Using with vim substitute command: lines created are not scanned in future steps. Can they be?

shruggy · 10-21-2021, 09:36 AM

Quote:

Originally Posted by dedec0

I know regex well, for vim and a few other programs. But for sed, it is sometimes strange to use. I never try it, unless it is something ready to run, or with changes i understand how to do. Can you show something to me?

sed is basically the same:

Code:

sed -E 's/.{72}\S{,8}/&\n/g' your_file

With indentation and guarding against short lines that would be

Code:

sed -E '/.{81,}/s/.{72}\S{,8}/&\n\t/g' your_file

shruggy · 10-21-2021, 09:48 AM

discard

dedec0 · 10-21-2021, 09:50 AM

Quote:

Originally Posted by shruggy

Yep. 72 characters is the soft limit. Or what the fmt man page calls goal. The splitting occurs after each 72 characters plus 0 to 8 non-spacing characters. So 80 characters is the hard limit. The {,8} quantifier is greedy, so it will try to consume as many characters as it can (up to 8). As soon as there's a white space among those 8 characters, the RE fires and splitting occurs. If there's no white space in that range, the splitting occurs after the 80th character anyway.

I am still not sure i understand it. Does this splitting happens only from 73 to 80 spaces, or before 81 if there was no space in this interval?

shruggy · 10-21-2021, 09:58 AM

Quote:

Originally Posted by dedec0

Does this splitting happens only from 73 to 80 spaces, or before 81 if there was no space in this interval?

The splitting happens before the first white space after the 72th character, or before the 81th character if there was no white space in that interval.

dedec0 · 10-21-2021, 10:01 AM

Quote:

Originally Posted by shruggy

sed is basically the same:

Code:

sed -E 's/.{72}\S{,8}/&\n/g' your_file

With indentation and guarding against short lines that would be

Code:

sed -E '/.{81,}/s/.{72}\S{,8}/&\n\t/g' your_file

The first command is pretty much what is written in Vim - just the magic/nonmagic changes, right?

But can you explain me the second command? Or i will have to study sed manual and/or something else about it. A pointer is welcome.

shruggy · 10-21-2021, 10:05 AM

Quote:

Originally Posted by dedec0

The first command is pretty much what is written in Vim - just the magic/nonmagic changes, right?

Right.

Quote:

Originally Posted by dedec0

But can you explain me the second command? Or i will have to study sed manual and/or something else about it. A pointer is welcome.

In Vim parlance, that would be

Code:

:g/.\{81,}/s/.\{72}\S\{,8}/&\r\t/g

dedec0 · 10-21-2021, 10:06 AM

Quote:

Originally Posted by shruggy

The splitting happens before the first white space after the 72th character, or before the 81th character if there was no white space in that interval.

If i understand you right, this is not my idea.

Considering your limits, but my wish: it should be in the *last* space after 72th, but before 80th; or before the 81th character.

With my limits: the *last* space before 80th character; or before the 81th character (which means no space was found until it).

shruggy · 10-21-2021, 10:17 AM

Quote:

Originally Posted by dedec0

Or i will have to study sed manual and/or something else about it. A pointer is welcome.

See 4.1 Addresses overview in the GNU sed manual.

shruggy · 10-21-2021, 10:17 AM

Quote:

Originally Posted by dedec0

Or i will have to study sed manual and/or something else about it. A pointer is welcome.

See 4.1 Addresses overview in the GNU sed manual.

shruggy · 10-21-2021, 10:38 AM

Quote:

Originally Posted by dedec0

With my limits: the *last* space before 80th character; or before the 81th character (which means no space was found until it).

Like this?

Code:

sed -E 's/.{,79}\s|\S{80}/&\n/g'

dedec0 · 10-21-2021, 01:13 PM

Quote:

Originally Posted by shruggy

Like this?

Code:

sed -E 's/.{,79}\s|\S{80}/&\n/g'

Exactly what i was thinking about writing, after seeing your command and explanation. Simpler than most things we have seen here. But i kind of worried if it would work, since any of you wrote that regex (in vim or not).

There is a problem with this: it will split lines shorter than N, if these lines have a space:

1. File:

Code:

1234567890 234567890 234567890
1234567890 234567890 234567890
12345678901234567890 234567890
12345678901234567890 234567890
12345678901234567890 234567890
33 66 999
33 66 999
33 66 999

2. Command:

Code:

:%!sed -E 's/.{,13}\s|\S{14}/&\n/g'

3. Result:

Code:

1234567890 
234567890 
234567890
1234567890 
234567890 
234567890
12345678901234
567890 
234567890
12345678901234
567890 
234567890
12345678901234
567890 
234567890
33 66 
999
33 66 
999
33 66 
999

shruggy · 10-21-2021, 01:21 PM

Quote:

Originally Posted by dedec0

There is a problem with this: it will split lines shorter than N, if these lines have a space:

Because that was your requirement. Compare

Quote:

Originally Posted by dedec0

Considering your limits, but my wish: it should be in the *last* space after 72th, but before 80th; or before the 81th character.

Quote:

Originally Posted by dedec0

With my limits: the *last* space before 80th character; or before the 81th character (which means no space was found until it).

So I did it to your limits. I understand your limits thusly: the *last* space before 80th character, even if it is the 2nd character in the line.

dedec0 · 10-21-2021, 01:28 PM

Quote:

Originally Posted by shruggy

sed is basically the same:

Code:

sed -E 's/.{72}\S{,8}/&\n/g' your_file

With indentation and guarding against short lines that would be

Code:

sed -E '/.{81,}/s/.{72}\S{,8}/&\n\t/g' your_file

shruggy, i think there are strange things in this regex you wrote. See:

Code:

/.{81,}/s/.{72}\S{,8}/&\n\t/g
^      ^ ^           ^     ^
1      2 3           4     5

My questions:

a. Should 2 be a slash or a backslash?

b. Assuming that 2 is a backslash, why this regex has 3 parts, with 4 slashes separating them, instead of just 2? (this is something i have never seen before, so i did not even bother to search about it)

If there are these mistakes in your messages, i suggest that you edit them. The commands may not solve my problem, but others may find them useful.

shruggy · 10-21-2021, 01:39 PM

No, the sed code is correct. It's not one regex, it's a sed command consisting out of a regex address and a substitute command.

Code:

/.{81,}/s/.{72}\S{,8}/&\n\t/g

BTW, have you seen my answer in #21?

dedec0 · 10-21-2021, 01:42 PM

Quote:

Originally Posted by shruggy

In Vim parlance, that would be

Code:

:g/.\{81,}/s/.\{72}\S\{,8}/&\r\t/g

New command for me: ":g". I will read its help.

But your suggestion did not work:

1. File, with one line only:

Code:

1234567890 234567890 234567890

2. Command:

Code:

:g/.\{15,}/s/.\{11}\S\{,4}/&\r\t/g

3. Result:

Code:

1234567890 2345
	67890 234567890
	# there is a tab inserted here!

The result is indented, but the indentation make the new line go beyound the demanded limit of 14 chars (in my example). And there is an extra tab, in the end of the file.

People, can you, please, write a sentence or a short paragraph to describe the workings of a regex or some command you send? This can be valuable, i think, and it is not hard.

shruggy · 10-21-2021, 01:52 PM

OK. Then how about

Code:

%:s/\v.{10}\S{,4}($)@!/&\r/g