#![allow(clippy::unwrap_used, clippy::panic)]
use zumen_base::{Budget, CancelToken, Cx, Diagnostics, Error, Kind, Limit, Limits};
use zumen_xml::{Document, Edit, Encoding, NewElement, Node};
const SOURCE: &str = "<?xml version='1.0' encoding='UTF-8'?>\r\n<!DOCTYPE 管理 SYSTEM 'synthetic.dtd'>\r\n<管理 xmlns:p='urn:test' version='1'><!--keep-->\r\n\t<項目 code='x'> old <![CDATA[<part>]]><?note x?> tail </項目>\r\n\t<空 属性 = \"保つ\" />\r\n\t<反復>first</反復>\r\n\t<反復>second</反復>\r\n</管理>\r\n";
#[test]
fn mutated_sources_stop_deterministically_before_or_after_writing() {
for encoding in [
Encoding::Utf8,
Encoding::ShiftJis,
Encoding::Utf16Le,
Encoding::Utf16Be,
] {
let seed = input(encoding, SOURCE);
for position in 0..seed.len() {
for byte in [0, b'<', b'/', 0xff] {
let mut source = seed.clone();
if let Some(value) = source.get_mut(position) {
*value = byte;
}
zumen_sdk::testkit::survives_with(&source, |source, cx| {
let doc = zumen_xml::read(source, cx)?;
let child = leaf("added", "<&>\n\r");
let written = zumen_xml::write(
&doc,
&[Edit::Append {
element: doc.root() as u64,
child: &child,
}],
cx,
)?;
zumen_xml::read(&written, cx)
})
.unwrap();
}
}
}
}
fn input(encoding: Encoding, source: &str) -> Vec<u8> {
match encoding {
Encoding::Utf8 => source.as_bytes().to_vec(),
Encoding::ShiftJis => {
let text = source.replace("UTF-8", "Shift_JIS");
let (data, _, changed) = encoding_rs::SHIFT_JIS.encode(&text);
assert!(!changed);
data.into_owned()
}
Encoding::Utf16Le => [
vec![0xff, 0xfe],
source
.replace("UTF-8", "UTF-16")
.encode_utf16()
.flat_map(u16::to_le_bytes)
.collect(),
]
.concat(),
Encoding::Utf16Be => [
vec![0xfe, 0xff],
source
.replace("UTF-8", "UTF-16")
.encode_utf16()
.flat_map(u16::to_be_bytes)
.collect(),
]
.concat(),
_ => panic!("未対応の文字コード"),
}
}
fn read(bytes: &[u8]) -> Document {
zumen_xml::read(
bytes,
&mut Cx::new(
&mut Budget::default(),
&mut Diagnostics::new(),
&CancelToken::new(),
),
)
.unwrap()
}
fn write(
doc: &Document,
edits: &[Edit<'_>],
limits: Limits,
cancel: &CancelToken,
) -> (Result<Vec<u8>, Error>, Diagnostics, Budget) {
let mut budget = Budget::new(limits);
let mut diag = Diagnostics::new();
let result = zumen_xml::write(doc, edits, &mut Cx::new(&mut budget, &mut diag, cancel));
(result, diag, budget)
}
fn index(doc: &Document, name: &str) -> u64 {
doc.elements()
.iter()
.position(|e| doc.cdata(e.name().raw()) == name)
.unwrap() as u64
}
fn value(doc: &Document, name: &str) -> String {
let element = doc.element(index(doc, name) as usize).unwrap();
element
.children()
.iter()
.filter_map(|node| match node {
Node::Text(range) => Some(doc.text(range.clone())),
Node::CData(range) => Some(doc.cdata(range.clone())),
Node::Element(_) => None,
})
.collect()
}
fn leaf(name: &str, value: &str) -> NewElement {
NewElement {
name: name.into(),
value: Some(value.into()),
children: Vec::new(),
}
}
#[test]
fn edits_only_target_bytes_in_all_encodings() {
for encoding in [
Encoding::Utf8,
Encoding::ShiftJis,
Encoding::Utf16Le,
Encoding::Utf16Be,
] {
let source = input(encoding, SOURCE);
let doc = read(&source);
let child = NewElement {
name: "p:追加".into(),
value: None,
children: vec![leaf("子", "入れ子")],
};
let second = leaf("追加2", "最後");
let edits = [
Edit::SetText {
element: index(&doc, "項目"),
value: "新<&>\n\r終",
},
Edit::SetText {
element: index(&doc, "空"),
value: "値",
},
Edit::Append {
element: doc.root() as u64,
child: &child,
},
Edit::Append {
element: doc.root() as u64,
child: &second,
},
];
let (result, diag, _) = write(&doc, &edits, Limits::default(), &CancelToken::new());
assert_eq!(diag.iter().count(), 0);
let bytes = result.unwrap();
let expected = SOURCE
.replace(
" old <![CDATA[<part>]]><?note x?> tail ",
"新<&>\r\n 終<?note x?>",
)
.replace("<空 属性 = \"保つ\" />", "<空 属性 = \"保つ\" >値</空>")
.replace(
"</管理>",
"\t<p:追加><子>入れ子</子></p:追加>\r\n\t<追加2>最後</追加2>\r\n</管理>",
);
assert_eq!(bytes, input(encoding, &expected));
let reread = read(&bytes);
assert_eq!(reread.encoding(), encoding);
assert_eq!(value(&reread, "項目"), "新<&>\n\r終");
assert_eq!(value(&reread, "空"), "値");
assert_eq!(value(&reread, "子"), "入れ子");
let prefixed = reread.element(index(&reread, "p:追加") as usize).unwrap();
assert_eq!(
reread.namespace(prefixed.name()).as_deref(),
Some("urn:test")
);
assert_eq!(doc.src(), source);
}
}
#[test]
fn preserves_bom_inline_layout_and_comments_without_text() {
let source = [
vec![0xef, 0xbb, 0xbf],
b"<r><a><!--same--><?x y?></a><b /></r>".to_vec(),
]
.concat();
let doc = read(&source);
let child = leaf("c", "x");
let edits = [
Edit::SetText {
element: index(&doc, "a"),
value: "ok",
},
Edit::Append {
element: index(&doc, "b"),
child: &child,
},
];
let result = write(&doc, &edits, Limits::default(), &CancelToken::new())
.0
.unwrap();
assert_eq!(
result,
[
vec![0xef, 0xbb, 0xbf],
b"<r><a><!--same--><?x y?>ok</a><b ><c>x</c></b></r>".to_vec()
]
.concat()
);
assert_eq!(
write(&doc, &[], Limits::default(), &CancelToken::new())
.0
.unwrap(),
source
);
}
#[test]
fn rejects_conflicts_children_bad_names_namespace_and_characters() {
let doc = read(b"<r><a>old</a></r>");
for (edits, code) in [
(
vec![
Edit::SetText {
element: 1,
value: "x",
},
Edit::SetText {
element: 1,
value: "y",
},
],
"xml.edit-range.invalid",
),
(
vec![Edit::SetText {
element: 0,
value: "x",
}],
"xml.edit-children.unsupported",
),
(
vec![Edit::SetText {
element: u64::MAX,
value: "x",
}],
"xml.edit-target.missing",
),
(
vec![Edit::SetText {
element: 1,
value: "\0",
}],
"xml.edit-character.invalid",
),
(
vec![Edit::SetText {
element: 1,
value: "\u{ffff}",
}],
"xml.edit-character.invalid",
),
] {
let (result, diag, _) = write(&doc, &edits, Limits::default(), &CancelToken::new());
assert!(result.is_err());
assert_eq!(diag.iter().next().unwrap().code().name(), code);
}
for (name, code) in [
("1bad", "xml.edit-name.invalid"),
("a/><b", "xml.edit-name.invalid"),
("q:a", "xml.edit-namespace.missing"),
("xmlns:a", "xml.edit-namespace.missing"),
] {
let child = leaf(name, "ok");
let (result, diag, _) = write(
&doc,
&[Edit::Append {
element: 0,
child: &child,
}],
Limits::default(),
&CancelToken::new(),
);
assert!(result.is_err());
assert_eq!(diag.iter().next().unwrap().code().name(), code);
}
let child = leaf("a", "x");
assert!(
write(
&doc,
&[
Edit::Append {
element: 1,
child: &child
},
Edit::SetText {
element: 1,
value: "x"
}
],
Limits::default(),
&CancelToken::new()
)
.0
.is_err()
);
assert_eq!(doc.src(), b"<r><a>old</a></r>");
}
#[test]
fn rejects_unrepresentable_or_changed_shift_jis_text() {
let doc = read(&input(
Encoding::ShiftJis,
"<?xml version='1.0' encoding='UTF-8'?><r><a>old</a></r>",
));
for value in ["😀", "¥", "‾"] {
let (result, diag, _) = write(
&doc,
&[Edit::SetText { element: 1, value }],
Limits::default(),
&CancelToken::new(),
);
assert!(result.is_err());
assert_eq!(
diag.iter().next().unwrap().code().name(),
"xml.edit-encoding.failed"
);
}
}
#[test]
fn shared_budgets_stop_before_output_and_restore_depth() {
let doc = read(b"<r><p/></r>");
let child = NewElement {
name: "c".into(),
value: None,
children: vec![leaf("d", "value")],
};
let edits = [Edit::Append {
element: 1,
child: &child,
}];
let (full, _, used) = write(&doc, &edits, Limits::default(), &CancelToken::new());
let full = full.unwrap();
for kind in [Kind::Work, Kind::DecodedBytes, Kind::Records] {
let limit = used.used(kind);
assert!(limit > 0);
assert_eq!(
write(
&doc,
&edits,
Limits::default().with(kind, Limit::Finite(limit)),
&CancelToken::new()
)
.0
.unwrap(),
full
);
for amount in [0, limit / 2, limit - 1] {
let (result, _, budget) = write(
&doc,
&edits,
Limits::default().with(kind, Limit::Finite(amount)),
&CancelToken::new(),
);
assert!(matches!(result, Err(Error::Stopped(_))));
assert_eq!(budget.used(Kind::Depth), 0);
}
}
assert!(
write(
&doc,
&edits,
Limits::default().with(Kind::Depth, Limit::Finite(3)),
&CancelToken::new()
)
.0
.is_err()
);
assert_eq!(
write(
&doc,
&edits,
Limits::default().with(Kind::Depth, Limit::Finite(4)),
&CancelToken::new()
)
.0
.unwrap(),
full
);
let cancel = CancelToken::new();
cancel.cancel();
assert!(matches!(
write(&doc, &edits, Limits::default(), &cancel).0,
Err(Error::Stopped(_))
));
assert_eq!(
write(&doc, &edits, Limits::default(), &CancelToken::new())
.0
.unwrap(),
full
);
}