Unicode error, q to Valloric

[Hitch]

@Valloric:

My Crew and I have stumbled across a wee error that seems to be a bit different than that which is reported here: http://code.google.com/p/sigil/issues/detail?id=316

in that if you manage to find the mysterious disppearing, Bill-Gates-is-trying-to-drive-me-insane Unicode for the soft hypen (more on that here: http://www.cs.tut.fi/~jkorpela/shy.html

And delete it, because the character is invisible to Sigil in the first place, Sigil does not recognize the document as changed, and therefore did not show the document-in-progress as having been changed (and needing to be saved). Saving the epub DID preserve the changes, however. There are bigger issues, though, so here's the boring history:

1. Created a mobi from a perfectly fine epub
2. Which, somehow, had made it to html stage some months back without one of us running our clip on it to remove MSFT's fracking invisible soft-hypens (in which they use the WRONG bloody unicode, anyway);
3. Client didn't proof it,
4. And it was uploaded to Amazon as-is,
5. Full of, SURPRISE!--soft-hyphens hither and yon.
6. We opened epub, and couldn't find that invisible code to save our lives, until
7. One of my Crew TL's figured out (genius) that if he copy-and-pasted one of the hyphenated words (that we could see in the mobis) from Sigil in CV into the S&R dialogue box, and then BACKSPACED (to delete the post-hyphen characters) to the spot where the hypen would be, it magically appeared, in the S&R dialog box, whereupon,
8. After deleting everything in FRONT of the offensive thing, we were able to s&r the entire epub, and delete all 972 occurrences.
9. Sigil did not recognize the document as being changed, after the 972 occurrences were deleted, so made no indication that the document needed saving. When we did save it, the occurrences were indeed deleted, which is happiness itself. BUT,
10. When I checked this in 0.4.0, not only was the character invisible in the xhtml, but,
11. In the s&r dialog box, it was displayed not as a -, (after the backspacing trick) but as a SPACE, and worse,
12. 0.4.0 "saw" it as a space, so you can't search and replace for it.

SO: Do you want me to expound upon this, or add this to the Issue 316 bug report, or make a new one, or...? And I'm happy to give you private copies of all the documents involved, but I never have been able to figure out how to upload docs at the GoogleCode thing and keep them private. I know you're seeking gainful employment world-wide (don't leave us! don't leave us!), but I'm happy to make as full a report as you need/want.

Hitch

[Valloric]

Create a new issue. Frankly, I'm not sure I understand what the problem is here (in fact, I'm sure I have no idea what you're talking about ), but do create a detailed report on the tracker, test procedure/test files and everything.

In the "Template" drop-down box in the New Issue form, there is a "Private issue" template. Use that to make the issue private.