usermanual issue #178 submitted.

I’ve taken this to a new discussion located here.
As far as the OP’s problem, this has been solved. I did not wish to further muddy these waters and have instead continued with a new post.

Terry

@Terry-R said in New line every n characters:

So Replace All with Wrap Around ticked starts at the first position in the file.
Replace All button with Wrap Around NOT ticked starts at the current position.

I disagree. AFAIK, if my cursor is on line 3 of a 5-line file, and there is a sequence that will match on all 5 lines, and I hit replace all with wrapping enabled, the first replacement will be on line 3, the second on line 4, the third on line 5, the fourth on line 1, and the fifth on line 2. I cannot think of a regex offhand that would cause that to be different than starting a replace-all at the begining of a file… but there might be one.

… experiment …

Nope, I was wrong.

Data:

one two three four five
fourteen something fifteen

regex = FIND:(?i-s)\b((?<!☺ )f\w+) => REPLACE:${1}☺

If my original assertion, and what is implied in the manual, were true, then starting with the cursor before four vs. before five and hitting Replace All would come up with different answers. Between the four and five, it would add a smiley to five if it really started wherever the cursor was, but would only add a smiley after four if it really started from the beginning.

With wrap-around enabled, it only puts the smiley after the four, not the five; without wrap-around, it puts it after the five if the cursor is between.

Is this something that needs to be amended?

Yes, if that’s what the manual says, then I think it could be better.

What does @PeterJones think, as our resident manual guru?

What I thought is different than what I could prove. I thought wrong.
usermanual issue #178 submitted.

So Replace All with Wrap Around ticked starts at the first position in the file.
Replace All button with Wrap Around NOT ticked starts at the current position.
From your example I see it works.

My earlier postings were not so great :-(
But I see that you got the point.

I’m not sure @guy038 's posting adds anything, except as another example of proof of how it works?

The manual does NOT appear to outline this VERY important change. At least not where Wrap Around and Replace All are described!
Is this something that needs to be amended?

Yes, if that’s what the manual says, then I think it could be better.
What does @PeterJones think, as our resident manual guru?

BTW, a posting somewhere here long ago (I think) said to think about Wrap around ticked as meaning “Entire document” in this (Replace All) circumstance, and I’ve always remembered that, as it seems to be appropriate.

Let’s consider this simple text, below, in a new Windows tab

123
456
789

with the last line without CRLF. So, the exact contents are the 13 bytes 1 2 3 CR LF 4 5 6 CR LF 7 8 9

FIRST test :

• Move the caret right after digit 5

• Run the regex S/R :

  • SEARCH (?s).{3}

  • REPLACE @

  • Untick the Wrap around option

  • Click on the Replace All button

we get the text :

123
45@@

without CRLF after the last @ char. Remember that the CR and LF characters count for ONE char because of the (?s) syntax !

• Then, move the caret at the very beginning of file, right before the digit 1

• Run the regex S/R :

  • SEARCH (?s).{3}

  • REPLACE @

  • The Wrap around option is still unticked

  • Click on the Replace All button

The text is now changed as :

@@@

without CRLF after the last @ char

So, it’s the final result if we suppose that the regex engine goes from the digit 6 to the digit 9, then from digit 1 to digit 5

SECOND test :

• Move, again, the caret right after digit 5

• Run the same regex S/R :

  • SEARCH (?s).{3}

  • REPLACE @

  • Tick the Wrap around option ( IMPORTANT )

  • Click on the Replace All button

we get the text :

@@@@9

without CRLF after the 9 digit

And this result proves that, although the caret was initially between digits 5 and 6, due to the Wrap around option ticked, the caret is internally moved at the very beginning of file and, then, the regex engine goes from digit 1 to digit 9 to process contents

As the total number of bytes is 13, this means that the search regex had sliced the text in 4 ranges of 3 chars, leaving the 9 digit alone, at the end ! Just logical ;-))

Best Regards,

guy038

So then the wrap around being ticked is actually redundant with a Replace All button press.

Sorry, I had read it differently the first time.

So Replace All with Wrap Around ticked starts at the first position in the file.
Replace All button with Wrap Around NOT ticked starts at the current position.

From your example I see it works. The manual does NOT appear to outline this VERY important change. At least not where Wrap Around and Replace All are described!

Is this something that needs to be amended?

Terry

Perhaps you don’t know that a Replace All search with Wrap around ticked does one transit through the file, starting at the very top, no matter where the caret starts out at?

I just checked the manual and there it says:
Replace All: Replaces all matches from the active location onward (following the Backward direction and Wrap around settings as appropriate).

That then seems to be wrong unless my interpretation is wrong of what active location means.

Terry

Perhaps you don’t know that a Replace All search with Wrap around ticked does one transit through the file, starting at the very top, no matter where the caret starts out at?

Oh, no I hadn’t realised that important fact. I suppose in tests I have always identified that as a potential issue when using the Replace button. Thank you for pointing out that important fact.

So then the wrap around being ticked is actually redundant with a Replace All button press.

Terry

Prove it thusly:

Find data for both cases: abcdefghijklmno
Replace data for both cases: @
Caret starting location for both cases: between the b and the c.

Reading my own posting I am stunned to see that I left off a very important part of the instructions, applicable to both Case 1 and 2:

• Put (?-s).{10} in the Find what box

And of course this doesn’t make any sense unless we’re in Regular expression for the Search mode.

Well, I use (?-s) as a “safety measure”. If I don’t intend line-ending characters to be matched by the ., I preamble with (?-s). For me, it’s a good habit to always specify (?s) or (?-s) when using .. And, I believe it is a good habit for forum postings as well, so that the . matches newline checkbox state doesn’t have to be provided as a separate item.

But sure, since there are no line-endings in the OP’s data…

Untick wrap around button as important the regex starts in very first position.

I don’t understand what is meant here. Well, I understand how to untick, but I don’t know what the rest means or how it is relevant.

Cursor MUST be in first position of file BEFORE regex run.

Hmmm, having Wrap around ticked negates the need for this, and is simpler than telling a user to move the caret.

Perhaps you don’t know that a Replace All search with Wrap around ticked does one transit through the file, starting at the very top, no matter where the caret starts out at?

Maybe you do know, but I’ll keep going for those that don’t…

Prove it thusly:

• Find data for both cases: abcdefghijklmno
• Replace data for both cases: @
• Caret starting location for both cases: between the b and the c.

Case 1:

• Tick Wrap around
• Press Replace all

Result: @klmno

Case 2:

• Untick Wrap around
• Press Replace all

Result: ab@mno

The key point here is that for Case 1 the search matches at start of file, NOT from the caret.

find: (?-s).{500}
repl: ${0}\r\n
Tick the Wrap around box, and press Replace All button

I would make some changes:

1. (?-s) is not required as no newline characters in file
2. Untick wrap around button as important the regex starts in very first position.
3. Cursor MUST be in first position of file BEFORE regex run.

All because OP stated it is all serial information, so ALL on one line.

Likely without these changes it will work out OK but if OP had moved through file first before starting the regex the cut points will be in the wrong positions.

My 2c worth
Terry

Well, it might be hard to just display it, but if you want to look at it that way I would suggest making a copy of it and then working with the copy in the following manner:

Set Search Mode in the Replace window to Regular expression

Then,

find: (?-s).{500}
repl: ${0}\r\n

Tick the Wrap around box, and press Replace All button.