This is a module of instances of RegexContext (defined in Text.Regex.Base.RegexLike). Nothing else is exported. This is usually imported via the Text.Regex.Base convenience package which itself is re-exported from newer Text.Regex.XXX modules provided by the different regex-xxx backends.

These instances work for all the supported types and backends interchangeably. These instances provide the different results that can be gotten from a match or matchM operation (often via the =~ and =~~ operators with combine makeRegex with match and matchM respectively). This module name is Context because they operators are context dependent: use them in a context that expects an Int and you get a count of matches, use them in a Bool context and get True if there is a match, etc.

RegexContext a b c takes a regular expression suppied in a type a generated by RegexMaker and a target text supplied in type b to a result type c using the match class function. The matchM class function works like match unless there is no match found, in which case it calls fail in the (arbitrary) monad context.

There are a few type synonyms from Text.Regex.Base.RegexLike that are used here:

-- | 0 based index from start of source, or (-1) for unused
type MatchOffset = Int
-- | non-negative length of a match
type MatchLength = Int
type MatchArray = Array Int (MatchOffset, MatchLength)
type MatchText source = Array Int (source, (MatchOffset, MatchLength))

There are also a few newtypes that used to prevent any possible overlap of types, which were not needed for GHC's late overlap detection but are needed for use in Hugs.

newtype AllSubmatches     f b = AllSubmatches     { getAllSubmatches     :: f b }
newtype AllTextSubmatches f b = AllTextSubmatches { getAllTextSubmatches :: f b }
newtype AllMatches        f b = AllMatches        { getAllMatches        :: f b }
newtype AllTextMatches    f b = AllTextMatches    { getAllTextMatches    :: f b }

The newtypes' f parameters are the containers, usually [] or Array Int, (where the arrays all have lower bound 0).

The two Submatches newtypes return only information on the first match. The other two newtypes return information on all the non-overlapping matches. The two Text newtypes are used to mark result types that contain the same type as the target text.

Where provided, noncaptured submatches will have a MatchOffset of (-1) and non-negative otherwise. The semantics of submatches depend on the backend and its compile and execution options. Where provided, MatchLength will always be non-negative. Arrays with no elements are returned with bounds of (1,0). Arrays with elements will have a lower bound of 0.

XXX THIS HADDOCK DOCUMENTATION IS OUT OF DATE XXX

These are for finding the first match in the target text:

RegexContext a b Bool : Whether there is any match or not.

RegexContext a b () : Useful as a guard with matchM or =~~ in a monad, since failure to match calls fail.

RegexContext a b b : This returns the text of the whole match. It will return empty from the Extract type class if there is no match. These are defined in each backend module, but documented here for convenience.

RegexContext a b (MatchOffset, MatchLength) : This returns the initial index and length of the whole match. MatchLength will always be non-negative, and 0 for a failed match.

RegexContext a b (MatchResult b) : The MatchResult structure with details for the match. This is the structure copied from the old JRegex pacakge.

RegexContext a b (b, b, b) : The text before the match, the text of the match, the text after the match

RegexContext a b (b, MatchText b, b) : The text before the match, the details of the match, and the text after the match

RegexContext a b (b, b, b, [b]) : The text before the match, the text of the match, the text after the match, and a list of the text of the 1st and higher sub-parts of the match. This is the same return value as used in the old Text.Regex API.

Two containers of the submatch offset information:

RegexContext a b MatchArray : Array of (MatchOffset, MatchLength) for all the sub matches. The whole match is at the intial 0th index. Noncaptured submatches will have a MatchOffset of (-1) The array will have no elements and bounds (1,0) if there is no match.

RegexContext a b (AllSubmatches [] (MatchOffset, MatchLength) : List of (MatchOffset, MatchLength) The whole match is the first element, the rest are the submatches (if any) in order. The list is empty if there is no match.

Two containers of the submatch text and offset information:

 RegexContext a b (AllTextSubmatches (Array Int) (b, (MatchOffset, MatchLength)))

 RegexContext a b (AllTextSubmatches [] (b, (MatchOffset, MatchLength)))

Two containers of the submatch text information:

 RegexContext a b (AllTextSubmatches [] b)

 RegexContext a b (AllTextSubmatches (Array Int) b)

These instances are for all the matches (non-overlapping). Note that backends are supposed to supply RegexLike instances for which the default matchAll and matchAllText stop searching after returning any successful but empty match.

RegexContext a b Int : The number of matches, non-negative.

Two containers for locations of all matches:

 RegexContext a b (AllMatches [] (MatchOffset, MatchLength))

 RegexContext a b (AllMatches (Array Int) (MatchOffset, MatchLength))

Two containers for the locations of all matches and their submatches:

 RegexContext a b [MatchArray]

 RegexContext a b (AllMatches (Array Int) MatchArray)

Two containers for the text and locations of all matches and their submatches:

 RegexContext a b [MatchText b]

 RegexContext a b (AllTextMatches (Array Int) (MatchText b))

Two containers for text of all matches: RegexContext a b (AllTextMatches [] b)

 RegexContext a b (AllTextMatches (Array Int) b)

Four containers for text of all matches and their submatches:

 RegexContext a b [[b]]

 RegexContext a b (AllTextMatches (Array Int) [b])

 RegexContext a b (AllTextMatches [] (Array Int b))

 RegexContext a b (AllTextMatches (Array Int) (Array Int b))

Unused matches are empty (defined via Extract)