Merge branch 'lukeshu/9p-logging'HEADmain

1 files changed, 159 insertions, 0 deletions

diff --git a/libfmt/quote.c b/libfmt/quote.c
new file mode 100644
index 0000000..c91e0b0
--- /dev/null
+++ b/libfmt/quote.c
@@ -0,0 +1,159 @@
+/* libfmt/quote.c - C-string quoting for pico-fmt
+ *
+ * Copyright (C) 2025  Luke T. Shumaker <lukeshu@lukeshu.com>
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+#include <string.h> /* for strnlen() */
+#include <stdint.h> /* for uint{n}_t() */
+
+#include <libfmt/fmt.h>
+
+enum quote {
+	QUOTE_NONE,   /* c */
+	QUOTE_SIMPLE, /* \c */
+	QUOTE_U4,     /* \uABCD */
+	QUOTE_U8,     /* \UABCDABCD */
+};
+
+static inline enum quote needs_quote(uint32_t ch) {
+	if (ch == '\a' ||
+	    ch == '\b' ||
+	    ch == '\f' ||
+	    ch == '\n' ||
+	    ch == '\r' ||
+	    ch == '\t' ||
+	    ch == '\v' ||
+	    ch == '\\' ||
+	    ch == '\'' ||
+	    ch == '"' ||
+	    ch == '?')
+		return QUOTE_SIMPLE;
+	else if (' ' <= ch && ch <= '~')
+		return QUOTE_NONE;
+	else if (ch < 0x10000)
+		return QUOTE_U4;
+	else
+		return QUOTE_U8;
+}
+
+/**
+ * Quote a string to ASCII-only C syntax.  Valid UTF-8 is quoted as
+ * short C-escape characters, \uABCD or \UABCDABCD; invalid UTF-8 is
+ * quoted as \xAB.
+ */
+static void libfmt_conv_quote(struct fmt_state *state) {
+    uint32_t ch;
+    uint8_t chlen;
+
+    const char *in = va_arg(*state->args, char*);
+    size_t in_len = strnlen(in, (state->flags & FMT_FLAG_PRECISION) ? state->precision : (size_t)-1);
+
+    size_t out_len = 2;
+    for (size_t pos = 0; pos < in_len;) {
+	    if      ((in[pos] & 0b10000000) == 0b00000000) { ch = in[pos] & 0b01111111; chlen = 1; }
+	    else if ((in[pos] & 0b11100000) == 0b11000000) { ch = in[pos] & 0b00011111; chlen = 2; }
+	    else if ((in[pos] & 0b11110000) == 0b11100000) { ch = in[pos] & 0b00001111; chlen = 3; }
+	    else if ((in[pos] & 0b11111000) == 0b11110000) { ch = in[pos] & 0b00000111; chlen = 4; }
+	    else goto measure_invalid_utf8;
+	    if ((ch == 0 && chlen != 1) || pos + chlen > in_len) goto measure_invalid_utf8;
+	    for (uint8_t i = 1; i < chlen; i++) {
+		    if ((in[pos+i] & 0b11000000) != 0b10000000) goto measure_invalid_utf8;
+		    ch = (ch << 6) | (in[pos+i] & 0b00111111);
+	    }
+	    if (ch > 0x10FFFF) goto measure_invalid_utf8;
+	    pos += chlen;
+
+	    switch (needs_quote(ch)) {
+	    case QUOTE_NONE   : out_len +=  1; break;
+	    case QUOTE_SIMPLE : out_len +=  2; break;
+	    case QUOTE_U4     : out_len +=  6; break;
+	    case QUOTE_U8     : out_len += 10; break;
+	    }
+	    continue;
+    measure_invalid_utf8:
+	    pos++;
+	    out_len += 4; /* \xAB */
+    }
+
+    if (!(state->flags & FMT_FLAG_LEFT)) {
+	    for (size_t i = 0; i + out_len < state->width; i++) {
+		    fmt_state_putchar(state, ' ');
+	    }
+    }
+
+    fmt_state_putchar(state, '"');
+    for (size_t pos = 0; pos < in_len;) {
+	    if      ((in[pos] & 0b10000000) == 0b00000000) { ch = in[pos] & 0b01111111; chlen = 1; }
+	    else if ((in[pos] & 0b11100000) == 0b11000000) { ch = in[pos] & 0b00011111; chlen = 2; }
+	    else if ((in[pos] & 0b11110000) == 0b11100000) { ch = in[pos] & 0b00001111; chlen = 3; }
+	    else if ((in[pos] & 0b11111000) == 0b11110000) { ch = in[pos] & 0b00000111; chlen = 4; }
+	    else goto output_invalid_utf8;
+	    if ((ch == 0 && chlen != 1) || pos + chlen > in_len) goto output_invalid_utf8;
+	    for (uint8_t i = 1; i < chlen; i++) {
+		    if ((in[pos+i] & 0b11000000) != 0b10000000) goto output_invalid_utf8;
+		    ch = (ch << 6) | (in[pos+i] & 0b00111111);
+	    }
+	    if (ch > 0x10FFFF) goto output_invalid_utf8;
+	    pos += chlen;
+
+	    switch (needs_quote(ch)) {
+	    case QUOTE_NONE:
+		    fmt_state_putchar(state, ch);
+		    break;
+	    case QUOTE_SIMPLE:
+		    fmt_state_putchar(state, '\\');
+		    switch (ch) {
+		    case '\a': fmt_state_putchar(state, 'a'); break;
+		    case '\b': fmt_state_putchar(state, 'b'); break;
+		    case '\f': fmt_state_putchar(state, 'f'); break;
+		    case '\n': fmt_state_putchar(state, 'n'); break;
+		    case '\r': fmt_state_putchar(state, 'r'); break;
+		    case '\t': fmt_state_putchar(state, 't'); break;
+		    case '\v': fmt_state_putchar(state, 'v'); break;
+		    case '\\': fmt_state_putchar(state, '\\'); break;
+		    case '\'': fmt_state_putchar(state, '\''); break;
+		    case '"': fmt_state_putchar(state, '"'); break;
+		    case '?': fmt_state_putchar(state, '?'); break;
+		    }
+		    break;
+	    case QUOTE_U4:
+		    fmt_state_putchar(state, '\\');
+		    fmt_state_putchar(state, 'u');
+		    fmt_state_putchar(state, (ch >> 12) & 0xF);
+		    fmt_state_putchar(state, (ch >>  8) & 0xF);
+		    fmt_state_putchar(state, (ch >>  4) & 0xF);
+		    fmt_state_putchar(state, (ch >>  0) & 0xF);
+		    break;
+	    case QUOTE_U8:
+		    fmt_state_putchar(state, '\\');
+		    fmt_state_putchar(state, 'U');
+		    fmt_state_putchar(state, (ch >> 28) & 0xF);
+		    fmt_state_putchar(state, (ch >> 24) & 0xF);
+		    fmt_state_putchar(state, (ch >> 20) & 0xF);
+		    fmt_state_putchar(state, (ch >> 16) & 0xF);
+		    fmt_state_putchar(state, (ch >> 12) & 0xF);
+		    fmt_state_putchar(state, (ch >>  8) & 0xF);
+		    fmt_state_putchar(state, (ch >>  4) & 0xF);
+		    fmt_state_putchar(state, (ch >>  0) & 0xF);
+		    break;
+	    }
+	    continue;
+    output_invalid_utf8:
+	    fmt_state_putchar(state, '\\');
+	    fmt_state_putchar(state, 'x');
+	    fmt_state_putchar(state, (in[pos] >> 4) & 0xF);
+	    fmt_state_putchar(state, (in[pos] >> 0) & 0xF);
+	    pos++;
+    }
+    fmt_state_putchar(state, '"');
+
+    for (size_t i = 0; i + out_len < state->width; i++) {
+	    fmt_state_putchar(state, ' ');
+    }
+}
+
+[[gnu::constructor]]
+static void libfmt_install_quote(void) {
+	fmt_install('q', libfmt_conv_quote);
+}