Edit Book: Inconsistency in Differences between books

[chaot]

With F&R (Find and Replace) (in MR also called S&R (Search and Replace)) I want to reduce double spaces to one, with command Replace all. I repeated the procedure several times, waited, no changes!

Find: *SPACE**SPACE*
Replace: *SPACE*


1212 occurrences

Click: See what changed


Calculating differences, please wait This window does not appear!?


No changes found In fact, there was also nothing changed.


My e-reader displays the locally occurring spaces in the words, but not the double spaces between.

A similar (cosmetic) procedure I have already done some time before, successfully carried out in another book, there were more than 150000 occurrences.

Interna: Roche (F)

[theducks]

If in doubt, COPY (select, the ctrl-F) the search value (space pair), the underlying migt be another form of whitespace

REGEX \s\s might work, BUT it might also get line ends

[chaot]

REGEX \s\s might work, BUT it might also get line ends

I chose this approach.

Run Fix HTML - all files & Beautify all files

Find: \s\s
Replace: *SPACE*


Replaced 3494 occurrences


Differences between books (excerpt)


Result: Optical confusion

Run again Fix HTML - all files & Beautify all files


Seems OK

Checkbook was before, during and after the procedure OK.

Or do you see/know something that I overlooked? Line end problems!?

[jackie_w]

Originally Posted by chaot

Or do you see/know something that I overlooked?

Something to be careful of with your regex is that you might end up replacing consecutive non-breaking-spaces with a plain space. That might be important if nbsps have been used to create vertical whitespace scenebreaks. E.g.
Before: <p>*nbsp**nbsp*</p>
After: <p>*plainspace*</p>

The 'Before' paragraph will look like a blank line. The 'After' will be completely invisible.

[BetterRed]

I think the pink blobs in the third shot probably indicate the presence of a 'special' space character. In codeview window, position the cursor immediately after a blob and the character details will be displayed at the right end of the status line.

What does the fourth shot look like in Preview - it appears to have a space between almost every character - but if that's what you want I guess it's OK.

To address the issue jackie-w raises you could try replacing the nbsp at line beginnings with a marker, do the global replace, and then replace the markers with nbsp.

BR

[chaot]

Originally Posted by theducks

If in doubt, COPY (select, the ctrl-F) the search value (space pair)

Also possible. It works once, then again different, to say, once there are 1361 occurences, then 291 and so on!?

To explain here my attempts with Find: \s\s respectively *SPACE**SPACE* (your and my way ) and Replace all would be very awkward. But the result is quite clear: It works unfortunately only unstable.

Whether that applies only to this book mentioned here I do not know?

... the underlying mig[h]t be another form of whitespace

???

[chaot]

Originally Posted by jackie_w

Something to be careful of with your regex is that you might end up replacing consecutive non-breaking-spaces with a plain space. That might be important if nbsps have been used to create vertical whitespace scenebreaks. E.g.
Before: <p>*nbsp**nbsp*</p>
After: <p>*plainspace*</p>

The 'Before' paragraph will look like a blank line. The 'After' will be completely invisible.

Consecutive non-breaking-spaces do not have a syntax?

The meaning of *nbsp* (non blanc spaces) is not clear to me. A *nbsp* must be syntactically somehow recognizable and different from the *plainspace* - or?

Originally Posted by BetterRed

I think the pink blobs in the third shot probably indicate the presence of a 'special' space character. In codeview window, position the cursor immediately after a blob and the character details will be displayed at the right end of the status line.

Differences between books (excerpt) (Image with pink blobs)
Never mind which white or blanc spaces or plainspaces, the regular ones, the ones in some words or the double ones between the words: all called SPACE (in status line).

---

EDIT Soon my head is also an empty space - yet it hums.

There are NO-BREAK SPACES, colored yellow.


Code View (Yellow space)


NO-BREAK SPACE (Cursor behind yellow space)

---

What does the fourth shot look like in Preview - it appears to have a space between almost every character - but if that's what you want I guess it's OK.

No! What I want is: One (normal) white space between words, none in the words.


without correction of double white spaces


with correction of double white spaces Looks exactly the same.

To address the issue jackie-w raises you could try replacing the nbsp at line beginnings with a marker, do the global replace, and then replace the markers with nbsp.

??? Further explanations desired.

[jackie_w]

Originally Posted by chaot

The meaning of *nbsp* (non blanc spaces) is not clear to me. A *nbsp* must be syntactically somehow recognizable and different from the *plainspace* - or?

By *nbsp* I meant 'non-breaking-space', a.k.a. the unicode char \xa0 - also often found as the html entity &nbsp;. The calibre Editor Check function will complain if your HTML contains entities. The Beautify function will always replace all html entities with their unicode char equivalent. The Editor visually differentiates normal space and unicode \xa0 by highlighting the latter (as it also does for the mdash and ndash unicode chars to distinguish them from a normal hyphen).

A Regex search for \s will match both a normal space and an \xa0 (and other 'whitespace' chars for that matter, e.g. newline). If you only want to search for '2 or more consecutive normal spaces' then perhaps a regex search for

[ ]{2,}

would be better than \s\s (there's a normal space between those square-brackets)

[chaot]

Originally Posted by jackie_w

...
A Regex search for \s will match both a normal space and an \xa0 (and other 'whitespace' chars for that matter, e.g. newline). If you only want to search for '2 or more consecutive normal spaces' then perhaps a regex search for

[ ]{2,}

would be better than \s\s (there's a normal space between those square-brackets)

Brief interim report: Regex functions! And we see once again differences of occurences #3.

Replaced 1117 occurrences of [ ]{2,}