Question

The following lines are arbitrary regexp which I need to use in lua.

['\";=]
!^(?:(?:[a-z]{3,10}\s+(?:\w{3,7}?://[\w\-\./]*(?::\d+)?)?/[^?#]*(?:\?[^#\s]*)?(?:#[\S]*)?|connect (?:\d{1,3}\.){3}\d{1,3}\.?(?::\d+)?|options \*)\s+[\w\./]+|get /[^?#]*(?:\?[^#\s]*)?(?:#[\S]*)?)$
'(?i:(?:c(?:o(?:n(?:t(?:entsmartz|actbot/)|cealed defense|veracrawler)|mpatible(?: ;(?: msie|\.)|-)|py(?:rightcheck|guard)|re-project/1.0)|h(?:ina(?: local browse 2\.|claw)|e(?:rrypicker|esebot))|rescent internet toolpak)|w(?:e(?:b(?: (?:downloader|by mail)|(?:(?:altb|ro)o|bandi)t|emailextract?|vulnscan|mole)|lls search ii|p Search 00)|i(?:ndows(?:-update-agent| xp 5)|se(?:nut)?bot)|ordpress(?: hash grabber|\/4\.01)|3mir)|m(?:o(?:r(?:feus fucking scanner|zilla)|zilla\/3\.mozilla\/2\.01$|siac 1.)|i(?:crosoft (?:internet explorer\/5\.0$|url control)|ssigua)|ailto:craftbot\@yahoo\.com|urzillo compatible)|p(?:ro(?:gram shareware 1\.0\.|duction bot|webwalker)|a(?:nscient\.com|ckrat)|oe-component-client|s(?:ycheclone|urf)|leasecrawl\/1\.|cbrowser|e 1\.4|mafind)|e(?:mail(?:(?:collec|harves|magne)t|(?: extracto|reape)r|(siphon|spider)|siphon|wolf)|(?:collecto|irgrabbe)r|ducate search vxb|xtractorpro|o browse)|t(?:(?: ?h ?a ?t ?' ?s g ?o ?t ?t ?a ? h ?u ?r ?|his is an exploi|akeou)t|oata dragostea mea pentru diavola|ele(?:port pro|soft)|uring machine)|a(?:t(?:(?:omic_email_hunt|spid)er|tache|hens)|d(?:vanced email extractor|sarobot)|gdm79\@mail\.ru|miga-aweb\/3\.4|utoemailspider| href=)|^(?:(google|i?explorer?\.exe|(ms)?ie( [0-9.]+)?\ ?(compatible( browser)?)?)$|www\.weblogs\.com|(?:jakart|vi)a|microsoft url|user-Agent)|s(?:e(?:archbot [e-mail]|curity scan)|(?:tress tes|urveybo)t|\.t\.a\.l\.k\.e\.r\.|afexplorer tl|itesnagger|hai)|n(?:o(?:kia-waptoolkit.* googlebot.*googlebot| browser)|e(?:(?:wt activeX; win3|uralbot\/0\.)2|ssus)|ameofagent|ikto)|f(?:a(?:(?:ntombrows|stlwspid)er|xobot)|(?:ranklin locato|iddle)r|ull web bot|loodgate|oobar/)|i(?:n(?:ternet(?: (?:exploiter sux|ninja)|-exprorer)|dy library)|sc systems irc search 2\.1)|g(?:ameBoy, powered by nintendo|rub(?: crawler|-client)|ecko\/25)|(myie2|libwen-us|murzillo compatible|webaltbot|wisenutbot)|b(?:wh3_user_agent|utch__2\.1\.1|lack hole|ackdoor)|d(?:ig(?:imarc webreader|out4uagent)|ts agent)|(?:(script|sql) inject|$botname/$botvers)ion|(msie .+; .*windows xp|compatible \; msie)|h(?:l_ftien_spider|hjhj@yahoo|anzoweb)|(?:8484 boston projec|xmlrpc exploi)t|u(?:nder the rainbow 2\.|ser-agent:)|(sogou develop spider|sohu agent)|(?:(?:d|e)browse|demo bot)|zeus(?: .*webster pro)?|[a-z]surf[0-9][0-9]|v(?:adixbot|oideye)|larbin@unspecified|\bdatacha0s\b|kenjin spider|; widows|rsync|\\\r))'

And there are many others where these came from…..
Point as you might noticed, the first case only the " is escaped with \" bot not the '
Hence,

rex_pcre.new('['\";=]')

Won’t work.

rex_pcre.new("['\";=]")

Should work, however, parts in the regex such as \-.

I also cannot use

[[ ]]

as there are regexp which ends with ] (first example)
breaking the lines as in

rex_pcre.new( [[
['\";=]
]])

won’t work for me in cases such as the third one which ends with ) and also raised an error of unexpected symbol.

in sum I am searching for such for the r"UNESCAPED STRING" of Python or the @"UNESCAPED STRING" of C#..

I assume there isn’t such, but wonder what is the way to get a similar functionality, given the fact, I only consume those value (regular expression) and have no control on how to compose them originally..

Here is my current solution
I simply try to compile the line, with [[ ]], if fail, move to " and then to “‘”/

EscapeRegEx = function (xp)
    -- try with [[ ]]
    local opening = '[['
    local closing = ']]'
    local codeline = "rex_pcre.new(" .. opening .. xp .. closing .. ")"
    _, err = loadstring(codeline)
    if not err then return codeline end

    -- then try with "
    opening = '"'
    closing = '"'
    codeline = "rex_pcre.new(" .. opening .. xp .. closing .. ")"
    _, err = loadstring(codeline)
    if not err then return codeline end

    -- then try with '
    opening = "'"
    closing = "'"
    codeline = "rex_pcre.new(" .. opening .. xp .. closing .. ")"
    _, err = loadstring(codeline)
    if not err then return codeline end

end

Answer 1

You can use longer versions of the long brackets:

[=========[the regex goes in here]=========]

The opening long bracket will only be matched by a closing long bracket of the same length.

See this for more details; you can also do a similar thing to get nested multi-line comments.