# reg.test -- # # This file contains a collection of tests for one or more of the Tcl # built-in commands. Sourcing this file into Tcl runs the tests and # generates output for errors. No output means no errors were found. # (Don't panic if you are seeing this as part of the reg distribution # and aren't using Tcl -- reg's own regression tester also knows how # to read this file, ignoring the Tcl-isms.) # # Copyright (c) 1998, 1999 Henry Spencer. All rights reserved. # # RCS: @(#) $Id: reg.test,v 1.16.2.1 2003/10/06 14:30:06 dgp Exp $ if {[lsearch [namespace children] ::tcltest] == -1} { package require tcltest 2 namespace import -force ::tcltest::* } # All tests require the testregexp command, return if this # command doesn't exist ::tcltest::testConstraint testregexp \ [expr {[info commands testregexp] != {}}] ::tcltest::testConstraint localeRegexp 0 # This file uses some custom procedures, defined below, for regexp regression # testing. The name of the procedure indicates the general nature of the # test: # e compile error expected # f match failure expected # m successful match # i successful match with -indices (used in checking things like # nonparticipating subexpressions) # p unsuccessful match with -indices (!!) (used in checking # partial-match reporting) # There is also "doing" which sets up title and major test number for each # block of tests. # The first 3 arguments are constant: a minor number (which often gets # a letter or two suffixed to it internally), some flags, and the RE itself. # For e, the remaining argument is the name of the compile error expected, # less the leading "REG_". For the rest, the next argument is the string # to try the match against. Remaining arguments are the substring expected # to be matched, and any substrings expected to be matched by subexpressions. # (For f, these arguments are optional, and if present are ignored except # that they indicate how many subexpressions should be present in the RE.) # It is an error for the number of subexpression arguments to be wrong. # Cases involving nonparticipating subexpressions, checking where empty # substrings are located, etc. should be done using i and p. # The flag characters are complex and a bit eclectic. Generally speaking, # lowercase letters are compile options, uppercase are expected re_info # bits, and nonalphabetics are match options, controls for how the test is # run, or testing options. The one small surprise is that AREs are the # default, and you must explicitly request lesser flavors of RE. The flags # are as follows. It is admitted that some are not very mnemonic. # There are some others which are purely debugging tools and are not # useful in this file. # # - no-op (placeholder) # + provide fake xy equivalence class and ch collating element # % force small state-set cache in matcher (to test cache replace) # ^ beginning of string is not beginning of line # $ end of string is not end of line # * test is Unicode-specific, needs big character set # # & test as both ARE and BRE # b BRE # e ERE # a turn advanced-features bit on (error unless ERE already) # q literal string, no metacharacters at all # # i case-independent matching # o ("opaque") no subexpression capture # p newlines are half-magic, excluded from . and [^ only # w newlines are half-magic, significant to ^ and $ only # n newlines are fully magic, both effects # x expanded RE syntax # t incomplete-match reporting # # A backslash-_a_lphanumeric seen # B ERE/ARE literal-_b_race heuristic used # E backslash (_e_scape) seen within [] # H looka_h_ead constraint seen # I _i_mpossible to match # L _l_ocale-specific construct seen # M unportable (_m_achine-specific) construct seen # N RE can match empty (_n_ull) string # P non-_P_OSIX construct seen # Q {} _q_uantifier seen # R back _r_eference seen # S POSIX-un_s_pecified syntax seen # T prefers shortest (_t_iny) # U saw original-POSIX botch: unmatched right paren in ERE (_u_gh) # The one area we can't easily test is memory-allocation failures (which # are hard to provoke on command). Embedded NULs also are not tested at # the moment, but this is a historical accident which should be fixed. # test procedures and related set ask "about" set xflags "xflags" set testbypassed 0 # re_info abbreviation mapping table set infonames(A) "REG_UBSALNUM" set infonames(B) "REG_UBRACES" set infonames(E) "REG_UBBS" set infonames(H) "REG_ULOOKAHEAD" set infonames(I) "REG_UIMPOSSIBLE" set infonames(L) "REG_ULOCALE" set infonames(M) "REG_UUNPORT" set infonames(N) "REG_UEMPTYMATCH" set infonames(P) "REG_UNONPOSIX" set infonames(Q) "REG_UBOUNDS" set infonames(R) "REG_UBACKREF" set infonames(S) "REG_UUNSPEC" set infonames(T) "REG_USHORTEST" set infonames(U) "REG_UPBOTCH" set infonameorder "RHQBAUEPSMLNIT" ;# must match bit order, lsb first # set major test number and description proc doing {major desc} { global prefix description testbypassed if {$testbypassed != 0} { puts stdout "!!! bypassed $testbypassed tests in\ $prefix, `$description'" } set prefix reg-$major set description "reg $desc" set testbypassed 0 } # build test number (internal) proc tno {testid} { return [join $testid .] } # build description, with possible modifiers (internal) proc desc {testid} { global description set d $description if {[llength $testid] > 1} { set d "([lreplace $testid 0 0]) $d" } return $d } # build trailing options and flags argument from a flags string (internal) proc flags {fl} { global xflags set args [list] set flags "" foreach f [split $fl ""] { switch -exact -- $f { "i" { lappend args "-nocase" } "x" { lappend args "-expanded" } "n" { lappend args "-line" } "p" { lappend args "-linestop" } "w" { lappend args "-lineanchor" } "-" { } default { append flags $f } } } if {[string compare $flags ""] != 0} { lappend args -$xflags $flags } return $args } # build info-flags list from a flags string (internal) proc infoflags {fl} { global infonames infonameorder set ret [list] foreach f [split $infonameorder ""] { if {[string first $f $fl] >= 0} { lappend ret $infonames($f) } } return $ret } # compilation error expected proc e {testid flags re err} { global prefix ask errorCode # Tcl locale stuff doesn't do the ch/xy test fakery yet if {[string first "+" $flags] >= 0} { # This will register as a skipped test test $prefix.[tno $testid] [desc $testid] localeRegexp {} {} return } # if &, test as both ARE and BRE set amp [string first "&" $flags] if {$amp >= 0} { set f [string range $flags 0 [expr $amp - 1]] append f [string range $flags [expr $amp + 1] end] e [linsert $testid end ARE] ${f} $re $err e [linsert $testid end BRE] ${f}b $re $err return } set cmd [concat [list testregexp -$ask] [flags $flags] [list $re]] set run "list \[catch \{$cmd\}\] \[lindex \$errorCode 1\]" test $prefix.[tno $testid] [desc $testid] \ {testregexp} $run [list 1 REG_$err] } # match failure expected proc f {testid flags re target args} { global prefix description ask # Tcl locale stuff doesn't do the ch/xy test fakery yet if {[string first "+" $flags] >= 0} { # This will register as a skipped test test $prefix.[tno $testid] [desc $testid] localeRegexp {} {} return } # if &, test as both ARE and BRE set amp [string first "&" $flags] if {$amp >= 0} { set f [string range $flags 0 [expr $amp - 1]] append f [string range $flags [expr $amp + 1] end] eval [linsert $args 0 f [linsert $testid end ARE] ${f} $re \ $target] eval [linsert $args 0 f [linsert $testid end BRE] ${f}b $re \ $target] return } set f [flags $flags] set infoflags [infoflags $flags] set ccmd [concat [list testregexp -$ask] $f [list $re]] set nsub [expr [llength $args] - 1] if {$nsub == -1} { # didn't tell us number of subexps set ccmd "lreplace \[$ccmd\] 0 0" set info [list $infoflags] } else { set info [list $nsub $infoflags] } lappend testid "compile" test $prefix.[tno $testid] [desc $testid] {testregexp} $ccmd $info set testid [lreplace $testid end end "execute"] set ecmd [concat [list testregexp] $f [list $re $target]] test $prefix.[tno $testid] [desc $testid] {testregexp} $ecmd 0 } # match expected, internal routine that does the work # parameters like the "real" routines except they don't have "opts", # which is a possibly-empty list of switches for the regexp match attempt # The ! flag is used to indicate expected match failure (for REG_EXPECT, # which wants argument testing even in the event of failure). proc matchexpected {opts testid flags re target args} { global prefix description ask regBug if {[info exists regBug] && $regBug} { # This will register as a skipped test test $prefix.[tno $testid] [desc $testid] knownBug {format 0} {1} return } # Tcl locale stuff doesn't do the ch/xy test fakery yet if {[string first "+" $flags] >= 0} { # This will register as a skipped test test $prefix.[tno $testid] [desc $testid] localeRegexp {} {} return } # if &, test as both BRE and ARE set amp [string first "&" $flags] if {$amp >= 0} { set f [string range $flags 0 [expr $amp - 1]] append f [string range $flags [expr $amp + 1] end] eval [concat [list matchexpected $opts \ [linsert $testid end ARE] ${f} $re $target] $args] eval [concat [list matchexpected $opts \ [linsert $testid end BRE] ${f}b $re $target] $args] return } set f [flags $flags] set infoflags [infoflags $flags] set ccmd [concat [list testregexp -$ask] $f [list $re]] set ecmd [concat [list testregexp] $opts $f [list $re $target]] set nsub [expr [llength $args] - 1] set names [list] set refs "" for {set i 0} {$i <= $nsub} {incr i} { if {$i == 0} { set name match } else { set name sub$i } lappend names $name append refs " \$$name" set $name "" } if {[string first "o" $flags] >= 0} { ;# REG_NOSUB kludge set nsub 0 ;# unsigned value cannot be -1 } if {[string first "t" $flags] >= 0} { ;# REG_EXPECT incr nsub -1 ;# the extra does not count } set ecmd [concat $ecmd $names] set erun "list \[$ecmd\] $refs" set retcode [list 1] if {[string first "!" $flags] >= 0} { set retcode [list 0] } set result [concat $retcode $args] set info [list $nsub $infoflags] lappend testid "compile" test $prefix.[tno $testid] [desc $testid] {testregexp} $ccmd $info set testid [lreplace $testid end end "execute"] test $prefix.[tno $testid] [desc $testid] {testregexp} $erun $result } # match expected (no missing, empty, or ambiguous submatches) # m testno flags re target mat submat ... proc m {args} { eval matchexpected [linsert $args 0 [list]] } # match expected (full fanciness) # i testno flags re target mat submat ... proc i {args} { eval matchexpected [linsert $args 0 [list "-indices"]] } # partial match expected # p testno flags re target mat "" ... # Quirk: number of ""s must be one more than number of subREs. proc p {args} { set f [lindex $args 1] ;# add ! flag set args [lreplace $args 1 1 "!$f"] eval matchexpected [linsert $args 0 [list "-indices"]] } # test is a knownBug proc knownBug {args} { set ::regBug 1 uplevel #0 $args set ::regBug 0 } # the tests themselves # support functions and preliminary misc. # This is sensitive to changes in message wording, but we really have to # test the code->message expansion at least once. test reg-0.1 "regexp error reporting" { list [catch {regexp (*) ign} msg] $msg } {1 {couldn't compile regular expression pattern: quantifier operand invalid}} doing 1 "basic sanity checks" m 1 & abc abc abc f 2 & abc def m 3 & abc xyabxabce abc doing 2 "invalid option combinations" e 1 qe a INVARG e 2 qa a INVARG e 3 qx a INVARG e 4 qn a INVARG e 5 ba a INVARG doing 3 "basic syntax" i 1 &NS "" a {0 -1} m 2 NS a| a a m 3 - a|b a a m 4 - a|b b b m 5 NS a||b b b m 6 & ab ab ab doing 4 "parentheses" m 1 - (a)e ae ae a m 2 o (a)e ae m 3 b {\(a\)b} ab ab a m 4 - a((b)c) abc abc bc b m 5 - a(b)(c) abc abc b c e 6 - a(b EPAREN e 7 b {a\(b} EPAREN # sigh, we blew it on the specs here... someday this will be fixed in POSIX, # but meanwhile, it's fixed in AREs m 8 eU a)b a)b a)b e 9 - a)b EPAREN e 10 b {a\)b} EPAREN m 11 P a(?:b)c abc abc e 12 e a(?:b)c BADRPT i 13 S a()b ab {0 1} {1 0} m 14 SP a(?:)b ab ab i 15 S a(|b)c ac {0 1} {1 0} m 16 S a(b|)c abc abc b doing 5 "simple one-char matching" # general case of brackets done later m 1 & a.b axb axb f 2 &n "a.b" "a\nb" m 3 & {a[bc]d} abd abd m 4 & {a[bc]d} acd acd f 5 & {a[bc]d} aed f 6 & {a[^bc]d} abd m 7 & {a[^bc]d} aed aed f 8 &p "a\[^bc]d" "a\nd" doing 6 "context-dependent syntax" # plus odds and ends e 1 - * BADRPT m 2 b * * * m 3 b {\(*\)} * * * e 4 - (*) BADRPT m 5 b ^* * * e 6 - ^* BADRPT f 7 & ^b ^b m 8 b x^ x^ x^ f 9 I x^ x m 10 n "\n^" "x\nb" "\n" f 11 bS {\(^b\)} ^b m 12 - (^b) b b b m 13 & {x$} x x m 14 bS {\(x$\)} x x x m 15 - {(x$)} x x x m 16 b {x$y} "x\$y" "x\$y" f 17 I {x$y} xy m 18 n "x\$\n" "x\n" "x\n" e 19 - + BADRPT e 20 - ? BADRPT doing 7 "simple quantifiers" m 1 &N a* aa aa i 2 &N a* b {0 -1} m 3 - a+ aa aa m 4 - a?b ab ab m 5 - a?b b b e 6 - ** BADRPT m 7 bN ** *** *** e 8 & a** BADRPT e 9 & a**b BADRPT e 10 & *** BADRPT e 11 - a++ BADRPT e 12 - a?+ BADRPT e 13 - a?* BADRPT e 14 - a+* BADRPT e 15 - a*+ BADRPT doing 8 "braces" m 1 NQ "a{0,1}" "" "" m 2 NQ "a{0,1}" ac a e 3 - "a{1,0}" BADBR e 4 - "a{1,2,3}" BADBR e 5 - "a{257}" BADBR e 6 - "a{1000}" BADBR e 7 - "a{1" EBRACE e 8 - "a{1n}" BADBR m 9 BS "a{b" "a\{b" "a\{b" m 10 BS "a{" "a\{" "a\{" m 11 bQ "a\\{0,1\\}b" cb b e 12 b "a\\{0,1" EBRACE e 13 - "a{0,1\\" BADBR m 14 Q "a{0}b" ab b m 15 Q "a{0,0}b" ab b m 16 Q "a{0,1}b" ab ab m 17 Q "a{0,2}b" b b m 18 Q "a{0,2}b" aab aab m 19 Q "a{0,}b" aab aab m 20 Q "a{1,1}b" aab ab m 21 Q "a{1,3}b" aaaab aaab f 22 Q "a{1,3}b" b m 23 Q "a{1,}b" aab aab f 24 Q "a{2,3}b" ab m 25 Q "a{2,3}b" aaaab aaab f 26 Q "a{2,}b" ab m 27 Q "a{2,}b" aaaab aaaab doing 9 "brackets" m 1 & {a[bc]} ac ac m 2 & {a[-]} a- a- m 3 & {a[[.-.]]} a- a- m 4 &L {a[[.zero.]]} a0 a0 m 5 &LM {a[[.zero.]-9]} a2 a2 m 6 &M {a[0-[.9.]]} a2 a2 m 7 &+L {a[[=x=]]} ax ax m 8 &+L {a[[=x=]]} ay ay f 9 &+L {a[[=x=]]} az e 10 & {a[0-[=x=]]} ERANGE m 11 &L {a[[:digit:]]} a0 a0 e 12 & {a[[:woopsie:]]} ECTYPE f 13 &L {a[[:digit:]]} ab e 14 & {a[0-[:digit:]]} ERANGE m 15 &LP {[[:<:]]a} a a m 16 &LP {a[[:>:]]} a a e 17 & {a[[..]]b} ECOLLATE e 18 & {a[[==]]b} ECOLLATE e 19 & {a[[::]]b} ECTYPE e 20 & {a[[.a} EBRACK e 21 & {a[[=a} EBRACK e 22 & {a[[:a} EBRACK e 23 & {a[} EBRACK e 24 & {a[b} EBRACK e 25 & {a[b-} EBRACK e 26 & {a[b-c} EBRACK m 27 &M {a[b-c]} ab ab m 28 & {a[b-b]} ab ab m 29 &M {a[1-2]} a2 a2 e 30 & {a[c-b]} ERANGE e 31 & {a[a-b-c]} ERANGE m 32 &M {a[--?]b} a?b a?b m 33 & {a[---]b} a-b a-b m 34 & {a[]b]c} a]c a]c m 35 EP {a[\]]b} a]b a]b f 36 bE {a[\]]b} a]b m 37 bE {a[\]]b} "a\\]b" "a\\]b" m 38 eE {a[\]]b} "a\\]b" "a\\]b" m 39 EP {a[\\]b} "a\\b" "a\\b" m 40 eE {a[\\]b} "a\\b" "a\\b" m 41 bE {a[\\]b} "a\\b" "a\\b" e 42 - {a[\Z]b} EESCAPE m 43 & {a[[b]c} "a\[c" "a\[c" m 44 EMP* {a[\u00fe-\u0507][\u00ff-\u0300]b} \ "a\u0102\u02ffb" "a\u0102\u02ffb" doing 10 "anchors and newlines" m 1 & ^a a a f 2 &^ ^a a i 3 &N ^ a {0 -1} i 4 & {a$} aba {2 2} f 5 {&$} {a$} a i 6 &N {$} ab {2 1} m 7 &n ^a a a m 8 &n "^a" "b\na" "a" i 9 &w "^a" "a\na" {0 0} i 10 &n^ "^a" "a\na" {2 2} m 11 &n {a$} a a m 12 &n "a\$" "a\nb" "a" i 13 &n "a\$" "a\na" {0 0} i 14 N ^^ a {0 -1} m 15 b ^^ ^ ^ i 16 N {$$} a {1 0} m 17 b {$$} "\$" "\$" m 18 &N {^$} "" "" f 19 &N {^$} a i 20 &nN "^\$" "a\n\nb" {2 1} m 21 N {$^} "" "" m 22 b {$^} "\$^" "\$^" m 23 P {\Aa} a a m 24 ^P {\Aa} a a f 25 ^nP {\Aa} "b\na" m 26 P {a\Z} a a m 27 {$P} {a\Z} a a f 28 {$nP} {a\Z} "a\nb" e 29 - ^* BADRPT e 30 - {$*} BADRPT e 31 - {\A*} BADRPT e 32 - {\Z*} BADRPT doing 11 "boundary constraints" m 1 &LP {[[:<:]]a} a a m 2 &LP {[[:<:]]a} -a a f 3 &LP {[[:<:]]a} ba m 4 &LP {a[[:>:]]} a a m 5 &LP {a[[:>:]]} a- a f 6 &LP {a[[:>:]]} ab m 7 bLP {\} a a f 10 bLP {a\>} ab m 11 LP {\ya} a a f 12 LP {\ya} ba m 13 LP {a\y} a a f 14 LP {a\y} ab m 15 LP {a\Y} ab a f 16 LP {a\Y} a- f 17 LP {a\Y} a f 18 LP {-\Y} -a m 19 LP {-\Y} -% - f 20 LP {\Y-} a- e 21 - {[[:<:]]*} BADRPT e 22 - {[[:>:]]*} BADRPT e 23 b {\<*} BADRPT e 24 b {\>*} BADRPT e 25 - {\y*} BADRPT e 26 - {\Y*} BADRPT m 27 LP {\ma} a a f 28 LP {\ma} ba m 29 LP {a\M} a a f 30 LP {a\M} ab f 31 ILP {\Ma} a f 32 ILP {a\m} a doing 12 "character classes" m 1 LP {a\db} a0b a0b f 2 LP {a\db} axb f 3 LP {a\Db} a0b m 4 LP {a\Db} axb axb m 5 LP "a\\sb" "a b" "a b" m 6 LP "a\\sb" "a\tb" "a\tb" m 7 LP "a\\sb" "a\nb" "a\nb" f 8 LP {a\sb} axb m 9 LP {a\Sb} axb axb f 10 LP "a\\Sb" "a b" m 11 LP {a\wb} axb axb f 12 LP {a\wb} a-b f 13 LP {a\Wb} axb m 14 LP {a\Wb} a-b a-b m 15 LP {\y\w+z\y} adze-guz guz m 16 LPE {a[\d]b} a1b a1b m 17 LPE "a\[\\s]b" "a b" "a b" m 18 LPE {a[\w]b} axb axb doing 13 "escapes" e 1 & "a\\" EESCAPE m 2 - {a\