Merge branch 'fixes'

Author: Byron

gix-attributes/src/lib.rs

@@ -6,7 +6,8 @@
     doc = ::document_features::document_features!()
 )]
 #![cfg_attr(all(doc, feature = "document-features"), feature(doc_cfg, doc_auto_cfg))]
-#![deny(missing_docs, rust_2018_idioms, unsafe_code)]
+#![deny(missing_docs, rust_2018_idioms)]
+#![forbid(unsafe_code)]
 
 pub use gix_glob as glob;
 use kstring::{KString, KStringRef};

gix-attributes/src/name.rs

@@ -1,8 +1,7 @@
+use crate::{Name, NameRef};
 use bstr::{BStr, BString, ByteSlice};
 use kstring::KStringRef;
 
-use crate::{Name, NameRef};
-
 impl<'a> NameRef<'a> {
     /// Turn this ref into its owned counterpart.
     pub fn to_owned(self) -> Name {

gix-attributes/src/parse.rs

@@ -1,10 +1,9 @@
 use std::borrow::Cow;
 
+use crate::{name, AssignmentRef, Name, NameRef, StateRef};
 use bstr::{BStr, ByteSlice};
 use kstring::KStringRef;
 
-use crate::{name, AssignmentRef, Name, NameRef, StateRef};
-
 /// The kind of attribute that was parsed.
 #[derive(PartialEq, Eq, Debug, Hash, Ord, PartialOrd, Clone)]
 #[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]

gix-attributes/src/search/mod.rs

@@ -1,7 +1,6 @@
-use std::collections::HashMap;
-
 use kstring::KString;
 use smallvec::SmallVec;
+use std::collections::HashMap;
 
 use crate::{Assignment, AssignmentRef};
 

gix-attributes/src/state.rs

@@ -1,29 +1,24 @@
-use bstr::{BStr, ByteSlice};
-use kstring::{KString, KStringRef};
-
 use crate::{State, StateRef};
+use bstr::{BStr, BString, ByteSlice};
 
 /// A container to encapsulate a tightly packed and typically unallocated byte value that isn't necessarily UTF8 encoded.
 #[derive(PartialEq, Eq, Debug, Hash, Ord, PartialOrd, Clone)]
 #[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
-pub struct Value(KString);
+// TODO: This should be some sort of 'smallbstring' - but can't use `kstring` here due to UTF8 requirement. 5% performance boost possible.
+//       What's really needed here is a representation that displays as string when serialized which helps with JSON.
+//       Maybe `smallvec` with display and serialization wrapper would do the trick?
+pub struct Value(BString);
 
 /// A reference container to encapsulate a tightly packed and typically unallocated byte value that isn't necessarily UTF8 encoded.
 #[derive(PartialEq, Eq, Debug, Hash, Ord, PartialOrd, Clone, Copy)]
 #[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
-pub struct ValueRef<'a>(#[cfg_attr(feature = "serde", serde(borrow))] KStringRef<'a>);
+pub struct ValueRef<'a>(#[cfg_attr(feature = "serde", serde(borrow))] &'a [u8]);
 
 /// Lifecycle
 impl<'a> ValueRef<'a> {
     /// Keep `input` as our value.
     pub fn from_bytes(input: &'a [u8]) -> Self {
-        Self(KStringRef::from_ref(
-            // SAFETY: our API makes accessing that value as `str` impossible, so illformed UTF8 is never exposed as such.
-            #[allow(unsafe_code)]
-            unsafe {
-                std::str::from_utf8_unchecked(input)
-            },
-        ))
+        Self(input)
     }
 }
 
@@ -42,7 +37,7 @@ impl ValueRef<'_> {
 
 impl<'a> From<&'a str> for ValueRef<'a> {
     fn from(v: &'a str) -> Self {
-        ValueRef(v.into())
+        ValueRef(v.as_bytes())
     }
 }
 
@@ -54,7 +49,7 @@ impl<'a> From<ValueRef<'a>> for Value {
 
 impl From<&str> for Value {
     fn from(v: &str) -> Self {
-        Value(KString::from_ref(v))
+        Value(v.as_bytes().into())
     }
 }
 

gix-attributes/tests/parse/mod.rs

@@ -1,4 +1,5 @@
 use bstr::BString;
+use gix_attributes::state::ValueRef;
 use gix_attributes::{parse, StateRef};
 use gix_glob::pattern::Mode;
 use gix_testtools::fixture_bytes;
@@ -275,6 +276,19 @@ fn attributes_can_have_values() {
     );
 }
 
+#[test]
+fn attributes_can_have_illformed_utf8() {
+    assert_eq!(
+        byte_line(b"p a=one b=\xC3\x28\x41 c=d "),
+        (
+            pattern("p", Mode::NO_SUB_DIR, None),
+            vec![value("a", "one"), byte_value("b", b"\xC3\x28\x41"), value("c", "d")],
+            1
+        ),
+        "illformed UTF8 is fully supported"
+    );
+}
+
 #[test]
 fn attributes_see_state_adjustments_over_value_assignments() {
     assert_eq!(
@@ -325,6 +339,10 @@ fn value<'b>(attr: &str, value: &'b str) -> (BString, StateRef<'b>) {
     (attr.into(), StateRef::Value(value.into()))
 }
 
+fn byte_value<'b>(attr: &str, value: &'b [u8]) -> (BString, StateRef<'b>) {
+    (attr.into(), StateRef::Value(ValueRef::from_bytes(value)))
+}
+
 fn pattern(name: &str, flags: gix_glob::pattern::Mode, first_wildcard_pos: Option<usize>) -> parse::Kind {
     parse::Kind::Pattern(gix_glob::Pattern {
         text: name.into(),
@@ -344,6 +362,17 @@ fn line(input: &str) -> ExpandedAttribute {
     try_line(input).unwrap()
 }
 
+fn byte_line(input: &[u8]) -> ExpandedAttribute {
+    try_byte_line(input).unwrap()
+}
+
+fn try_byte_line(input: &[u8]) -> Result<ExpandedAttribute, parse::Error> {
+    let mut lines = gix_attributes::parse(input);
+    let res = expand(lines.next().unwrap())?;
+    assert!(lines.next().is_none(), "expected only one line");
+    Ok(res)
+}
+
 fn lenient_lines(input: &str) -> Vec<ExpandedAttribute> {
     gix_attributes::parse(input.as_bytes())
         .map(expand)

gix-attributes/tests/search/mod.rs

@@ -270,7 +270,7 @@ fn given_attributes_are_made_available_in_given_order() -> crate::Result {
 fn size_of_outcome() {
     assert_eq!(
         std::mem::size_of::<Outcome>(),
-        904,
+        840,
         "it's quite big, shouldn't change without us noticing"
     )
 }

gix-dir/src/walk/classify.rs

@@ -169,6 +169,7 @@ pub fn path(
                         .map(|platform| platform.excluded_kind())
                 })
                 .map_err(Error::ExcludesAccess)?
+                .filter(|_| filename_start_idx > 0)
             {
                 out.status = entry::Status::Ignored(excluded);
             }
@@ -256,6 +257,7 @@ pub fn path(
     if let Some(excluded) = ctx
         .excludes
         .as_mut()
+        .filter(|_| !rela_path.is_empty())
         .map_or(Ok(None), |stack| {
             stack
                 .at_entry(rela_path.as_bstr(), is_dir, ctx.objects)

gix-dir/tests/fixtures/many.sh

@@ -351,3 +351,83 @@ git clone submodule multiple-submodules
   git submodule add ../submodule a/b
   git commit -m "add modules"
 )
+
+git clone submodule one-ignored-submodule
+(cd one-ignored-submodule
+  git submodule add ../submodule submodule
+  echo '/submodule/' > .gitignore
+  echo '*' > submodule/.gitignore
+  git commit -m "add seemingly ignored submodule"
+)
+
+git init slash-in-root-and-negated
+(cd slash-in-root-and-negated
+  cat <<'EOF' >.gitignore
+/
+!file
+!*.md
+!.github
+!.github/**
+EOF
+  touch file readme.md
+  mkdir .github
+  touch .github/workflow.yml
+  git add .github readme.md .gitignore
+  git commit -m "init"
+)
+
+git init star-in-root-and-negated
+(cd star-in-root-and-negated
+  cat <<'EOF' >.gitignore
+*
+!file
+!.gitignore
+!*.md
+!.github
+!.github/**
+EOF
+  touch file readme.md
+  mkdir .github
+  touch .github/workflow.yml
+  git add .github readme.md .gitignore
+  git commit -m "init"
+)
+
+git init slash-in-subdir-and-negated
+(cd slash-in-subdir-and-negated
+  mkdir sub
+  (cd sub
+    cat <<'EOF' >.gitignore
+/
+!file
+!*.md
+!.github
+!.github/**
+EOF
+    touch file readme.md
+    mkdir .github
+    touch .github/workflow.yml
+    git add .github readme.md .gitignore
+    git commit -m "init"
+  )
+)
+
+git init star-in-subdir-and-negated
+(cd star-in-subdir-and-negated
+  mkdir sub
+  (cd sub
+    cat <<'EOF' >.gitignore
+*
+!file
+!.gitignore
+!*.md
+!.github
+!.github/**
+EOF
+    touch file readme.md
+    mkdir .github
+    touch .github/workflow.yml
+    git add .github readme.md .gitignore
+    git commit -m "init"
+  )
+)