/[pcre]/code/trunk/testdata/testinput2
ViewVC logotype

Diff of /code/trunk/testdata/testinput2

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 471 by ph10, Fri Dec 11 15:11:55 2009 UTC revision 550 by ph10, Sun Oct 10 16:24:11 2010 UTC
# Line 2  Line 2 
2      of PCRE's API, error diagnostics, and the compiled code of some patterns.      of PCRE's API, error diagnostics, and the compiled code of some patterns.
3      It also checks the non-Perl syntax the PCRE supports (Python, .NET,      It also checks the non-Perl syntax the PCRE supports (Python, .NET,
4      Oniguruma). Finally, there are some tests where PCRE and Perl differ,      Oniguruma). Finally, there are some tests where PCRE and Perl differ,
5      either because PCRE can't be compatible, or there is potential Perl      either because PCRE can't be compatible, or there is a possible Perl
6      bug. --/      bug. --/
7    
8  /-- Originally, the Perl 5.10 things were in here too, but now I have separated  /-- Originally, the Perl 5.10 and 5.11 things were in here too, but now I have
9      many (most?) of them out into test 11. However, there may still be some      separated many (most?) of them out into test 11. However, there may still
10      that were overlooked. --/      be some that were overlooked. --/
11    
12  /(a)b|/I  /(a)b|/I
13    
# Line 51  Line 51 
51    
52  /(?X)[\B]/  /(?X)[\B]/
53    
54    /(?X)[\R]/
55    
56    /(?X)[\X]/
57    
58    /[\B]/BZ
59    
60    /[\R]/BZ
61    
62    /[\X]/BZ
63    
64  /[z-a]/  /[z-a]/
65    
66  /^*/  /^*/
# Line 344  Line 354 
354      *** Failers      *** Failers
355      a      a
356    
357  /This one is here because I think Perl 5.005_02 gets the setting of $1 wrong/I  /This one is here because Perl behaves differently; see also the following/I
358    
359  /^(a\1?){4}$/I  /^(a\1?){4}$/I
360        aaaa
361      aaaaaa      aaaaaa
362    
363    /Perl does not fail these two for the final subjects. Neither did PCRE until/
364    /release 8.01. The problem is in backtracking into a subpattern that contains/
365    /a recursive reference to itself. PCRE has now made these into atomic patterns./
366    
367    /^(xa|=?\1a){2}$/
368        xa=xaa
369        ** Failers
370        xa=xaaa
371    
372    /^(xa|=?\1a)+$/
373        xa=xaa
374        ** Failers
375        xa=xaaa
376    
377  /These are syntax tests from Perl 5.005/I  /These are syntax tests from Perl 5.005/I
378    
# Line 2264  a random value. /Ix Line 2289  a random value. /Ix
2289  /a+b?(*THEN)c+(*FAIL)/C  /a+b?(*THEN)c+(*FAIL)/C
2290      aaabccc      aaabccc
2291    
 /a(*PRUNE:XXX)b/  
   
2292  /a(*MARK)b/  /a(*MARK)b/
2293    
2294  /(?i:A{1,}\6666666666)/  /(?i:A{1,}\6666666666)/
# Line 3175  a random value. /Ix Line 3198  a random value. /Ix
3198  /()i(?(1)a)/SI  /()i(?(1)a)/SI
3199      ia      ia
3200    
3201    /(?i)a(?-i)b|c/BZ
3202        XabX
3203        XAbX
3204        CcC
3205        ** Failers
3206        XABX
3207    
3208    /(?i)a(?s)b|c/BZ
3209    
3210    /(?i)a(?s-i)b|c/BZ
3211    
3212    /^(ab(c\1)d|x){2}$/BZ
3213        xabcxd
3214    
3215    /^(?&t)*+(?(DEFINE)(?<t>.))$/BZ
3216    
3217    /^(?&t)*(?(DEFINE)(?<t>.))$/BZ
3218    
3219    / -- The first four of these are not in the Perl 5.10 test because Perl
3220         documents that the use of \K in assertions is "not well defined". The
3221         last is here because Perl gives the match as "b" rather than "ab". I
3222         believe this to be a Perl bug. --/
3223    
3224    /(?=a\Kb)ab/
3225        ab
3226    
3227    /(?!a\Kb)ac/
3228        ac
3229    
3230    /^abc(?<=b\Kc)d/
3231        abcd
3232    
3233    /^abc(?<!b\Kq)d/
3234        abcd
3235    
3236    /(?>a\Kb)z|(ab)/
3237        ab
3238    
3239    /----------------------/
3240    
3241    /(?P<L1>(?P<L2>0|)|(?P>L2)(?P>L1))/
3242    
3243    /abc(*MARK:)pqr/
3244    
3245    /abc(*:)pqr/
3246    
3247    /abc(*FAIL:123)xyz/
3248    
3249    /--- This should, and does, fail. In Perl, it does not, which I think is a
3250         bug because replacing the B in the pattern by (B|D) does make it fail. ---/
3251    
3252    /A(*COMMIT)B/+K
3253        ACABX
3254    
3255    /--- These should be different, but in Perl 5.11 are not, which I think
3256         is a bug in Perl. ---/
3257    
3258    /A(*THEN)B|A(*THEN)C/K
3259        AC
3260    
3261    /A(*PRUNE)B|A(*PRUNE)C/K
3262        AC
3263    
3264    /--- A whole lot of tests of verbs with arguments are here rather than in test
3265         11 because Perl doesn't seem to follow its specification entirely
3266         correctly. ---/
3267    
3268    /--- Perl 5.11 sets $REGERROR on the AC failure case here; PCRE does not. It is
3269         not clear how Perl defines "involved in the failure of the match". ---/
3270    
3271    /^(A(*THEN:A)B|C(*THEN:B)D)/K
3272        AB
3273        CD
3274        ** Failers
3275        AC
3276        CB
3277    
3278    /--- Check the use of names for success and failure. PCRE doesn't show these
3279    names for success, though Perl does, contrary to its spec. ---/
3280    
3281    /^(A(*PRUNE:A)B|C(*PRUNE:B)D)/K
3282        AB
3283        CD
3284        ** Failers
3285        AC
3286        CB
3287    
3288    /--- An empty name does not pass back an empty string. It is the same as if no
3289    name were given. ---/
3290    
3291    /^(A(*PRUNE:)B|C(*PRUNE:B)D)/K
3292        AB
3293        CD
3294    
3295    /--- PRUNE goes to next bumpalong; COMMIT does not. ---/
3296    
3297    /A(*PRUNE:A)B/K
3298        ACAB
3299    
3300    /(*MARK:A)(*PRUNE:B)(C|X)/K
3301        C
3302        D
3303    
3304    /(*MARK:A)(*THEN:B)(C|X)/K
3305        C
3306        D
3307    
3308    /--- This should fail, as the skip causes a bump to offset 3 (the skip) ---/
3309    
3310    /A(*MARK:A)A+(*SKIP)(B|Z) | AC/xK
3311        AAAC
3312    
3313    /--- Same --/
3314    
3315    /A(*MARK:A)A+(*MARK:B)(*SKIP:B)(B|Z) | AC/xK
3316        AAAC
3317    
3318    /--- This should fail; the SKIP advances by one, but when we get to AC, the
3319         PRUNE kills it. ---/
3320    
3321    /A(*PRUNE:A)A+(*SKIP:A)(B|Z) | AC/xK
3322        AAAC
3323    
3324    /A(*:A)A+(*SKIP)(B|Z) | AC/xK
3325        AAAC
3326    
3327    /--- This should fail, as a null name is the same as no name ---/
3328    
3329    /A(*MARK:A)A+(*SKIP:)(B|Z) | AC/xK
3330        AAAC
3331    
3332    /--- This fails in PCRE, and I think that is in accordance with Perl's
3333         documentation, though in Perl it succeeds. ---/
3334    
3335    /A(*MARK:A)A+(*SKIP:B)(B|Z) | AAC/xK
3336        AAAC
3337    
3338    /--- Mark names can be duplicated ---/
3339    
3340    /A(*:A)B|X(*:A)Y/K
3341        AABC
3342        XXYZ
3343    
3344    /^A(*:A)B|^X(*:A)Y/K
3345        ** Failers
3346        XAQQ
3347    
3348    /--- A check on what happens after hitting a mark and them bumping along to
3349    something that does not even start. Perl reports tags after the failures here,
3350    though it does not when the individual letters are made into something
3351    more complicated. ---/
3352    
3353    /A(*:A)B|XX(*:B)Y/K
3354        AABC
3355        XXYZ
3356        ** Failers
3357        XAQQ
3358        XAQQXZZ
3359        AXQQQ
3360        AXXQQQ
3361    
3362    /--- COMMIT at the start of a pattern should be the same as an anchor. Perl
3363    optimizations defeat this. So does the PCRE optimization unless we disable it
3364    with \Y. ---/
3365    
3366    /(*COMMIT)ABC/
3367        ABCDEFG
3368        ** Failers
3369        DEFGABC\Y
3370    
3371    /--- Repeat some tests with added studying. ---/
3372    
3373    /A(*COMMIT)B/+KS
3374        ACABX
3375    
3376    /A(*THEN)B|A(*THEN)C/KS
3377        AC
3378    
3379    /A(*PRUNE)B|A(*PRUNE)C/KS
3380        AC
3381    
3382    /^(A(*THEN:A)B|C(*THEN:B)D)/KS
3383        AB
3384        CD
3385        ** Failers
3386        AC
3387        CB
3388    
3389    /^(A(*PRUNE:A)B|C(*PRUNE:B)D)/KS
3390        AB
3391        CD
3392        ** Failers
3393        AC
3394        CB
3395    
3396    /^(A(*PRUNE:)B|C(*PRUNE:B)D)/KS
3397        AB
3398        CD
3399    
3400    /A(*PRUNE:A)B/KS
3401        ACAB
3402    
3403    /(*MARK:A)(*PRUNE:B)(C|X)/KS
3404        C
3405        D
3406    
3407    /(*MARK:A)(*THEN:B)(C|X)/KS
3408        C
3409        D
3410    
3411    /A(*MARK:A)A+(*SKIP)(B|Z) | AC/xKS
3412        AAAC
3413    
3414    /A(*MARK:A)A+(*MARK:B)(*SKIP:B)(B|Z) | AC/xKS
3415        AAAC
3416    
3417    /A(*PRUNE:A)A+(*SKIP:A)(B|Z) | AC/xKS
3418        AAAC
3419    
3420    /A(*:A)A+(*SKIP)(B|Z) | AC/xKS
3421        AAAC
3422    
3423    /A(*MARK:A)A+(*SKIP:)(B|Z) | AC/xKS
3424        AAAC
3425    
3426    /A(*MARK:A)A+(*SKIP:B)(B|Z) | AAC/xKS
3427        AAAC
3428    
3429    /A(*:A)B|XX(*:B)Y/KS
3430        AABC
3431        XXYZ
3432        ** Failers
3433        XAQQ
3434        XAQQXZZ
3435        AXQQQ
3436        AXXQQQ
3437    
3438    /(*COMMIT)ABC/
3439        ABCDEFG
3440        ** Failers
3441        DEFGABC\Y
3442    
3443    /^(ab (c+(*THEN)cd) | xyz)/x
3444        abcccd
3445    
3446    /^(ab (c+(*PRUNE)cd) | xyz)/x
3447        abcccd
3448    
3449    /^(ab (c+(*FAIL)cd) | xyz)/x
3450        abcccd
3451    
3452    /--- Perl 5.11 gets some of these wrong ---/
3453    
3454    /(?>.(*ACCEPT))*?5/
3455        abcde
3456    
3457    /(.(*ACCEPT))*?5/
3458        abcde
3459    
3460    /(.(*ACCEPT))5/
3461        abcde
3462    
3463    /(.(*ACCEPT))*5/
3464        abcde
3465    
3466    /A\NB./BZ
3467        ACBD
3468        *** Failers
3469        A\nB
3470        ACB\n
3471    
3472    /A\NB./sBZ
3473        ACBD
3474        ACB\n
3475        *** Failers
3476        A\nB
3477    
3478    /A\NB/<crlf>
3479        A\nB
3480        A\rB
3481        ** Failers
3482        A\r\nB
3483    
3484    /\R+b/BZ
3485    
3486    /\R+\n/BZ
3487    
3488    /\R+\d/BZ
3489    
3490    /\d*\R/BZ
3491    
3492    /\s*\R/BZ
3493    
3494    /-- Perl treats this one differently, not failing the second string. I believe
3495        that is a bug in Perl. --/
3496    
3497    /^((abc|abcx)(*THEN)y|abcd)/
3498        abcd
3499        *** Failers
3500        abcxy
3501    
3502  /-- End of testinput2 --/  /-- End of testinput2 --/
Legend:
Removed from v.471  
changed lines
  Added in v.550
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12