new regex question ( trailing 0 or 1 instance permultations

[cybmole]

I could not see this covered in the regex sticky

I want to write a single find/replace that will zap all of

line-height:1.2;
line-height:1.3
line-height:1.2em;
line-height:1.2em
line-height: 1.2;
line-height: 1.3
line-height: 1.2em;
line-height: 1.2em

i.e. remove all fixed line heights in range 1-2 em from CSS.
NBthey may or may not have an explicit em & they may nor may not have a space after the first colon, and they may or may not have a trailing ;

I've got this far:
line-height:\s?1\.\d
I can't seem to also capture the last bits ?

update:
this works but it feels like a cheat, because it just grabs everything up to the next bit of white space. you could probably write a valid CSS with no white space between statements.
line-height:\s?1\.\d(.*)\s

Anyway, I'd like to learn the proper way to find and include zero or one instance of a given string

[Jellby]

How about this?

\s*line-height:\s*1(\.[0123]\d*)?(em)?\s*;

"line-height:" with any number of possible preceding spaces and any number of spaces after,
a "1",
possibly, a period, a 0, 1, 2 or 3, and any number of additional digits,
possibly, "em",
a semicolon preceded with any number of possible spaces.

That should catch "1em", "1.1333333", "1.2em", etc., but not not "1.5em" or "0"

[cybmole]

sounds great in theory but is not finding all the matches in my simple test file ?
e.g. it is skipping this line :
line-height:1.3
i think it needs a tweak to allow for no trailing semicolon ?

for testing, I created a blank stylesheet & put in

line-height:1.2;
line-height:1.3
line-height:1.2em;
line-height:1.2em
line-height: 1.9;
line-height: 1.4
line-height: 1.2em;
line-height: 1.2em
line-height:1.9;
line-height:1.4
line-height:1.2em;
line-height:1.2em

also your code works for the same reason as my cheat, it looks for a final mandated whitespace. if I remove that from your code then the find picks out only the bold part of
line-height:1.2em

the hard bit is getting the optional components to be included in what is captured, and allowing the final bit of the target to be an optional string

[Jellby]

The semicolon is only optional if it's the final element in a style. Maybe this?

\s*line-height:\s*1(\.[0123]\d*)?(em)?\s*[;}]

and check "multiline", or whatever there is to include newlines in \s. This, however, will not work if you want to remove everything (you should not remove the } ). You could preprocess the file and first replace every } with ;}, now every rule should be ended by a semicolon (and there will possibly be two consecutive semicolons in some cases, which you could remove afterwards).

[Doitsu]

How about this quick and dirty regex: line-height:\s*1\.\d+[em;]*

[cybmole]

Originally Posted by Jellby

The semicolon is only optional if it's the final element in a style. Maybe this?

\s*line-height:\s*1(\.[0123]\d*)?(em)?\s*[;}]

...

i know about the final element rule, but my test case examples are all from actual book syntax

your new code still skips this case
line-height:1.3

[cybmole]

Originally Posted by Doitsu

How about this quick and dirty regex: line-height:\s*1\.\d+[em;]

it stops too soon e.g. it captures only the bold below
line-height:1.2em;

[Doitsu]

Originally Posted by cybmole

it stops too soon e.g. it captures only the bold below
line-height:1.2em;

What kind of editor do you use to test these regular expression?
The expression that I suggested will capture all test values in Notepad+, and, after adding an asterisk at the end, also with EmEdit.

[cybmole]

i am testing with sigil !

I want to be able to use the solution in sigil

forum name is a clue

[PeterT]

Originally Posted by cybmole

it stops too soon e.g. it captures only the bold below
line-height:1.2em;

You're missing the * after the [em;]

[Doitsu]

Originally Posted by PeterT

You're missing the * after the [em;]

You're right. I corrected my initial post. The following expression finds all test case strings:

line-height:\s*1\.\d+[em;]*

[Jellby]

Originally Posted by cybmole

i know about the final element rule, but my test case examples are all from actual book syntax

your new code still skips this case
line-height:1.3

That's because that line is wrong, it's missing the semicolon. If it's in a real book, it would fail in ADE I guess.

Originally Posted by Doitsu

You're right. I corrected my initial post. The following expression finds all test case strings:

line-height:\s*1\.\d+[em;]*

It would also find 1.5mm

How about just ending with ";?" ?

\s*line-height:\s*1(\.[0123]\d*)?(em)?\s*;?

[cybmole]

Originally Posted by PeterT

You're missing the * after the [em;]

with the * added, it still stops here;
line-height:1.2em;

[cybmole]

Originally Posted by Jellby

That's because that line is wrong, it's missing the semicolon. If it's in a real book, it would fail in ADE I guess.



It would also find 1.5mm

How about just ending with ";?" ?

\s*line-height:\s*1(\.[0123]\d*)?(em)?\s*;?

stops here
line-height:1.2em;

PS
ADE is quite happy for a real book to contain this in CSS - * I have seen it many times - no trailing ;

.calibre99
{
lineheight:1.2
}

but hey, I never claimed it was an easy puzzle

clearly " remove all fixed line heights from a given book" is not a trivial thing to automate

[Jellby]

Originally Posted by cybmole

stops here
line-height:1.2em;

Hmm... I can't see why, but I have no Sigil to test.
Ah, well, I sort of see why... Isn't there an option for "greedy" matching?

ADE is quite happy for a real book to contain this in CSS - * I have seen it many times - no trailing ;

.calibre99
{
lineheight:1.2
}

Yes, that's fine, and that should be catched by my previous suggestion (with multiline enabled), but this is not:

line-height:1.3
line-height:1.2em;

I understood that you were testing with that.