language-servers/roc
1
0
Fork 0
mirror of https://github.com/roc-lang/roc.git synced 2025-09-27 13:59:08 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 6

thrashing

Browse source
This commit is contained in:
Dan Knutson 2021-10-24 11:35:45 -05:00
parent 46365da73a
commit 13fd72f3d8
2 changed files with 21 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

17
compiler/builtins/bitcode/src/str.zig
View file

@ -1511,11 +1511,13 @@ pub fn strTrim(string: RocStr) callconv(.C) RocStr {
return RocStr.empty(); return RocStr.empty();
} }
// SIGSEGV is after this
const leading_bytes = countLeadingWhitespaceBytes(string); const leading_bytes = countLeadingWhitespaceBytes(string);
const trailing_bytes = countTrailingWhitespaceBytes(string); const trailing_bytes = countTrailingWhitespaceBytes(string);
const new_len = string.len() - leading_bytes - trailing_bytes; const new_len = string.len() - leading_bytes - trailing_bytes;
if (new_len == 0) { if (new_len <= 0) {
return RocStr.empty(); return RocStr.empty();
} }
@ -1525,8 +1527,8 @@ pub fn strTrim(string: RocStr) callconv(.C) RocStr {
// could also just inline the unsafe reallocate call // could also just inline the unsafe reallocate call
// SIGSEGV is not from this branch // SIGSEGV is not from this branch
if (string.isRefcountOne()) { if (string.isRefcountOne() and !string.isSmallStr()) {
const dest = string.str_bytes orelse unreachable; const dest = string.str_bytes orelse return RocStr.empty();
const source = dest + leading_bytes; const source = dest + leading_bytes;
@memcpy(dest, source, new_len); @memcpy(dest, source, new_len);
return string.reallocate(new_len); return string.reallocate(new_len);
@ -1542,7 +1544,7 @@ fn countLeadingWhitespaceBytes(string: RocStr) usize {
var iter = unicode.Utf8View.initUnchecked(bytes).iterator(); var iter = unicode.Utf8View.initUnchecked(bytes).iterator();
while (iter.nextCodepoint()) |codepoint| { while (iter.nextCodepoint()) |codepoint| {
if (isWhitespace(codepoint)) { if (isWhitespace(codepoint)) {
byte_count += unicode.utf8CodepointSequenceLength(codepoint) catch unreachable; byte_count += unicode.utf8CodepointSequenceLength(codepoint) catch break;
} else { } else {
break; break;
} }
@ -1558,7 +1560,7 @@ fn countTrailingWhitespaceBytes(string: RocStr) usize {
var iter = ReverseUtf8View.initUnchecked(bytes).iterator(); var iter = ReverseUtf8View.initUnchecked(bytes).iterator();
while (iter.nextCodepoint()) |codepoint| { while (iter.nextCodepoint()) |codepoint| {
if (isWhitespace(codepoint)) { if (isWhitespace(codepoint)) {
byte_count += unicode.utf8CodepointSequenceLength(codepoint) catch unreachable; byte_count += unicode.utf8CodepointSequenceLength(codepoint) catch break;
} else { } else {
break; break;
} }
@ -1598,6 +1600,11 @@ const ReverseUtf8Iterator = struct {
it.i -= 1; it.i -= 1;
} }
// TODO this should be unnecessary; it means invalid utf8
if (it.i < 0) {
return null;
}
const cp_len = unicode.utf8ByteSequenceLength(it.bytes[it.i]) catch unreachable; const cp_len = unicode.utf8ByteSequenceLength(it.bytes[it.i]) catch unreachable;
const slice = it.bytes[it.i .. it.i + cp_len]; const slice = it.bytes[it.i .. it.i + cp_len];
it.i -= 1; it.i -= 1;

9
compiler/test_gen/src/gen_str.rs
View file

@ -1005,3 +1005,12 @@ fn str_trim_hello_world() {
RocStr RocStr
); );
} }
#[test]
fn str_trim_hello_world_both_large() {
assert_evals_to!(
indoc!(r#"Str.trim " hello world world ""#),
RocStr::from("hello world world"),
RocStr
);
}