xref: /aosp_15_r20/external/pcre/testdata/testinput11 (revision 22dc650d8ae982c6770746019a6f94af92b0f024)

# This set of tests is for the 16-bit and 32-bit libraries' basic (non-UTF)
# features that are not compatible with the 8-bit library, or which give
# different output in 16-bit or 32-bit mode. The output for the two widths is
# different, so they have separate output files.

#forbid_utf
#newline_default LF ANY ANYCRLF

/[^\x{c4}]/IB

/\x{100}/I

/  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*                          # optional leading comment
(?:    (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|
" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
)                    # initial word
(?:  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|
" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
)  )* # further okay, if led by a period
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  @  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*    (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                           # initial subdomain
(?:                                  #
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.                        # if led by a period...
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                     #   ...further okay
)*
# address
|                     #  or
(?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|
" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
)             # one word, optionally followed by....
(?:
[^()<>@,;:".\\\[\]\x80-\xff\000-\010\012-\037]  |  # atom and space parts, or...
\(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)       |  # comments, or...

" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
# quoted strings
)*
<  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*                     # leading <
(?:  @  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*    (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                           # initial subdomain
(?:                                  #
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.                        # if led by a period...
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                     #   ...further okay
)*

(?:  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  ,  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  @  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*    (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                           # initial subdomain
(?:                                  #
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.                        # if led by a period...
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                     #   ...further okay
)*
)* # further okay, if led by comma
:                                # closing colon
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  )? #       optional route
(?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|
" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
)                    # initial word
(?:  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|
" (?:                      # opening quote...
[^\\\x80-\xff\n\015"]                #   Anything except backslash and quote
|                     #    or
\\ [^\x80-\xff]           #   Escaped something (something != CR)
)* "  # closing quote
)  )* # further okay, if led by a period
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  @  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*    (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                           # initial subdomain
(?:                                  #
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  \.                        # if led by a period...
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*   (?:
[^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]+    # some number of atom characters...
(?![^(\040)<>@,;:".\\\[\]\000-\037\x80-\xff]) # ..not followed by something that could be part of an atom
|   \[                         # [
(?: [^\\\x80-\xff\n\015\[\]] |  \\ [^\x80-\xff]  )*    #    stuff
\]                        #           ]
)                     #   ...further okay
)*
#       address spec
(?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*  > #                  trailing >
# name and address
)  (?: [\040\t] |  \(
(?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  |  \( (?:  [^\\\x80-\xff\n\015()]  |  \\ [^\x80-\xff]  )* \)  )*
\)  )*                       # optional trailing comment
/Ix

/[\h]/B
    >\x09<

/[\h]+/B
    >\x09\x20\xa0<

/[\v]/B

/[^\h]/B

/\h+/I
    \x{1681}\x{200b}\x{1680}\x{2000}\x{202f}\x{3000}
    \x{3001}\x{2fff}\x{200a}\xa0\x{2000}

/[\h\x{dc00}]+/IB
    \x{1681}\x{200b}\x{1680}\x{2000}\x{202f}\x{3000}
    \x{3001}\x{2fff}\x{200a}\xa0\x{2000}

/\H+/I
    \x{1680}\x{180e}\x{167f}\x{1681}\x{180d}\x{180f}
    \x{2000}\x{200a}\x{1fff}\x{200b}
    \x{202f}\x{205f}\x{202e}\x{2030}\x{205e}\x{2060}
    \xa0\x{3000}\x9f\xa1\x{2fff}\x{3001}

/[\H\x{d800}]+/
    \x{1680}\x{180e}\x{167f}\x{1681}\x{180d}\x{180f}
    \x{2000}\x{200a}\x{1fff}\x{200b}
    \x{202f}\x{205f}\x{202e}\x{2030}\x{205e}\x{2060}
    \xa0\x{3000}\x9f\xa1\x{2fff}\x{3001}

/\v+/I
    \x{2027}\x{2030}\x{2028}\x{2029}
    \x09\x0e\x84\x86\x85\x0a\x0b\x0c\x0d

/[\v\x{dc00}]+/IB
    \x{2027}\x{2030}\x{2028}\x{2029}
    \x09\x0e\x84\x86\x85\x0a\x0b\x0c\x0d

/\V+/I
    \x{2028}\x{2029}\x{2027}\x{2030}
    \x85\x0a\x0b\x0c\x0d\x09\x0e\x84\x86

/[\V\x{d800}]+/
    \x{2028}\x{2029}\x{2027}\x{2030}
    \x85\x0a\x0b\x0c\x0d\x09\x0e\x84\x86

/\R+/I,bsr=unicode
    \x{2027}\x{2030}\x{2028}\x{2029}
    \x09\x0e\x84\x86\x85\x0a\x0b\x0c\x0d

/\x{d800}\x{d7ff}\x{dc00}\x{dc00}\x{dcff}\x{dd00}/I
    \x{d800}\x{d7ff}\x{dc00}\x{dc00}\x{dcff}\x{dd00}

/[^\x{80}][^\x{ff}][^\x{100}][^\x{1000}][^\x{ffff}]/B

/[^\x{80}][^\x{ff}][^\x{100}][^\x{1000}][^\x{ffff}]/Bi

/[^\x{100}]*[^\x{1000}]+[^\x{ffff}]??[^\x{8000}]{4,}[^\x{7fff}]{2,9}?[^\x{100}]{5,6}+/B

/[^\x{100}]*[^\x{1000}]+[^\x{ffff}]??[^\x{8000}]{4,}[^\x{7fff}]{2,9}?[^\x{100}]{5,6}+/Bi

/(*:0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF)XX/mark
    XX

/(*:0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDE)XX/mark
    XX

/\u0100/B,alt_bsux,allow_empty_class,match_unset_backref

/[\u0100-\u0200]/B,alt_bsux,allow_empty_class,match_unset_backref

/\ud800/B,alt_bsux,allow_empty_class,match_unset_backref

/^\x{ffff}+/i
    \x{ffff}

/^\x{ffff}?/i
    \x{ffff}

/^\x{ffff}*/i
    \x{ffff}

/^\x{ffff}{3}/i
    \x{ffff}\x{ffff}\x{ffff}

/^\x{ffff}{0,3}/i
    \x{ffff}

/[^\x00-a]{12,}[^b-\xff]*/B

/[^\s]*\s* [^\W]+\W+ [^\d]*?\d0 [^\d\w]{4,6}?\w*A/B

/a*[b-\x{200}]?a#a*[b-\x{200}]?b#[a-f]*[g-\x{200}]*#[g-\x{200}]*[a-c]*#[g-\x{200}]*[a-h]*/B

/^[\x{1234}\x{4321}]{2,4}?/
    \x{1234}\x{1234}\x{1234}

# Check maximum non-UTF character size for the 16-bit library.

/\x{ffff}/
    A\x{ffff}B

/\x{10000}/

/\o{20000}/

# Check maximum character size for the 32-bit library. These will all give
# errors in the 16-bit library.

/\x{110000}/

/\x{7fffffff}/

/\x{80000000}/

/\x{ffffffff}/

/\x{100000000}/

/\o{17777777777}/

/\o{20000000000}/

/\o{37777777777}/

/\o{40000000000}/

/\x{7fffffff}\x{7fffffff}/I

/\x{80000000}\x{80000000}/I

/\x{ffffffff}\x{ffffffff}/I

# Non-UTF characters

/.{2,3}/
    \x{400000}\x{400001}\x{400002}\x{400003}

/\x{400000}\x{800000}/IBi

# Check character ranges

/[\H]/IB

/[\V]/IB

/(*THEN:\[A]{65501})/expand

# We can use pcre2test's utf8_input modifier to create wide pattern characters,
# even though this test is run when UTF is not supported.

/ab������z/utf8_input
    ab������z
    ab\x{7fffffff}z

/ab�������z/utf8_input
    ab�������z
    ab\x{ffffffff}z

/ab�Az/utf8_input
    ab�Az
    ab\x{80000041}z

/(?i:A{1,}\6666666666)/
    A\x{1b6}6666666

# End of testinput11