python - Why is this regex experiencing catastrophic backtracking? -


i've read the articles , other questions on catastrophic backtracking in regular expressions , how can caused nested + , * quantifiers. however, regexes still encountering catastrophic backtracking without nested quantifiers. can me understand why?

i wrote these regexes search a specific type of rhyme in lines of welsh poetry. rhyme consists of consonants in beginning of line being repeated @ end, , there must space between beginning , end consonants. i've removed vowels, there 2 exceptions make these regexes ugly. first, there allowed consonants in middle don't repeat, , if there any, it's different type of rhyme. second, letters m, n, r, h, , v allowed interrupt rhyme (appear in beginning not in end or vice versa), can't ignored because rhyme consists of letters.

my script automatically builds regex each line , tests it. works rest of time, 1 line giving catastrophic backtracking. line's text without vowels is:

nn  frvvn  frv v

the regex automatically finds nn frvvn rhymes frv v, tries again last letter (the n in frvvn) required in back. if it's not required, rhyme can shortened. here's regex:

^(?p<s_letters>         # starting letters [mnrhv]*?\s*n{0,2}      # number of optional letters or number                         # of spaces can come between rhyming letters [mnrhv]*?\s*n{0,2} [mnrhv]*?\s*f{1,2} [mnrhv]*?\s*[rr]?(?:\s*[rr])? # r can rhyme r, that's                               # not relevant here (i think) [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*n{1,2} [mnrhv\s]*?) (?p<m_letters>          # middle letters [^\s]*?(?p<caesura>\s)  # caesura (end of rhyme)                         # first space after rhyme      .*)                     # end letters come late possible (?p<e_letters>          # end group [mnrhv]*?\s*n{0,2} [mnrhv]*?\s*n{0,2} [mnrhv]*?\s*f{1,2} [mnrhv]*?\s*[rr]?(?:\s*[rr])? [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*n{1,2} [mnrhv\s]*?)$

even though doesn't have nested quantifiers, still takes forever run. regexes other lines generated in same way run quickly. why this?

i'm not seeing nested quantifiers, seeing lot of ambiguities cause high-exponent polynomial runtime. example, consider part of regex:

[mnrhv]*?\s*[rr]?(?:\s*[rr])? # r can rhyme r, that's                               # not relevant here (i think) [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*v{0,2} [mnrhv]*?\s*n{1,2} [mnrhv\s]*?) (?p<m_letters>          # middle letters [^\s]*?(?p<caesura>\s)  # caesura (end of rhyme)

suppose regex engine @ point, , text it's seeing huge block of ns. ns can divided between following parts of regex:

[mnrhv]*?\s*[rr]?(?:\s*[rr])? ^^^^^^^^^  [mnrhv]*?\s*v{0,2} ^^^^^^^^^  [mnrhv]*?\s*v{0,2} ^^^^^^^^^ [mnrhv]*?\s*n{1,2} ^^^^^^^^^   ^^^^^^ [mnrhv\s]*?) ^^^^^^^^^^^ (?p<m_letters> [^\s]*?(?p<caesura>\s) ^^^^^^^

if number of ns n, there o(n**6) ways divide ns, since there 6 *? blocks match n here, , in between optional or matches n.

are \s parts mandatory? if so, might able improve runtime putting + instead of * on them.


Comments

Post a Comment

Popular posts from this blog

powershell Start-Process exit code -1073741502 when used with Credential from a windows service environment -

i'm running strange behavior powershell start-process call. here call: $process = start-process ` "c:\somepath\mybinary.exe" ` -passthru ` -credential $defaultcredential ` -wait ` -workingdirectory "c:\somepath" ` -loaduserprofile if ($process.exitcode -ne 0) { #do } this call return exit code - 1073741502 . after quick search, exit code seems related generic error when program not load required dll (aka. status_dll_init_failed ). when run without -credential $credential program runs correctly. in order isolate problem, manually launched some.exe in prompt target credential , runs smoothly. so problem seems come way start-process cmdlet launch process. i found potential solutions problem tried apply no luck : link , link . would have idea of what's going on here ? edit 1: run proc mon monitoring program activities when launched directly or via powershell script. problem seems occur when loading kernelbase
Read more

twig - Using Twigbridge in a Laravel 5.1 Package -

i'm trying create laravel 5.1 package uses twigbridge instead of .php views. can normal .php views display fine, when try make use of .twig templates i'm met error. possible use twigbridge templates when creating views laravel package or pretty stuck blade , php ? edit 1: loading twig template without extending works. it's when start using {% extends 'file' %} error comes out. i error error loading /virtual/laravel/packages/username/mypackage/src/views/sample.twig: template "templates.master" not defined () in "/virtual/laravel/packages/username/mypackage/src/views/sample.twig" @ line 1. my folders username |_ mypackage |_ src |_ controllers - samplecontroller.php |_ views |_ templates - master.twig - sample.twig twig template sample.twig {% extends 'templates.master' %} {% block content %} {{ parent()
Read more

c# - LINQ join Entities from HashSet's, Join vs Dictionary vs HashSet performance -

i have hashset's each stores t, have written test application compares different relation algorithms can think of, i'm not pleased results getting. does there exist more efficient ways of achieving object relations, once testing? using system; using system.collections.generic; using system.linq; using system.text; using system.threading.tasks; using system.diagnostics; namespace linqtests { class program { static void main(string[] args) { hashset<user> usertable = new hashset<user>(); hashset<userproperty> userpropertytable = new hashset<userproperty>(); #region lets create dummy data. console.writeline("please wait while creating dummy data, can take while..."); console.writeline(""); int rows = 1000000; for(int x = 0; x < rows; x++) { random rnd = new random(); // ad
Read more