Support backcapture references in 's replacement string

Author: malmaud

base/exports.jl

@@ -1337,6 +1337,7 @@ export
     # notation for certain types
     @b_str,  # byte vector
     @r_str,  # regex
+    @s_str,  # regex substitution string
     @v_str,  # version number
 
     # documentation

base/pcre.jl

@@ -140,6 +140,23 @@ function substring_number_from_name(re, name)
         (Ptr{Void}, Cstring), re, name)
 end
 
+function substring_length_bynumber(match_data, number)
+  s = Ref{Csize_t}()
+  rc = ccall((:pcre2_substring_length_bynumber_8, PCRE_LIB), Cint,
+        (Ptr{Void}, UInt32, Ref{Csize_t}), match_data, number, s)
+  rc < 0 && error("PCRE error: $(err_message(rc))")
+  convert(Int, s[])
+end
+
+function substring_copy_bynumber(match_data, number, buf, buf_size)
+  s = Ref{Csize_t}(buf_size)
+  rc = ccall((:pcre2_substring_copy_bynumber_8, PCRE_LIB), Cint,
+             (Ptr{Void}, UInt32, Ptr{UInt8}, Ref{Csize_t}),
+             match_data, number, buf, s)
+  rc < 0 && error("PCRE error: $(err_message(rc))")
+  convert(Int, s[])
+end
+
 function capture_names(re)
     name_count = info(re, INFO_NAMECOUNT, UInt32)
     name_entry_size = info(re, INFO_NAMEENTRYSIZE, UInt32)

base/regex.jl

@@ -208,14 +208,86 @@ end
 search(s::AbstractString, r::Regex, idx::Integer) =
     throw(ArgumentError("regex search is only available for bytestrings; use bytestring(s) to convert"))
 search(s::AbstractString, r::Regex) = search(s,r,start(s))
-_search(s::AbstractString, r::Regex, idx::Integer) = match(r, s, idx)
 
-first(m::RegexMatch) = m.offset
-last(m::RegexMatch) = m.offset + length(m.match) - 1
+immutable SubstitutionString{T<:AbstractString} <: AbstractString
+    string::T
+end
+
+endof(s::SubstitutionString) = endof(s.string)
+next(s::SubstitutionString, idx::Int) = next(s.string, idx)
+function show(io::IO, s::SubstitutionString)
+    print(io, "s")
+    show(io, s.string)
+end
+
+macro s_str(string) SubstitutionString(string) end
+
+replace_err(repl) = error("Bad replacement string: $repl")
+
+function _write_capture(io, re, group)
+    len = PCRE.substring_length_bynumber(re.match_data, group)
+    ensureroom(io, len+1)
+    PCRE.substring_copy_bynumber(re.match_data, group,
+        pointer(io.data, io.ptr), len+1)
+    io.ptr += len
+end
 
-function _replace(io, repl::AbstractString, str, m::RegexMatch)
-    write(io, "repl")
-    write(io, repl)
+function _replace(io, repl_s::SubstitutionString, str, r, re)
+    const SUB_CHAR = '\\'
+    const GROUP_CHAR = 'g'
+    const LBRACKET = '<'
+    const RBRACKET = '>'
+    repl = repl_s.string
+    i = start(repl)
+    e = endof(repl)
+    while i <= e
+        if repl[i] == SUB_CHAR
+            next_i = nextind(repl, i)
+            next_i > e && replace_err(repl)
+            if repl[next_i] == SUB_CHAR
+                write(io, SUB_CHAR, SUB_CHAR)
+                i = nextind(repl, next_i)
+            elseif isnumber(repl[next_i])
+                group = parse(Int, repl[next_i])
+                i = nextind(repl, next_i)
+                while i <= e
+                    if isnumber(repl[i])
+                        group = 10group + parse(Int, repl[i])
+                        i = nextind(repl, i)
+                    else
+                        break
+                    end
+                end
+                _write_capture(io, re, group)
+            elseif repl[next_i] == GROUP_CHAR
+                i = nextind(repl, next_i)
+                if i > e || repl[i] != LBRACKET
+                    replace_err(repl)
+                end
+                i = nextind(repl, i)
+                i > e && replace_err(repl)
+                groupstart = i
+                while repl[i] != RBRACKET
+                    i = nextind(repl, i)
+                    i > e && replace_err(repl)
+                end
+                #groupname = repl[groupstart:prevind(repl, i)]
+                groupname = SubString(repl, groupstart, prevind(repl, i))
+                if isnumber(groupname)
+                    _write_capture(io, re, parse(Int, groupname))
+                else
+                    _write_capture(io, re,
+                        PCRE.substring_number_from_name(re.regex, groupname))
+                end
+                i = nextind(repl, i)
+            else
+                replace_err(repl)
+            end
+        else
+            write(io, repl[i])
+            i = nextind(repl, i)
+        end
+    end
 end
 
 immutable RegexMatchIterator

base/string.jl

@@ -1327,11 +1327,9 @@ function _rsplit{T<:AbstractString,U<:Array}(str::T, splitter, limit::Integer, k
 end
 #rsplit(str::AbstractString) = rsplit(str, _default_delims, 0, false)
 
-_replace(io, repl, str, r) = write(io, repl)
-_replace(io, repl::Function, str, r) =
+_replace(io, repl, str, r, pattern) = write(io, repl)
+_replace(io, repl::Function, str, r, pattern) =
     write(io, repl(SubString(str, first(r), last(r))))
-_search(str, pattern, offset) = search(str, pattern, offset)
-
 
 function replace(str::ByteString, pattern, repl, limit::Integer)
     n = 1
@@ -1340,10 +1338,11 @@ function replace(str::ByteString, pattern, repl, limit::Integer)
     r = search(str,pattern,i)
     j, k = first(r), last(r)
     out = IOBuffer()
+    ensureroom(out, floor(Int, 1.2sizeof(str)))
     while j != 0
         if i == a || i <= k
             write_sub(out, str.data, i, j-i)
-            _replace(out, repl, str, r)
+            _replace(out, repl, str, r, pattern)
         end
         if k<j
             i = j
@@ -1354,8 +1353,7 @@ function replace(str::ByteString, pattern, repl, limit::Integer)
         if j > e
             break
         end
-        r = _search(str,pattern,k)
-        r == nothing && break
+        r = search(str,pattern,k)
         j, k = first(r), last(r)
         n == limit && break
         n += 1

test/regex.jl

@@ -39,6 +39,10 @@ show(buf, r"")
 @test_throws ArgumentError search(utf32("this is a test"), r"test")
 
 # Named subpatterns
-m = match(r"(?<a>.)(.)(?<b>.)", "xyz")
-@test (m[:a], m[2], m["b"]) == ("x", "y", "z")
-@test sprint(show, m) == "RegexMatch(\"xyz\", a=\"x\", 2=\"y\", b=\"z\")"
+let m = match(r"(?<a>.)(.)(?<b>.)", "xyz")
+    @test (m[:a], m[2], m["b"]) == ("x", "y", "z")
+    @test sprint(show, m) == "RegexMatch(\"xyz\", a=\"x\", 2=\"y\", b=\"z\")"
+end
+
+# Backcapture reference in substitution string
+@test replace("abcde", r"(..)(?P<byname>d)", s"\g<byname>xy\1") == "adxybce"