CMake regex match

Question

I have a CMake string variable:

set(var "String0 String_1 String_2")

I need to select whatever is before the first whitespace from this variable ("String0") and make a new variable with this content.

I used CMake's REGEX MATCH method to do this and used this regex: '\S\w*'. I have tested that regex on an online regex interpreter and it worked.

I have written this code in CMake (after setting the variable of course):

string(REGEX MATCH "\S\w*" NEW_VAR "${VAR}")

When I do this the script complains about invalid skipping of characters (S and w). So, next, I tried escaping both the slashes:

string(REGEX MATCH "\\S\\w*" NEW_VAR "${VAR}")

Now NEW_VAR equal 'S' instead of "String0" as I was expecting.

Please help me correct this, since I have very little experience with regular expressions and CMake.

CMake has its own regex specification, and this specification lacks of character classes like \S and \w. — Tsyvarev
– Tsyvarev, Commented Sep 10, 2018 at 9:44
thank you. I have found a regex that works. Not sure holw good it is for all cases but from what I've tested seems to work — Kennedy
– Kennedy, Commented Sep 10, 2018 at 10:48
On Stack Overflow we tend to NOT answer on the question in the question post itself. Instead, you are allowed to answer on your question. Please, move your last Edit paragrapgh into the answer post. — Tsyvarev
– Tsyvarev, Commented Sep 10, 2018 at 11:51

compor · Accepted Answer · 2018-09-10 09:48:45Z

4

You could also search for the location of the first occurrence of a whitespace character and select the substring from the start of this string till that position:

set(var "String_0 String_1 String_2")

string(FIND ${var} " " wsloc)
message(STATUS "position of first whitespace character: ${wsloc}")

string(SUBSTRING ${var} 0 ${wsloc} newvar)
message(STATUS "selected substring: \"${newvar}\"")

edited Sep 10, 2018 at 9:48

answered Sep 10, 2018 at 9:41

compor

2,3591 gold badge22 silver badges30 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Kennedy Over a year ago

thanks but I was looking for smth smaller like a line of code no more :D

compor Over a year ago

@Kennedy not stated in the question, but no worries. This is just 2 lines and uses no regex, so there might be some merit there; not sure what strings you're coming up against.

Kennedy Over a year ago

it worked for some examples i've tried in the interpreter. should be ok thanks :D

Alexis Wilke · Accepted Answer · 2023-04-22 18:15:09Z

0

I have used this and it works(not sure how good it is for all cases, though):

.[(A-z)|(a-z)|(0-9)]*

edited Apr 22, 2023 at 18:15

Alexis Wilke

21.2k11 gold badges111 silver badges183 bronze badges

answered Sep 24, 2018 at 8:23

Kennedy

3252 gold badges8 silver badges21 bronze badges

2 Comments

tripleee Over a year ago

You are almost certainly mixing two different regex features here. The stuff between [ and ] is simply an enumeration of characters, with no other special facilities than character ranges and negation. You are including ( and ) and | in the character class (multiple times), probably needlessly.

Alexis Wilke Over a year ago

As tripleee says, you would use .[A-Za-z0-9]*

Collectives™ on Stack Overflow

CMake regex match

2 Answers 2

3 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related