testing the drive is often done periodically after some number "n" of mounts.

KevinH