added support for share-safe requirement, options

Cargo.toml

@@ -4,7 +4,7 @@ description = """
 Mercurial repository changelog parser. It allows to get any revision
 with creation date, user, parents, branch and files.
 """
-edition = "2018"
+edition = "2021"
 homepage = "https://github.com/kilork/hg-parser"
 keywords = ["hg", "mercurial", "parser"]
 license = "GPL-2.0-or-later"
@@ -14,14 +14,14 @@ version = "0.8.0"
 rust-version = "1.78"
 
 [dependencies]
-bitflags = {version = "2.5", default-features = false}
-byteorder = {version = "1.3", default-features = false}
-chrono = {version = "0.4", features = ["clock"], default-features = false}
+bitflags = { version = "2.5", default-features = false }
+byteorder = { version = "1.3", default-features = false }
+chrono = { version = "0.4", features = ["clock"], default-features = false }
 flate2 = "1.0"
-lazy_static = {version = "1.4", default-features = false}
-lru-cache = {version = "0.1", default-features = false}
-nom = {version = "7.1", default-features = false, features = ["alloc"]}
-ordered-parallel-iterator = {version = "0.2", default-features = false}
-rayon = {version = "1.4", default-features = false}
-sha-1 = {version = "0.10", default-features = false}
-thiserror = "1"
+lazy_static = { version = "1.4", default-features = false }
+lru-cache = { version = "0.1", default-features = false }
+nom = { version = "7.1", default-features = false, features = ["alloc"] }
+ordered-parallel-iterator = { version = "0.2", default-features = false }
+rayon = { version = "1.4", default-features = false }
+sha-1 = { version = "0.10", default-features = false }
+thiserror = "1"

README.md

@@ -13,14 +13,16 @@ hg-parser = "0.8"
 
 ### Use case - Analyse revision log and export to ```git fast-import``` format
 
-```rust#ignore
-use hg_parser::{file_content, FileType, ManifestEntryDetails, MercurialRepository, Revision};
+```rust, no_run
+use std::{
+    env::args,
+    io::Write,
+    path::{Path, PathBuf},
+    string::ParseError,
+    time::Instant,
+};
 
-use std::env::args;
-use std::io::Write;
-use std::path::{Path, PathBuf};
-use std::string::ParseError;
-use std::time::Instant;
+use hg_parser::{file_content, FileType, ManifestEntryDetails, MercurialRepository, Revision};
 
 fn main() -> Result<(), Error> {
     let path: PathBuf = args().nth(1).expect("path not provided").parse()?;

examples/git-fast-import.rs

@@ -1,10 +1,12 @@
-use hg_parser::{file_content, FileType, ManifestEntryDetails, MercurialRepository, Revision};
+use std::{
+    env::args,
+    io::Write,
+    path::{Path, PathBuf},
+    string::ParseError,
+    time::Instant,
+};
 
-use std::env::args;
-use std::io::Write;
-use std::path::{Path, PathBuf};
-use std::string::ParseError;
-use std::time::Instant;
+use hg_parser::{file_content, FileType, ManifestEntryDetails, MercurialRepository, Revision};
 
 fn main() -> Result<(), Error> {
     let path: PathBuf = args().nth(1).expect("path not provided").parse()?;

src/error.rs

@@ -1,3 +1,5 @@
+use crate::types::RepositoryRequire;
+
 #[derive(Debug, thiserror::Error)]
 pub enum ErrorKind {
     /// Parser failed.
@@ -20,7 +22,7 @@ pub enum ErrorKind {
     InvalidDateTime(String),
     /// Requirement in ``.hg/requires`` is not supported.
     #[error("unknown requirement {0}")]
-    UnknownRequirement(String),
+    UnknownRequirement(RepositoryRequire),
     /// Manifest issue.
     #[error("manifest issue {0}")]
     Manifest(String),

src/lib.rs

@@ -22,16 +22,23 @@ mod revisionlog;
 mod types;
 
 use cache::{Cachable, Cache};
-pub use changeset::*;
-pub use error::ErrorKind;
 use manifest::Manifest;
 use path::{fncache_fsencode, simple_fsencode, MPath, MPathElement};
 use revisionlog::RevisionLog;
 use types::{MercurialTag, NodeHash, RepositoryRequire};
 
+pub use changeset::*;
+pub use error::ErrorKind;
 pub use manifest::{FileType, ManifestEntry, ManifestEntryDetails};
 pub use types::{Revision, RevisionRange};
 
+/// Options for mercurial repository.
+#[derive(Default)]
+pub struct MercurialRepositoryOptions {
+    /// Threat unknown requirements as warnings.
+    pub ignore_unknown_requirements: bool,
+}
+
 #[derive(Debug)]
 /// Mercurial repository. Top-level structure for access to change sets and tags.
 pub struct MercurialRepository {
@@ -44,9 +51,17 @@ pub struct MercurialRepository {
 impl MercurialRepository {
     /// Opens `MercurialRepository` at `root_path`.
     pub fn open<P: AsRef<Path>>(root_path: P) -> Result<MercurialRepository, ErrorKind> {
+        Self::open_with_options(root_path, Default::default())
+    }
+
+    /// Opens `MercurialRepository` at `root_path` with options.
+    pub fn open_with_options<P: AsRef<Path>>(
+        root_path: P,
+        options: MercurialRepositoryOptions,
+    ) -> Result<MercurialRepository, ErrorKind> {
         let base = root_path.as_ref().join(".hg");
 
-        let requires = MercurialRepository::load_requires(&base)?;
+        let requires = MercurialRepository::load_requires(&base, &options)?;
 
         let store = base.join("store");
 
@@ -142,13 +157,23 @@ impl MercurialRepository {
 
     fn load_requires<P: AsRef<Path>>(
         path: P,
-    ) -> Result<HashSet<RepositoryRequire>, std::io::Error> {
+        options: &MercurialRepositoryOptions,
+    ) -> Result<HashSet<RepositoryRequire>, ErrorKind> {
         let requires_path = path.as_ref().join("requires");
         let file = File::open(requires_path)?;
-        Ok(BufReader::new(file)
-            .lines()
-            .map(|x| x.unwrap().parse().expect("could not parse requirement"))
-            .collect())
+        let lines = BufReader::new(file).lines().map_while(Result::ok);
+        if options.ignore_unknown_requirements {
+            lines
+                .map(|x| match x.parse() {
+                    Err(ErrorKind::UnknownRequirement(r)) => Ok(r),
+                    other => other,
+                })
+                .collect()
+        } else {
+            Ok(lines
+                .map(|x| x.parse().expect("could not parse requirement"))
+                .collect())
+        }
     }
 
     fn fsencode_path(&self, elements: &[MPathElement]) -> PathBuf {

src/types.rs

@@ -1,23 +1,24 @@
 use std::{
-    fmt::{self, Debug, Display},
+    fmt::{self, Display},
     ops::{Add, Range, Sub},
     rc::Rc,
     str::FromStr,
     sync::Arc,
 };
 
-/// Repository requires flags.
-/// Repositories contain a file (``.hg/requires``) containing a list of
-/// features/capabilities that are *required* for clients to interface
-/// with the repository.
 use super::error::ErrorKind;
 use bitflags::bitflags;
 use chrono::{
     DateTime as ChronoDateTime, FixedOffset, Local, LocalResult, NaiveDateTime, TimeZone,
 };
 use sha1::{Digest, Sha1};
-#[derive(Debug, Copy, Clone, Eq, PartialEq, Hash)]
-pub(crate) enum RepositoryRequire {
+
+/// Repository requires flags.
+/// Repositories contain a file (``.hg/requires``) containing a list of
+/// features/capabilities that are *required* for clients to interface
+/// with the repository.
+#[derive(Debug, Clone, Eq, PartialEq, Hash)]
+pub enum RepositoryRequire {
     /// When present, revlogs are version 1 (**RevlogNG**).
     Revlogv1,
     /// The **store** repository layout is used.
@@ -45,6 +46,10 @@ pub(crate) enum RepositoryRequire {
     TreeManifest,
     /// The working directory is sparse (only contains a subset of files).
     ExpSparse,
+    /// Safe behaviour for all shares that access a repository.
+    ShareSafe,
+    /// Unknown requirement.
+    Unknown(String),
 }
 
 impl FromStr for RepositoryRequire {
@@ -64,7 +69,29 @@ impl FromStr for RepositoryRequire {
             "manifestv2" => Ok(Manifestv2),
             "treemanifest" => Ok(TreeManifest),
             "exp-sparse" => Ok(ExpSparse),
-            other => Err(ErrorKind::UnknownRequirement(other.into())),
+            "share-safe" => Ok(ShareSafe),
+            other => Err(ErrorKind::UnknownRequirement(Self::Unknown(other.into()))),
+        }
+    }
+}
+
+impl std::fmt::Display for RepositoryRequire {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        use RepositoryRequire::*;
+        match self {
+            Revlogv1 => "revlogv1".fmt(f),
+            Store => "store".fmt(f),
+            FnCache => "fncache".fmt(f),
+            Shared => "shared".fmt(f),
+            RelShared => "relshared".fmt(f),
+            DotEncode => "dotencode".fmt(f),
+            ParentDelta => "parentdelta".fmt(f),
+            GeneralDelta => "generaldelta".fmt(f),
+            Manifestv2 => "manifestv2".fmt(f),
+            TreeManifest => "treemanifest".fmt(f),
+            ExpSparse => "exp-sparse".fmt(f),
+            ShareSafe => "share-safe".fmt(f),
+            Unknown(s) => s.fmt(f),
         }
     }
 }
@@ -81,7 +108,7 @@ impl Revision {
 
     /// Return an open ended iterator from index.
     pub fn range(self) -> RevisionRange {
-        RevisionRange(self.0, std::u32::MAX)
+        RevisionRange(self.0, u32::MAX)
     }
 }
 
@@ -221,13 +248,13 @@ impl AsRef<[u8]> for NodeHash {
     }
 }
 
-impl Display for NodeHash {
+impl std::fmt::Display for NodeHash {
     fn fmt(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
         Display::fmt(&self.to_hex(), fmt)
     }
 }
 
-impl Debug for NodeHash {
+impl std::fmt::Debug for NodeHash {
     fn fmt(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
         write!(fmt, "NodeHash({})", self)
     }
@@ -303,7 +330,7 @@ impl Fragment {
     }
 }
 
-impl Debug for Fragment {
+impl std::fmt::Debug for Fragment {
     fn fmt(&self, fmt: &mut fmt::Formatter) -> fmt::Result {
         write!(
             fmt,

templates/README.md

@@ -13,7 +13,7 @@ hg-parser = "0.8"
 
 ### Use case - Analyse revision log and export to ```git fast-import``` format
 
-{{ codeblock "rust#ignore" ( read_to_str "examples/git-fast-import.rs" ) }}
+{{ codeblock "rust, no_run" ( read_to_str "examples/git-fast-import.rs" ) }}
 
 ## Implementation details