mirror of
https://github.com/storytold/cadcraft.git
synced 2026-11-20 12:55:43 -06:00
Compare commits
10
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
e5091e052e | ||
|
|
d267d6580e | ||
|
|
2974d74bdc | ||
|
|
007a24ef94 | ||
|
|
0dae816499 | ||
|
|
ed6377fc46 | ||
|
|
923f4174a6 | ||
|
|
173cdf6e8a | ||
|
|
78bcdce794 | ||
|
|
fd11b47915 |
Generated
+15
-2
@@ -746,6 +746,7 @@ version = "0.21.0"
|
||||
dependencies = [
|
||||
"cadcraft-geom",
|
||||
"craft-dxf",
|
||||
"craft-shx",
|
||||
"encoding_rs",
|
||||
"harfrust 0.14.0",
|
||||
"icu_segmenter",
|
||||
@@ -1173,6 +1174,11 @@ name = "craft-codepage"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=0325b2d73676d736e94097dbcef01219cf25c7db#0325b2d73676d736e94097dbcef01219cf25c7db"
|
||||
|
||||
[[package]]
|
||||
name = "craft-codepage"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=10f0fac31c3c7f063de8c359b39d247451324b48#10f0fac31c3c7f063de8c359b39d247451324b48"
|
||||
|
||||
[[package]]
|
||||
name = "craft-docpath"
|
||||
version = "0.1.0"
|
||||
@@ -1190,9 +1196,11 @@ dependencies = [
|
||||
[[package]]
|
||||
name = "craft-dxf"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=006e4df225fca57e95ead0fb1f91920900798a99#006e4df225fca57e95ead0fb1f91920900798a99"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=10f0fac31c3c7f063de8c359b39d247451324b48#10f0fac31c3c7f063de8c359b39d247451324b48"
|
||||
dependencies = [
|
||||
"craft-codepage 0.1.0 (git+https://github.com/storytold/craft-libs?rev=10f0fac31c3c7f063de8c359b39d247451324b48)",
|
||||
"encoding_rs",
|
||||
"kurbo",
|
||||
"log",
|
||||
"thiserror 2.0.21",
|
||||
]
|
||||
@@ -1293,7 +1301,7 @@ version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=0325b2d73676d736e94097dbcef01219cf25c7db#0325b2d73676d736e94097dbcef01219cf25c7db"
|
||||
dependencies = [
|
||||
"chrono",
|
||||
"craft-codepage",
|
||||
"craft-codepage 0.1.0 (git+https://github.com/storytold/craft-libs?rev=0325b2d73676d736e94097dbcef01219cf25c7db)",
|
||||
"encoding_rs",
|
||||
]
|
||||
|
||||
@@ -1307,6 +1315,11 @@ dependencies = [
|
||||
"thiserror 2.0.21",
|
||||
]
|
||||
|
||||
[[package]]
|
||||
name = "craft-shx"
|
||||
version = "0.1.0"
|
||||
source = "git+https://github.com/storytold/craft-libs?rev=b3307f7814629961841b80ab907116252e41b34d#b3307f7814629961841b80ab907116252e41b34d"
|
||||
|
||||
[[package]]
|
||||
name = "craft-tiff"
|
||||
version = "0.1.0"
|
||||
|
||||
+2
-1
@@ -14,8 +14,9 @@ repository = "https://github.com/storytold/cadcraft"
|
||||
# internal
|
||||
cadcraft-geom = { path = "crates/geom" }
|
||||
# DXF tags: the shared reader and writer (moved from crates/dxf; craft-libs#791).
|
||||
craft-dxf = { git = "https://github.com/storytold/craft-libs", rev = "006e4df225fca57e95ead0fb1f91920900798a99" }
|
||||
craft-dxf = { git = "https://github.com/storytold/craft-libs", rev = "10f0fac31c3c7f063de8c359b39d247451324b48" }
|
||||
craft-dwg = { git = "https://github.com/storytold/craft-libs", rev = "7bb6c90076a0177d01cc4c482e7fd8dc9f31a8bf" }
|
||||
craft-shx = { git = "https://github.com/storytold/craft-libs", rev = "b3307f7814629961841b80ab907116252e41b34d" }
|
||||
craft-image-codecs = { git = "https://github.com/storytold/craft-libs", rev = "ae5020c1ed922cf7976a3a5545f05b07f466be12" }
|
||||
craft-mime = { git = "https://github.com/storytold/craft-libs", rev = "0325b2d73676d736e94097dbcef01219cf25c7db" }
|
||||
cadcraft-color = { path = "crates/color" }
|
||||
|
||||
@@ -11,6 +11,7 @@ workspace = true
|
||||
[dependencies]
|
||||
cadcraft-geom = { workspace = true }
|
||||
craft-dxf = { workspace = true }
|
||||
craft-shx = { workspace = true }
|
||||
skrifa = { workspace = true }
|
||||
harfrust = { workspace = true }
|
||||
unicode-bidi = { workspace = true }
|
||||
|
||||
+34
-89
@@ -71,8 +71,14 @@ pub fn plain_mtext(s: &str) -> String {
|
||||
}
|
||||
i = j + 1;
|
||||
}
|
||||
// An unknown code has no argument: drop the backslash and the code only.
|
||||
_ => i += 2,
|
||||
// A trailing lone backslash is dropped.
|
||||
_ if chars.get(i + 1).is_none() => i += 1,
|
||||
// An unknown code or a malformed `\U+` / `\M+` is shown as typed, as AutoCAD
|
||||
// shows it: keep the backslash and read on from the code.
|
||||
_ => {
|
||||
out.push('\\');
|
||||
i += 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
'{' | '}' | '\r' => i += 1,
|
||||
@@ -220,11 +226,8 @@ const MAX_ATOMS: usize = 200_000;
|
||||
/// Scale of stacked text relative to the surrounding text.
|
||||
const STACK_SCALE: f64 = 0.7;
|
||||
|
||||
/// Codes that take an argument terminated by `;` (besides `\S`, which has its own scan). Any
|
||||
/// other code stands alone, so a missing `;` must not swallow the text after it.
|
||||
pub(crate) fn takes_arg(code: char) -> bool {
|
||||
matches!(code, 'f' | 'F' | 'H' | 'W' | 'Q' | 'T' | 'A' | 'C' | 'c' | 'p')
|
||||
}
|
||||
// Shared DXF escape handling and parameterized-command classification.
|
||||
pub(crate) use craft_dxf::mtext::{takes_arg, unescape_codes};
|
||||
|
||||
/// The line-spacing factor to lay out with: AutoCAD's 0.25..=4 range, 1 for a missing or
|
||||
/// non-finite value.
|
||||
@@ -232,77 +235,6 @@ pub(crate) fn sane_line_spacing(v: f64) -> f64 {
|
||||
if v.is_finite() && v > 0.0 { v.clamp(0.25, 4.0) } else { 1.0 }
|
||||
}
|
||||
|
||||
/// Four hex digits at `chars[at..at + 4]`.
|
||||
fn hex4(chars: &[char], at: usize) -> Option<u32> {
|
||||
let digits = chars.get(at..at.checked_add(4)?)?;
|
||||
digits.iter().try_fold(0u32, |acc, c| Some(acc * 16 + c.to_digit(16)?))
|
||||
}
|
||||
|
||||
/// Push a decoded character so that the tokenizers read it as text, never as markup.
|
||||
fn push_literal(out: &mut String, c: char) {
|
||||
if matches!(c, '\\' | '{' | '}') {
|
||||
out.push('\\');
|
||||
}
|
||||
out.push(c);
|
||||
}
|
||||
|
||||
/// Decode `\U+hhhh` (one UTF-16 unit; a surrogate pair gives one character, a lone surrogate
|
||||
/// U+FFFD) and `\M+nhhhh` (double-byte East Asian characters) to the characters they stand
|
||||
/// for. Malformed escapes are left for the tokenizers, which drop the backslash and the code.
|
||||
/// An escaped backslash (`\\U+0041`) is not an escape.
|
||||
pub(crate) fn unescape_codes(s: &str) -> std::borrow::Cow<'_, str> {
|
||||
if !s.contains("\\U+") && !s.contains("\\M+") {
|
||||
return std::borrow::Cow::Borrowed(s);
|
||||
}
|
||||
let chars: Vec<char> = s.chars().collect();
|
||||
let mut out = String::with_capacity(s.len());
|
||||
let mut i = 0;
|
||||
while let Some(&c) = chars.get(i) {
|
||||
if c != '\\' {
|
||||
out.push(c);
|
||||
i += 1;
|
||||
continue;
|
||||
}
|
||||
let (n, plus) = (chars.get(i + 1).copied(), chars.get(i + 2).copied());
|
||||
if n == Some('U') && plus == Some('+') {
|
||||
if let Some(unit) = hex4(&chars, i + 3) {
|
||||
i += 7;
|
||||
let low = |at: usize| {
|
||||
(chars.get(at) == Some(&'\\') && chars.get(at + 1) == Some(&'U') && chars.get(at + 2) == Some(&'+'))
|
||||
.then(|| hex4(&chars, at + 3))
|
||||
.flatten()
|
||||
.filter(|u| (0xDC00..0xE000).contains(u))
|
||||
};
|
||||
let ch = if (0xD800..0xDC00).contains(&unit) {
|
||||
match low(i) {
|
||||
Some(lo) => {
|
||||
i += 7;
|
||||
char::from_u32(0x10000 + ((unit - 0xD800) << 10) + (lo - 0xDC00))
|
||||
}
|
||||
None => None,
|
||||
}
|
||||
} else {
|
||||
char::from_u32(unit)
|
||||
};
|
||||
push_literal(&mut out, ch.unwrap_or('\u{fffd}'));
|
||||
continue;
|
||||
}
|
||||
} else if n == Some('M') && plus == Some('+') && chars.get(i + 3).is_some_and(char::is_ascii_digit) && hex4(&chars, i + 4).is_some() {
|
||||
push_literal(&mut out, craft_dxf::mif_escape_at(&chars, i).unwrap_or('\u{fffd}'));
|
||||
i += 8;
|
||||
continue;
|
||||
}
|
||||
// Any other backslash pair stays as it is (and an escaped backslash stays one).
|
||||
out.push('\\');
|
||||
i += 1;
|
||||
if let Some(&next) = chars.get(i) {
|
||||
out.push(next);
|
||||
i += 1;
|
||||
}
|
||||
}
|
||||
std::borrow::Cow::Owned(out)
|
||||
}
|
||||
|
||||
fn arg(chars: &[char], from: usize) -> (String, usize) {
|
||||
let mut j = from;
|
||||
let mut a = String::new();
|
||||
@@ -518,8 +450,15 @@ fn parse(contents: &str, base: Fmt) -> (Vec<Atom>, Vec<Fmt>) {
|
||||
let (_, n) = arg(&chars, i);
|
||||
i = n;
|
||||
}
|
||||
// An unknown code has no argument: only the backslash and the code are dropped.
|
||||
_ => {}
|
||||
// An unknown code or a malformed `\U+` / `\M+` is shown as typed, as AutoCAD
|
||||
// shows it: keep the backslash and read on from the code. A trailing lone
|
||||
// backslash is dropped.
|
||||
_ => {
|
||||
if chars.get(i - 1).is_some() {
|
||||
push_char(&mut word, '\\', cur);
|
||||
i -= 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
'{' => {
|
||||
@@ -956,7 +895,7 @@ mod tests {
|
||||
assert_eq!(plain_mtext("\\H2.5x;Big"), "Big");
|
||||
assert_eq!(plain_mtext("unterminated \\H2"), "unterminated ");
|
||||
assert_eq!(plain_mtext("\\U+0041B \\U+00E8"), "AB è");
|
||||
assert_eq!(plain_mtext("\\U+12"), "+12", "a short escape loses only the backslash and the code");
|
||||
assert_eq!(plain_mtext("\\U+12"), "\\U+12", "a short escape is shown as typed");
|
||||
}
|
||||
|
||||
#[test]
|
||||
@@ -978,18 +917,24 @@ mod tests {
|
||||
assert_eq!(plain_mtext("\\\\U+0041"), "\\U+0041");
|
||||
// `\M+n` double-byte characters, decoded like the DXF reader does.
|
||||
assert_eq!(plain_mtext("\\M+5BCFE x;y"), "\u{4ef6} x;y");
|
||||
// Malformed escapes lose only the backslash and the code.
|
||||
assert_eq!(plain_mtext("a\\U+12 b"), "a+12 b");
|
||||
assert_eq!(lay("a\\U+12 b", 0.0).lines, vec!["a+12 b"]);
|
||||
assert_eq!(lay("\\pxql;a\\U+12 b", 0.0).lines, vec!["a+12 b"]);
|
||||
assert_eq!(plain_mtext("a\\M+ b"), "a+ b");
|
||||
// Malformed escapes are shown as typed, backslash included, as AutoCAD 2024 shows them.
|
||||
for s in ["a\\U+12 b", "\\U+12x", "a\\U+12;b", "a\\U+123 b", "a\\U+ b", "a\\U+00G1b", "a\\M+ b"] {
|
||||
assert_eq!(plain_mtext(s), s);
|
||||
assert_eq!(lay(s, 0.0).lines, vec![s]);
|
||||
assert_eq!(lay(&format!("\\pxql;{s}"), 0.0).lines, vec![s]);
|
||||
}
|
||||
// A valid escape takes exactly four digits.
|
||||
assert_eq!(plain_mtext("a\\U+12345b"), "a\u{1234}5b");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn unknown_code_without_semicolon_keeps_the_text() {
|
||||
assert_eq!(plain_mtext("A\\Zbc d"), "Abc d");
|
||||
assert_eq!(lay("A\\Zbc d", 0.0).lines, vec!["Abc d"]);
|
||||
assert_eq!(lay("\\pxql;A\\Zbc d", 0.0).lines, vec!["Abc d"]);
|
||||
// An unknown code is shown as typed, as AutoCAD 2024 shows it.
|
||||
assert_eq!(plain_mtext("A\\Zbc d"), "A\\Zbc d");
|
||||
assert_eq!(lay("A\\Zbc d", 0.0).lines, vec!["A\\Zbc d"]);
|
||||
assert_eq!(lay("\\pxql;A\\Zbc d", 0.0).lines, vec!["A\\Zbc d"]);
|
||||
// The text after it is read as usual: here a paragraph break.
|
||||
assert_eq!(plain_mtext("A\\Z\\Pb"), "A\\Z\nb");
|
||||
// A trailing lone backslash is dropped.
|
||||
assert_eq!(plain_mtext("A\\"), "A");
|
||||
// Codes with an argument still take it up to the `;`, or to the end.
|
||||
|
||||
@@ -308,8 +308,8 @@ impl Paragraphs {
|
||||
let (_, n) = arg(&chars, i);
|
||||
i = n;
|
||||
}
|
||||
// An unknown code has no argument: only the backslash and the code go.
|
||||
_ => {}
|
||||
// An unknown code is text, as typed: read on from the code, as the layout does.
|
||||
_ => i -= 1,
|
||||
}
|
||||
}
|
||||
'\n' => {
|
||||
|
||||
@@ -1,313 +0,0 @@
|
||||
//! Shape source (`.shp`) to a compiled SHX file, as COMPILE does.
|
||||
//!
|
||||
//! A source is a list of definitions: a header line `*number,defbytes,NAME` followed by its
|
||||
//! specification bytes (on as many lines as needed, ending with 0). Numbers are decimal, or
|
||||
//! hexadecimal after a leading `0` (`012`, `08cfb`); parentheses only group; `;` starts a
|
||||
//! comment. A big font starts with `*BIGFONT nchars,nranges,first,last,…` and a Unicode font
|
||||
//! with `*UNIFONT,6,name` followed by `above,below,modes,encoding,type,0`. Shape 0 is a text
|
||||
//! font's record (`above,below,modes,0`).
|
||||
//!
|
||||
//! Operands are encoded by what they are: a subshape number takes two bytes in a Unicode font
|
||||
//! (and after `7,0` in a big font), displacements and bulges are two's complement, and an arc's
|
||||
//! direction byte (`-043`) keeps its sign in the high bit. `defbytes` is not checked: the
|
||||
//! program's own length is written.
|
||||
|
||||
use super::ShxError;
|
||||
use super::parse::{BIGFONT, SHAPES_10, UNIFONT};
|
||||
|
||||
/// Largest source compiled.
|
||||
const MAX_SOURCE: usize = 8 << 20;
|
||||
|
||||
#[derive(Clone, Copy, PartialEq)]
|
||||
enum Kind {
|
||||
Shapes,
|
||||
Big,
|
||||
Uni,
|
||||
}
|
||||
|
||||
struct Shape {
|
||||
number: u16,
|
||||
name: String,
|
||||
program: Vec<u8>,
|
||||
}
|
||||
|
||||
/// A number token: value (sign included).
|
||||
fn number(tok: &str, line: usize) -> Result<i64, ShxError> {
|
||||
let t: String = tok.chars().filter(|c| !matches!(c, '(' | ')') && !c.is_whitespace()).collect();
|
||||
let (neg, digits) = match t.strip_prefix('-') {
|
||||
Some(d) => (true, d),
|
||||
None => (false, t.strip_prefix('+').unwrap_or(&t)),
|
||||
};
|
||||
let err = || ShxError::Source(line, format!("not a number: {tok:?}"));
|
||||
let v = if digits.len() > 1 && digits.starts_with('0') {
|
||||
i64::from_str_radix(digits.get(1..).unwrap_or_default(), 16).map_err(|_| err())?
|
||||
} else {
|
||||
digits.parse::<i64>().map_err(|_| err())?
|
||||
};
|
||||
if v > 0xFFFF {
|
||||
return Err(err());
|
||||
}
|
||||
Ok(if neg { -v } else { v })
|
||||
}
|
||||
|
||||
/// The comma-separated numbers of a line (comment removed, blank fields skipped).
|
||||
fn numbers(text: &str, line: usize) -> Result<Vec<i64>, ShxError> {
|
||||
text.split(',').map(str::trim).filter(|t| !t.is_empty() && *t != "(" && *t != ")").map(|t| number(t, line)).collect()
|
||||
}
|
||||
|
||||
/// The spec-byte encoder: walks the tokens as the reader walks the bytes.
|
||||
struct Enc<'a> {
|
||||
toks: &'a [i64],
|
||||
i: usize,
|
||||
out: Vec<u8>,
|
||||
kind: Kind,
|
||||
line: usize,
|
||||
}
|
||||
|
||||
impl Enc<'_> {
|
||||
fn next(&mut self) -> Result<i64, ShxError> {
|
||||
let v = self.toks.get(self.i).copied().ok_or_else(|| ShxError::Source(self.line, "the shape ends inside a command".into()))?;
|
||||
self.i += 1;
|
||||
Ok(v)
|
||||
}
|
||||
fn byte(&mut self) -> Result<(), ShxError> {
|
||||
let v = self.next()?;
|
||||
let b = u8::try_from(v).map_err(|_| ShxError::Source(self.line, format!("{v} is not a byte")))?;
|
||||
self.out.push(b);
|
||||
Ok(())
|
||||
}
|
||||
/// Two's complement −128..127.
|
||||
fn signed(&mut self) -> Result<i64, ShxError> {
|
||||
let v = self.next()?;
|
||||
let b = i8::try_from(v).map_err(|_| ShxError::Source(self.line, format!("{v} is not −128..127")))?;
|
||||
self.out.extend(b.to_le_bytes());
|
||||
Ok(v)
|
||||
}
|
||||
/// Sign in the high bit (an arc's direction byte).
|
||||
fn sign_magnitude(&mut self) -> Result<(), ShxError> {
|
||||
let v = self.next()?;
|
||||
let m = u8::try_from(v.unsigned_abs())
|
||||
.ok()
|
||||
.filter(|m| *m < 0x80)
|
||||
.ok_or_else(|| ShxError::Source(self.line, format!("{v} is not an octant byte")))?;
|
||||
self.out.push(if v < 0 { m | 0x80 } else { m });
|
||||
Ok(())
|
||||
}
|
||||
fn word(&mut self) -> Result<(), ShxError> {
|
||||
let v = self.next()?;
|
||||
let w = u16::try_from(v).map_err(|_| ShxError::Source(self.line, format!("{v} is not a shape number")))?;
|
||||
self.out.extend(w.to_be_bytes());
|
||||
Ok(())
|
||||
}
|
||||
|
||||
fn program(mut self) -> Result<Vec<u8>, ShxError> {
|
||||
while self.i < self.toks.len() {
|
||||
let code = self.next()?;
|
||||
let c = u8::try_from(code).map_err(|_| ShxError::Source(self.line, format!("{code} is not a code")))?;
|
||||
self.out.push(c);
|
||||
match c {
|
||||
0 => break,
|
||||
3 | 4 => self.byte()?,
|
||||
7 => match self.kind {
|
||||
Kind::Uni => self.word()?,
|
||||
Kind::Big => {
|
||||
let first = self.toks.get(self.i).copied().unwrap_or(0);
|
||||
if first == 0 {
|
||||
self.byte()?;
|
||||
self.word()?;
|
||||
for _ in 0..4 {
|
||||
self.byte()?;
|
||||
}
|
||||
} else {
|
||||
self.byte()?;
|
||||
}
|
||||
}
|
||||
Kind::Shapes => self.byte()?,
|
||||
},
|
||||
8 => {
|
||||
self.signed()?;
|
||||
self.signed()?;
|
||||
}
|
||||
9 => loop {
|
||||
let (x, y) = (self.signed()?, self.signed()?);
|
||||
if x == 0 && y == 0 {
|
||||
break;
|
||||
}
|
||||
},
|
||||
10 => {
|
||||
self.byte()?;
|
||||
self.sign_magnitude()?;
|
||||
}
|
||||
11 => {
|
||||
for _ in 0..4 {
|
||||
self.byte()?;
|
||||
}
|
||||
self.sign_magnitude()?;
|
||||
}
|
||||
12 => {
|
||||
for _ in 0..3 {
|
||||
self.signed()?;
|
||||
}
|
||||
}
|
||||
13 => loop {
|
||||
let (x, y) = (self.signed()?, self.signed()?);
|
||||
if x == 0 && y == 0 {
|
||||
break;
|
||||
}
|
||||
self.signed()?;
|
||||
},
|
||||
_ => {}
|
||||
}
|
||||
}
|
||||
if self.out.last() != Some(&0) {
|
||||
self.out.push(0);
|
||||
}
|
||||
Ok(self.out)
|
||||
}
|
||||
}
|
||||
|
||||
/// Compile shape source into an SHX file (a shape file or text font, a big font, or a Unicode
|
||||
/// font, as its first definition says).
|
||||
pub fn compile(source: &str) -> Result<Vec<u8>, ShxError> {
|
||||
if source.len() > MAX_SOURCE {
|
||||
return Err(ShxError::TooLarge);
|
||||
}
|
||||
let mut kind = Kind::Shapes;
|
||||
let mut ranges: Vec<(u16, u16)> = Vec::new();
|
||||
let mut uni_info: Option<(String, Vec<u8>)> = None;
|
||||
// Definitions: header (number, name, line) and the numbers that follow it.
|
||||
let mut defs: Vec<(i64, String, usize, Vec<i64>)> = Vec::new();
|
||||
let mut uni_header_pending = false;
|
||||
for (no, raw) in source.lines().enumerate() {
|
||||
let line = no + 1;
|
||||
let text = raw.split(';').next().unwrap_or_default().trim();
|
||||
if text.is_empty() {
|
||||
continue;
|
||||
}
|
||||
if let Some(head) = text.strip_prefix('*') {
|
||||
let upper = head.trim_start().to_ascii_uppercase();
|
||||
if let Some(rest) = upper.strip_prefix("BIGFONT") {
|
||||
kind = Kind::Big;
|
||||
let v = numbers(rest, line)?;
|
||||
let n = usize::try_from(v.get(1).copied().unwrap_or(0)).unwrap_or(0);
|
||||
for k in 0..n {
|
||||
let lo = v.get(2 + 2 * k).copied().unwrap_or(0);
|
||||
let hi = v.get(3 + 2 * k).copied().unwrap_or(lo);
|
||||
let to = |x: i64| u16::try_from(x).map_err(|_| ShxError::Source(line, "bad escape range".into()));
|
||||
ranges.push((to(lo)?, to(hi)?));
|
||||
}
|
||||
continue;
|
||||
}
|
||||
if upper.starts_with("UNIFONT") {
|
||||
kind = Kind::Uni;
|
||||
let name = head.splitn(3, ',').nth(2).unwrap_or_default().trim().to_string();
|
||||
uni_info = Some((name, Vec::new()));
|
||||
uni_header_pending = true;
|
||||
continue;
|
||||
}
|
||||
uni_header_pending = false;
|
||||
let mut parts = head.splitn(3, ',');
|
||||
let num = number(parts.next().unwrap_or_default().trim(), line)?;
|
||||
let _defbytes = parts.next();
|
||||
let name = parts.next().unwrap_or_default().trim().to_string();
|
||||
defs.push((num, name, line, Vec::new()));
|
||||
continue;
|
||||
}
|
||||
let v = numbers(text, line)?;
|
||||
if uni_header_pending {
|
||||
if let Some((_, info)) = uni_info.as_mut() {
|
||||
for x in v {
|
||||
info.push(u8::try_from(x).map_err(|_| ShxError::Source(line, "bad font record".into()))?);
|
||||
}
|
||||
}
|
||||
continue;
|
||||
}
|
||||
match defs.last_mut() {
|
||||
Some(d) => d.3.extend(v),
|
||||
None => return Err(ShxError::Source(line, "specification bytes before any *number,defbytes,NAME line".into())),
|
||||
}
|
||||
}
|
||||
let mut shapes: Vec<Shape> = Vec::new();
|
||||
for (num, name, line, toks) in defs {
|
||||
let number = u16::try_from(num).map_err(|_| ShxError::Source(line, format!("shape number {num} out of range")))?;
|
||||
let program = if number == 0 {
|
||||
// The font record: plain bytes.
|
||||
toks.iter().map(|x| u8::try_from(*x).map_err(|_| ShxError::Source(line, "bad font record".into()))).collect::<Result<Vec<u8>, _>>()?
|
||||
} else {
|
||||
Enc { toks: &toks, i: 0, out: Vec::new(), kind, line }.program()?
|
||||
};
|
||||
if shapes.iter().any(|s| s.number == number) {
|
||||
return Err(ShxError::Source(line, format!("shape {number} is defined twice")));
|
||||
}
|
||||
shapes.push(Shape { number, name, program });
|
||||
}
|
||||
if shapes.is_empty() {
|
||||
return Err(ShxError::Empty);
|
||||
}
|
||||
shapes.sort_by_key(|s| s.number);
|
||||
let record = |s: &Shape| -> Result<Vec<u8>, ShxError> {
|
||||
let mut r: Vec<u8> = s.name.bytes().filter(|b| *b != 0).collect();
|
||||
r.push(0);
|
||||
r.extend(&s.program);
|
||||
if r.len() > usize::from(u16::MAX) { Err(ShxError::TooLarge) } else { Ok(r) }
|
||||
};
|
||||
let count = u16::try_from(shapes.len()).map_err(|_| ShxError::TooLarge)?;
|
||||
let len16 = |r: &[u8]| u16::try_from(r.len()).unwrap_or(u16::MAX);
|
||||
let mut out = Vec::new();
|
||||
match kind {
|
||||
Kind::Shapes => {
|
||||
out.extend(SHAPES_10);
|
||||
out.extend(shapes.first().map_or(0, |s| s.number).to_le_bytes());
|
||||
out.extend(shapes.last().map_or(0, |s| s.number).to_le_bytes());
|
||||
out.extend(count.to_le_bytes());
|
||||
let recs: Vec<Vec<u8>> = shapes.iter().map(record).collect::<Result<_, _>>()?;
|
||||
for (s, r) in shapes.iter().zip(&recs) {
|
||||
out.extend(s.number.to_le_bytes());
|
||||
out.extend(len16(r).to_le_bytes());
|
||||
}
|
||||
for r in &recs {
|
||||
out.extend(r);
|
||||
}
|
||||
out.extend(b"EOF");
|
||||
}
|
||||
Kind::Big => {
|
||||
out.extend(BIGFONT);
|
||||
let nranges = u16::try_from(ranges.len()).map_err(|_| ShxError::TooLarge)?;
|
||||
out.extend(count.to_le_bytes());
|
||||
out.extend(count.to_le_bytes());
|
||||
out.extend(nranges.to_le_bytes());
|
||||
for (lo, hi) in &ranges {
|
||||
out.extend(lo.to_le_bytes());
|
||||
out.extend(hi.to_le_bytes());
|
||||
}
|
||||
let recs: Vec<Vec<u8>> = shapes.iter().map(record).collect::<Result<_, _>>()?;
|
||||
let mut off = out.len() + 8 * recs.len();
|
||||
for (s, r) in shapes.iter().zip(&recs) {
|
||||
out.extend(s.number.to_le_bytes());
|
||||
out.extend(len16(r).to_le_bytes());
|
||||
out.extend(u32::try_from(off).map_err(|_| ShxError::TooLarge)?.to_le_bytes());
|
||||
off += r.len();
|
||||
}
|
||||
for r in &recs {
|
||||
out.extend(r);
|
||||
}
|
||||
}
|
||||
Kind::Uni => {
|
||||
out.extend(UNIFONT);
|
||||
let (name, info) = uni_info.unwrap_or_default();
|
||||
out.extend((u32::from(count) + 1).to_le_bytes());
|
||||
let mut rec: Vec<u8> = name.bytes().filter(|b| *b != 0).collect();
|
||||
rec.push(0);
|
||||
rec.extend(&info);
|
||||
out.extend(len16(&rec).to_le_bytes());
|
||||
out.extend(&rec);
|
||||
for s in &shapes {
|
||||
let r = record(s)?;
|
||||
out.extend(s.number.to_le_bytes());
|
||||
out.extend(len16(&r).to_le_bytes());
|
||||
out.extend(r);
|
||||
}
|
||||
}
|
||||
}
|
||||
Ok(out)
|
||||
}
|
||||
@@ -26,7 +26,6 @@
|
||||
//! above, below, modes, encoding, embedding, 0) and then each shape as (number, length,
|
||||
//! record).
|
||||
|
||||
mod compile;
|
||||
mod parse;
|
||||
mod program;
|
||||
mod table;
|
||||
@@ -39,7 +38,7 @@ use std::sync::{Arc, PoisonError, RwLock};
|
||||
|
||||
use cadcraft_geom::Vec2;
|
||||
|
||||
pub use compile::compile;
|
||||
pub use craft_shx::{ShxError, compile};
|
||||
pub use table::{
|
||||
FontProvider, add_drawing_dir, clear_registered, drawing_dirs, find, font_map, generation, installed, register, search_path, set_font_dirs,
|
||||
set_font_map, set_provider, set_search_path,
|
||||
@@ -47,35 +46,6 @@ pub use table::{
|
||||
pub use text::ShxText;
|
||||
pub(crate) use text::{shape, width};
|
||||
|
||||
/// Why a font file can't be read or a shape source can't be compiled.
|
||||
#[derive(Clone, Debug, PartialEq, Eq)]
|
||||
pub enum ShxError {
|
||||
/// Not an SHX file (unknown signature).
|
||||
NotShx,
|
||||
/// The file ends inside its header or index.
|
||||
Truncated,
|
||||
/// The file is larger, or has more of something, than CADCraft reads.
|
||||
TooLarge,
|
||||
/// The file is well-formed but holds no shape.
|
||||
Empty,
|
||||
/// Shape source error: line and message.
|
||||
Source(usize, String),
|
||||
}
|
||||
|
||||
impl std::fmt::Display for ShxError {
|
||||
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
|
||||
match self {
|
||||
ShxError::NotShx => f.write_str("not an SHX shape or font file"),
|
||||
ShxError::Truncated => f.write_str("the SHX file is truncated"),
|
||||
ShxError::TooLarge => f.write_str("the SHX file is too large"),
|
||||
ShxError::Empty => f.write_str("the SHX file has no shapes"),
|
||||
ShxError::Source(line, msg) => write!(f, "line {line}: {msg}"),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
impl std::error::Error for ShxError {}
|
||||
|
||||
/// The kind of SHX file.
|
||||
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
|
||||
pub enum ShxKind {
|
||||
|
||||
@@ -99,7 +99,7 @@ fn utf8_wins_when_valid() {
|
||||
/// A code page CADCraft doesn't know (or none) reads as ANSI_1252, as before.
|
||||
#[test]
|
||||
fn unknown_code_page_falls_back_to_ansi_1252() {
|
||||
for cp in ["ANSI_1361", "EBCDIC", "", "\u{1}"] {
|
||||
for cp in ["ANSI_UNKNOWN", "EBCDIC", "", "\u{1}"] {
|
||||
let d = cadcraft_io::read_dxf(&ascii_dxf("AC1015", cp, b"Caf\xE9", b"B", b"\x80 \xD8")).unwrap();
|
||||
assert_eq!(names(&d), ("Café".into(), "B".into(), "€ Ø".into()), "{cp:?}");
|
||||
}
|
||||
@@ -209,7 +209,9 @@ fn binary_dxf_honours_the_code_page() {
|
||||
// a pre-R2007 binary file has no way to escape to UTF-8, so its strings are never read as it.
|
||||
assert_eq!(read(&bin("AC1015", "ANSI_1251", b"\xD1\xB2")), "СІ");
|
||||
assert_eq!(read(&bin("AC1021", "ANSI_1251", "Привет".as_bytes())), "Привет");
|
||||
assert_eq!(read(&bin("AC1015", "ANSI_1361", b"Caf\xE9")), "Café");
|
||||
assert_eq!(read(&bin("AC1015", "ANSI_UNKNOWN", b"Caf\xE9")), "Café");
|
||||
assert_eq!(read(&bin("AC1015", "ANSI_1361", b"\xD0\x65\x8B\x69")), "한글");
|
||||
assert_eq!(read(&bin("AC1015", "ANSI_1361", b"Caf\xE9")), "Caf\u{FFFD}");
|
||||
}
|
||||
|
||||
/// Truncated and invalid multibyte sequences become U+FFFD; nothing panics and the structure of
|
||||
|
||||
Reference in New Issue
Block a user