Make getindex for String check if all indices are valid

Closes #22548

fixes a bug with use of prevind in dates/io.jl

NEWS.md

@@ -100,6 +100,9 @@ Breaking changes
 
 This section lists changes that do not have deprecation warnings.
 
+  * `getindex(s::String, r::UnitRange{Int})` now throws `UnicodeError` if `last(r)`
+    is not a valid index into `s` ([#22572]).
+
   * `ntuple(f, n::Integer)` throws `ArgumentError` if `n` is negative.
     Previously an empty tuple was returned ([#21697]).
 

base/dates/io.jl

@@ -330,7 +330,7 @@ function DateFormat(f::AbstractString, locale::DateLocale=ENGLISH)
 
     letters = String(collect(keys(CONVERSION_SPECIFIERS)))
     for m in eachmatch(Regex("(?<!\\\\)([\\Q$letters\\E])\\1*"), f)
-        tran = replace(f[prev_offset:m.offset - 1], r"\\(.)", s"\1")
+        tran = replace(f[prev_offset:prevind(f, m.offset)], r"\\(.)", s"\1")
 
         if !isempty(prev)
             letter, width = prev

base/repl/LineEdit.jl

@@ -317,7 +317,7 @@ function refresh_multi_line(termbuf::TerminalBuffer, terminal::UnixTerminal, buf
             # in this case, we haven't yet written the cursor position
             line_pos -= slength # '\n' gets an extra pos
             if line_pos < 0 || !moreinput
-                num_chars = (line_pos >= 0 ? llength : strwidth(l[1:(line_pos + slength)]))
+                num_chars = (line_pos >= 0 ? llength : strwidth(l[1:prevind(l, line_pos + slength + 1)]))
                 curs_row, curs_pos = divrem(lindent + num_chars - 1, cols)
                 curs_row += cur_row
                 curs_pos += 1

base/repl/REPL.jl

@@ -882,7 +882,7 @@ function setup_interface(
                     end
                     # Check if input line starts with "julia> ", remove it if we are in prompt paste mode
                     jl_prompt_len = 7
-                    if (firstline || isprompt_paste) && (oldpos + jl_prompt_len <= sizeof(input) && input[oldpos:oldpos+jl_prompt_len-1] == JULIA_PROMPT)
+                    if (firstline || isprompt_paste) && startswith(SubString(input, oldpos), JULIA_PROMPT)
                         isprompt_paste = true
                         oldpos += jl_prompt_len
                     # If we are prompt pasting and current statement does not begin with julia> , skip to next line

base/strings/string.jl

@@ -235,20 +235,25 @@ isvalid(s::String, i::Integer) =
 
 function getindex(s::String, r::UnitRange{Int})
     isempty(r) && return ""
-    i, j = first(r), last(r)
     l = sizeof(s)
+    i = first(r)
     if i < 1 || i > l
         throw(BoundsError(s, i))
     end
     @inbounds si = codeunit(s, i)
     if is_valid_continuation(si)
         throw(UnicodeError(UTF_ERR_INVALID_INDEX, i, si))
     end
+    j = last(r)
     if j > l
-        throw(BoundsError())
+        throw(BoundsError(s, j))
     end
-    j = nextind(s,j)-1
-    unsafe_string(pointer(s,i), j-i+1)
+    @inbounds sj = codeunit(s, j)
+    if is_valid_continuation(sj)
+        throw(UnicodeError(UTF_ERR_INVALID_INDEX, j, sj))
+    end
+    j = nextind(s,j)
+    unsafe_string(pointer(s,i), j-i)
 end
 
 function search(s::String, c::Char, i::Integer = 1)

test/strings/basic.jl

@@ -159,9 +159,10 @@ end
 @test first('\x00':'\x7f') === '\x00'
 @test last('\x00':'\x7f') === '\x7f'
 
-# make sure substrings handle last code unit even if not start of codepoint
+# make sure substrings do not accept code unit if it is not start of codepoint
 let s = "x\u0302"
-    @test s[1:3] == s
+    @test_throws UnicodeError s[1:3]
+    @test s[1:2]==s
 end
 
 # issue #9781