root/ncdu-zig
1
0
Fork 0
mirror of https://code.blicky.net/yorhel/ncdu.git synced 2026-01-15 02:08:40 -09:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix panic when shortening strings with unicode variation selectors

Browse source 
Fixes #199.
That's not to say it handles variation selectors or combining marks
well, though. This is kind of messy. :(
This commit is contained in:
Yorhel 2022-04-16 20:02:15 +02:00
parent 2e4f0f0bce
commit 0fc14173f2
1 changed files with 8 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
src/ui.zig
View file

@ -128,20 +128,21 @@ pub fn shorten(in: [:0]const u8, max_width: u32) [:0] const u8 {
var total_width: u32 = 0; var total_width: u32 = 0;
var prefix_width: u32 = 0; var prefix_width: u32 = 0;
var prefix_end: u32 = 0; var prefix_end: u32 = 0;
var prefix_done = false;
var it = std.unicode.Utf8View.initUnchecked(in).iterator(); var it = std.unicode.Utf8View.initUnchecked(in).iterator();
while (it.nextCodepoint()) |cp| { while (it.nextCodepoint()) |cp| {
// XXX: libc assumption: wchar_t is a Unicode point. True for most modern libcs? // XXX: libc assumption: wchar_t is a Unicode point. True for most modern libcs?
// (The "proper" way is to use mbtowc(), but I'd rather port the musl wcwidth implementation to Zig so that I *know* it'll be Unicode. // (The "proper" way is to use mbtowc(), but I'd rather port the musl wcwidth implementation to Zig so that I *know* it'll be Unicode.
// On the other hand, ncurses also use wcwidth() so that would cause duplicated code. Ugh) // On the other hand, ncurses also use wcwidth() so that would cause duplicated code. Ugh)
const cp_width_ = c.wcwidth(cp); const cp_width_ = c.wcwidth(cp);
const cp_width = @intCast(u32, if (cp_width_ < 0) 1 else cp_width_); const cp_width = @intCast(u32, if (cp_width_ < 0) 0 else cp_width_);
const cp_len = std.unicode.utf8CodepointSequenceLength(cp) catch unreachable; const cp_len = std.unicode.utf8CodepointSequenceLength(cp) catch unreachable;
total_width += cp_width; total_width += cp_width;
if (prefix_width + cp_width <= @divFloor(max_width-1, 2)-1) { if (!prefix_done and prefix_width + cp_width <= @divFloor(max_width-1, 2)-1) {
prefix_width += cp_width; prefix_width += cp_width;
prefix_end += cp_len; prefix_end += cp_len;
continue; } else
} prefix_done = true;
} }
if (total_width <= max_width) return in; if (total_width <= max_width) return in;
@ -154,7 +155,7 @@ pub fn shorten(in: [:0]const u8, max_width: u32) [:0] const u8 {
it = std.unicode.Utf8View.initUnchecked(in[prefix_end..]).iterator(); it = std.unicode.Utf8View.initUnchecked(in[prefix_end..]).iterator();
while (it.nextCodepoint()) |cp| { while (it.nextCodepoint()) |cp| {
const cp_width_ = c.wcwidth(cp); const cp_width_ = c.wcwidth(cp);
const cp_width = @intCast(u32, if (cp_width_ < 0) 1 else cp_width_); const cp_width = @intCast(u32, if (cp_width_ < 0) 0 else cp_width_);
const cp_len = std.unicode.utf8CodepointSequenceLength(cp) catch unreachable; const cp_len = std.unicode.utf8CodepointSequenceLength(cp) catch unreachable;
start_width += cp_width; start_width += cp_width;
start_len += cp_len; start_len += cp_len;
@ -185,6 +186,8 @@ test "shorten" {
try t("ＡaＢＣＤＥＦＧＨ", 8, "Ａ...Ｈ"); // could optimize this, but w/e try t("ＡaＢＣＤＥＦＧＨ", 8, "Ａ...Ｈ"); // could optimize this, but w/e
try t("ＡＢＣＤＥＦＧaＨ", 8, "Ａ...aＨ"); try t("ＡＢＣＤＥＦＧaＨ", 8, "Ａ...aＨ");
try t("ＡＢＣＤＥＦＧＨ", 15, "ＡＢＣ...ＦＧＨ"); try t("ＡＢＣＤＥＦＧＨ", 15, "ＡＢＣ...ＦＧＨ");
try t("❤︎a❤︎a❤︎a", 5, "❤︎...︎a"); // Variation selectors; not great, there's an additional U+FE0E before 'a'.
try t("ą́ą́ą́ą́ą́ą́", 5, "ą́...̨́ą́"); // Combining marks, similarly bad.
} }
// ncurses_refs.c // ncurses_refs.c