File Splitting-->Rename Files-->Link filenames broken

[GrannyGrump]

This might already be discussed somewhere, but I didn't find it.

My tale:
I created my html text in Notepad++, one big file for the entire book, with Sigil chapter-break split markers included.
All my footnote tags and ID's are included, and look like this. They all use ch01.xhtml as the filename, because the file has not been split yet.

<a href="../Text/ch01.xhtml#fnt01" id="fnc01">[1]

The last footnotes occur in what will be Chapter 70.


Load this into Sigil v 7.4 or 7.3 (or copy-paste). (I am using the portable version from PortableApps.com, in case this has any bearing on the subject.)
Go to Edit->Split On Markers. Successfully split into 73 files.

Sigil has set EVERY footnote link and back link to this:

<a href="../Text/ch01_0069.xhtml#fnt01" id="fnc01">[1]

, with the exception of the links in ch01_0069.xhtml, which it (incorrectly) left set at ch01.xhtml.

I must manually edit each one to match the correct chapter name/number.

If I use bulk-rename to ch01; ch02, ch03, etc; it then sets the filenames in all the links to ch70.

So of course, every footnote call goes to the footnotes in Chapter 69/70. Every backlink returns to the anchor ID in Chapter 69/70. (Excepting the ones in Chapter 69/70, which travel to and from Chapter 01.)


The same thing happens if I edit the filenames in the links to the "future" filenames before I load the text into Sigil.

Is this a normal outcome, or have I done something wrong? For future use, do I need to split the file up and set the links BEFORE I load into Sigil?

EDITED TO ADD: I used to do this successfully in v 6.x. When creating the file in Notepad++, I would set the link filenames to the "future" filename, and Sigil would leave them alone. But now it is updating links when files are re-named, and something has gone wrong with my world...

[Toxaris]

Since it is one big file, you can omit the filename from the links. That might do it after splitting.

[theducks]

Originally Posted by Toxaris

Since it is one big file, you can omit the filename from the links. That might do it after splitting.

My experience is splitting a Relative link, always breaks things. Orphans: ..\Text\#foo

This (change) has to be a really odd side effect of the OP's work flow. (AFAIK this is the first report in 3 years)

[DiapDealer]

I can always manage to find ways to make Sigil trounce some links when doing a lot of splitting/merging/renaming. I've just never been able to come up with a step-by-step to reproduce it.

[eschwartz]

Try seeing how calibre handles it in a conversion?

[GrannyGrump]

Originally Posted by Toxaris

Since it is one big file, you can omit the filename from the links. That might do it after splitting.

Well, that suggestion got my hopes up, but it didn't work. All the tags got assigned to ch70 again.

eschwartz said

Try seeing how calibre handles it in a conversion?

I'm not sure what you mean. Could you give a brief explanation?

EDITED TO ADD: Just on the off chance it might make a difference, I tried again with my *non-portable* version 7.2 installation. The same thing happened.

I can't believe I'm the only victim! Has nobody had this happen with any of the 7.x versions?

[eschwartz]

If you turn it into a one-file epub, then do an epub-to-epub conversion, calibre will split at the chapter breaks, as found by xpath, using the header tags.

[GrannyGrump]

Ah, I see (understandable when I finally pay attention to what you say!)

Should I remove the Sigil split markers before I try the Calibre conversion?

[theducks]

Granny
All I do to programatically split a file is insert a Sigil Split Marker using REGEX

To split before a <h3 tag (add more terms to the REGEX if H3 is used elsewhere)

search : <h3
Replace: <hr class="sigil_split_marker" /> <h3

when complete (I do a save for good times), the press F6 to cause the split.

I have never had a mangle like you show. (XP, W7, or Linux)

[eschwartz]

Originally Posted by GrannyGrump

Ah, I see (understandable when I finally pay attention to what you say!)

Should I remove the Sigil split markers before I try the Calibre conversion?

I don't believe calibre knows/cares what to do with them. They won't damage anything, but you might want to get rid of them simply because they are extra. (Sigil would get rid of them by turning them into file splits, but I think calibre would simply ignore them.)

[theducks]

Sigil split markers are just a <hr /> with a unique class name that Sigil will see and split at (and remove)

[eschwartz]

Originally Posted by theducks

Sigil split markers are just a <hr /> with a unique class name that Sigil will see and split at (and remove)

So it would show up as a big line at the end of every page?

Well, then, definitely get rid of it.

I think I agree with theducks that the best way to make those markers is with a regex, that way you don't create them until/unless it's necessary.

[GrannyGrump]

@theducks -- I have tried this on my home computer (Asus laptop running Win8 [ewww]) and on a borrowed Dell running Vista (double ewww). Identical results.

@eschwartz-- Thanks for that info. I will try the Calibre conversion this evening at home, and hopefully it works.

If not, I will have to come up with a better work-flow. This occasion isn't too bad, because there are only about 45 links, so not too much to handle manually. But I need to find out why it happens, and stop it from happening with a book with lots more links.

[GrannyGrump]

Well, I tried a few rough tests converting epub --> epub with Calibre. First attempt was different, but just as mangled as the Sigil output. But now I see the problem.

In both Sigil and Calibre, I see that I cannot start renumbering footnotes with #01 on each chapter. I will have to start with 1 and number straight through to 45. Then, it keeps the anchors attached to the correct splits/chapters.

I think I was expecting a miracle from the software. (hey, a person can dream, can't they?)

Thanks to all for your help. eschwartz, your Calibre tip helped lift the curtain of confusion.

[theducks]

Originally Posted by GrannyGrump

Well, I tried a few rough tests converting epub --> epub with Calibre. First attempt was different, but just as mangled as the Sigil output. But now I see the problem.

In both Sigil and Calibre, I see that I cannot start renumbering footnotes with #01 on each chapter. I will have to start with 1 and number straight through to 45. Then, it keeps the anchors attached to the correct splits/chapters.

I think I was expecting a miracle from the software. (hey, a person can dream, can't they?)

Thanks to all for your help. eschwartz, your Calibre tip helped lift the curtain of confusion.

The link id needs to start with a letter or FC flags errors
id='42' error
id='f42' good

Some Crafty REGEX run on a per file from the Calibre editor can rename those 'duplicate' ids

Find: \.html\#(\d+) <change the file type to match, both places
Replace: .html:#a\1 <- Change to a unique letter(s) for each file processed

REGEX can't to math, so you might come up with an Idea to use in the actual FN label [C2-1) (chapter2: note 1)