Forum: CFEngine Help Subject: Re: cfengine 3.1.5 pcre question Author: sauer Link to topic: https://cfengine.com/forum/read.php?3,21784,24146#msg-24146
Well, if this is going to be broughtback from the dead, someone may as well link to the Anchored v/s Unachored regexp documentation: http://cfengine.com/manuals/cf3-reference.html#Anchored-vs_002e-unanchored-regular-expressions Specifically, the comment pattern is unanchored - and the file is treated as a single multiline string. You need the end of line anchor if you want to only delete to the end of the line. The reason it's set up this way is so you can delete multi-line comments and similar (ie, you can use the /\*.*\*\// pattern (the reference manual regexp is much less accurate, IMHO) to delete /* C comments */ which span multiple lines). The dot matches *any* character, including an EOL. So using [^\n]* means "0 or more non-newline characters" and the pattern ends when it gets to the end of a line indicated by a \n or by the EOF. Using .* matches until the end of the file. If the file starts with a comment, therefore, the whole file gets discarded as a big ol' multi-line comment and you get no results back. You can probably do #.*$, or if that doesn't work, use \z instead of the $ (ie, use #.*\z to trim comments). The Perl/PCRE \z escape (lowercase z) matches just after the newline (which is probably \R to match any unicode line ending sequence) at the end of a line, and \z is independent of the multiline mode PCRE is using, so it's arguably "safer" than the $ in general. The upper-case \Z matches before and then again after the newline, so if \z makes things behave oddly in a specific case, try \Z. _______________________________________________ Help-cfengine mailing list Help-cfengine@cfengine.org https://cfengine.org/mailman/listinfo/help-cfengine
