prometheus-client-mmap 0.20.0 → 0.20.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: aaa63589f21ccefd35765317ce8e1b4ca2a11dc41d3391554fb7a5182616db66
-  data.tar.gz: d58b2c4813b211ced7bf7d3f6c2277a9d5195af3d7ee19ef15ce892f70f4bc4a
+  metadata.gz: f91fbb3a1899d0093d51d92ab8dceec80981c6160b3a09d35d90117f29e6c3d0
+  data.tar.gz: fea591800d5a3a2344082fd38df75f35aea8b5bf5394bcd7e20bf80d14dd48fd
 SHA512:
-  metadata.gz: 24ec66d0f72463e93881eede6333e714f222e48812cf172efd023d17359380513c368ede182476071b02f55ab138aea44b00a0f05c7334c012bcb9083d00e356
-  data.tar.gz: e674a37306171bad42464096958a4d32dd4c8d36f677b8197eb8b64aaf5944c6ec372d6b0a461de10c59204624520be472f78253438c1558b377d99c447074d7
+  metadata.gz: f6e98dda485f79a83fe614a21c52444fad33a69b208c4a6d637516c0390f79c9c54216dda50f7b151483d3e69fe4e7739ed0076984cb7497c59110f0f941b664
+  data.tar.gz: 30bc5288dd7247222128b99a4cf88bdde9ca8625682b0d013dc5ea9cd6ce12aa5285be4b3a875f9a933dcc23e943c4b2cccc9ad9c53e273754a4dac46ddee1ed

data/ext/fast_mmaped_file_rs/Cargo.lock

@@ -9,7 +9,6 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2c99f64d1e06488f620f932677e24bc6e2897582980441ae90a671415bd7ec2f"
 dependencies = [
  "cfg-if",
- "getrandom",
  "once_cell",
  "version_check",
 ]
@@ -161,8 +160,8 @@ dependencies = [
 name = "fast_mmaped_file_rs"
 version = "0.1.0"
 dependencies = [
- "ahash",
  "bstr",
+ "hashbrown",
  "indoc",
  "libc",
  "magnus",
@@ -212,6 +211,15 @@ version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d2fabcfbdc87f4758337ca535fb41a6d701b65693ce38287d856d1674551ec9b"
 
+[[package]]
+name = "hashbrown"
+version = "0.13.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "43a3c133739dddd0d2990f9a4bdf8eb4b21ef50e4851ca85ab661199821d510e"
+dependencies = [
+ "ahash",
+]
+
 [[package]]
 name = "hermit-abi"
 version = "0.3.1"

data/ext/fast_mmaped_file_rs/Cargo.toml

@@ -6,7 +6,7 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-ahash = "0.8"
+hashbrown = "0.13"
 libc = "0.2"
 magnus = { git = "https://github.com/matsadler/magnus", branch = "main", features = ["rb-sys"] }
 memmap2 = "0.5"

data/ext/fast_mmaped_file_rs/extconf.rb

@@ -1,7 +1,9 @@
 require "mkmf"
 require "rb_sys/mkmf"
 
-create_rust_makefile("prometheus-client-mmap/fast_mmaped_file_rs") do |r|
+create_rust_makefile("fast_mmaped_file_rs") do |r|
+  r.auto_install_rust_toolchain = false
+
   if enable_config('fail-on-warning')
     r.extra_rustflags = ["-Dwarnings"]
   end

data/ext/fast_mmaped_file_rs/src/file_entry.rs

@@ -1,4 +1,6 @@
 use magnus::{StaticSymbol, Symbol};
+use std::fmt::Write;
+use std::str;
 
 use crate::error::{MmapError, RubyError};
 use crate::file_info::FileInfo;
@@ -22,27 +24,55 @@ pub struct EntryData {
     pub pid: Option<String>,
 }
 
-impl EntryData {
-    pub fn new(raw_entry: &RawEntry, file_info: &FileInfo, pid_significant: bool) -> Result<Self> {
-        let mut buf = Vec::new();
+impl<'a> PartialEq<BorrowedData<'a>> for EntryData {
+    fn eq(&self, other: &BorrowedData) -> bool {
+        self.pid.as_deref() == other.pid && self.json == other.json
+    }
+}
+
+impl<'a> TryFrom<BorrowedData<'a>> for EntryData {
+    type Error = MmapError;
 
-        if buf.try_reserve_exact(raw_entry.encoded_len()).is_err() {
-            return Err(MmapError::OutOfMemory(raw_entry.encoded_len()));
+    fn try_from(borrowed: BorrowedData) -> Result<Self> {
+        let mut json = String::new();
+        if json.try_reserve_exact(borrowed.json.len()).is_err() {
+            return Err(MmapError::OutOfMemory(borrowed.json.len()));
         }
+        json.push_str(borrowed.json);
 
-        buf.resize(raw_entry.encoded_len(), 0);
-        buf.copy_from_slice(raw_entry.json());
+        Ok(Self {
+            json,
+            // Don't bother checking for allocation failure, typically ~10 bytes
+            pid: borrowed.pid.map(|p| p.to_string()),
+        })
+    }
+}
+
+/// A borrowed copy of the JSON string and pid for a `FileEntry`. We use this
+/// to check if a given string/pid combination is present in the `EntryMap`,
+/// copying them to owned values only when needed.
+#[derive(Hash, PartialEq, Eq, PartialOrd, Ord, Clone, Debug)]
+pub struct BorrowedData<'a> {
+    pub json: &'a str,
+    pub pid: Option<&'a str>,
+}
 
-        let json = String::from_utf8(buf)
+impl<'a> BorrowedData<'a> {
+    pub fn new(
+        raw_entry: &'a RawEntry,
+        file_info: &'a FileInfo,
+        pid_significant: bool,
+    ) -> Result<Self> {
+        let json = str::from_utf8(raw_entry.json())
             .map_err(|e| MmapError::Encoding(format!("invalid UTF-8 in entry JSON: {e}")))?;
 
         let pid = if pid_significant {
-            Some(file_info.pid.clone())
+            Some(file_info.pid.as_str())
         } else {
             None
         };
 
-        Ok(EntryData { json, pid })
+        Ok(Self { json, pid })
     }
 }
 
@@ -121,7 +151,8 @@ impl FileEntry {
 
             entry.append_entry(metrics_data, &mut out)?;
 
-            out.push_str(&format!(" {}\n", entry.meta.value));
+            writeln!(&mut out, " {}", entry.meta.value)
+                .map_err(|e| MmapError::Other(format!("Failed to append to output: {e}")))?;
 
             processed_count += 1;
         }
@@ -419,7 +450,9 @@ mod test {
                 .zip(file_infos)
                 .map(|(entry, info)| {
                     let meta = EntryMetadata::new(&entry, &info).unwrap();
-                    let data = EntryData::new(&entry, &info, meta.is_pid_significant()).unwrap();
+                    let borrowed =
+                        BorrowedData::new(&entry, &info, meta.is_pid_significant()).unwrap();
+                    let data = EntryData::try_from(borrowed).unwrap();
                     FileEntry { data, meta }
                 })
                 .collect();
@@ -526,7 +559,9 @@ mod test {
                 .map(|s| RawEntry::from_slice(s).unwrap())
                 .map(|entry| {
                     let meta = EntryMetadata::new(&entry, &info).unwrap();
-                    let data = EntryData::new(&entry, &info, meta.is_pid_significant()).unwrap();
+                    let borrowed =
+                        BorrowedData::new(&entry, &info, meta.is_pid_significant()).unwrap();
+                    let data = EntryData::try_from(borrowed).unwrap();
                     FileEntry { data, meta }
                 })
                 .collect();

data/ext/fast_mmaped_file_rs/src/map.rs

@@ -1,9 +1,11 @@
-use ahash::AHashMap;
+use hashbrown::hash_map::RawEntryMut;
+use hashbrown::HashMap;
 use magnus::{exception::*, Error, RArray};
+use std::hash::{BuildHasher, Hash, Hasher};
 use std::mem::size_of;
 
 use crate::error::MmapError;
-use crate::file_entry::{EntryData, EntryMetadata, FileEntry};
+use crate::file_entry::{BorrowedData, EntryData, EntryMetadata, FileEntry};
 use crate::file_info::FileInfo;
 use crate::raw_entry::RawEntry;
 use crate::util::read_u32;
@@ -17,12 +19,12 @@ use crate::{err, HEADER_SIZE};
 /// allows us to have multiple entries on the map for multiple pids using the
 /// same string.
 #[derive(Default, Debug)]
-pub struct EntryMap(AHashMap<EntryData, EntryMetadata>);
+pub struct EntryMap(HashMap<EntryData, EntryMetadata>);
 
 impl EntryMap {
     /// Construct a new EntryMap.
     pub fn new() -> Self {
-        Self(AHashMap::new())
+        Self(HashMap::new())
     }
 
     /// Given a list of files, read each one into memory and parse the metrics it contains.
@@ -95,12 +97,28 @@ impl EntryMap {
     /// Check if the `EntryMap` already contains the JSON string.
     /// If yes, update the associated value, if not insert the
     /// entry into the map.
-    pub fn merge_or_store(&mut self, data: EntryData, meta: EntryMetadata) {
-        if let Some(existing) = self.0.get_mut(&data) {
-            existing.merge(&meta);
-        } else {
-            self.0.insert(data, meta);
+    pub fn merge_or_store(&mut self, data: BorrowedData, meta: EntryMetadata) -> Result<()> {
+        // Manually hash the `BorrowedData` and perform an equality check on the
+        // key. This allows us to perform the comparison without allocating a
+        // new `EntryData` that may not be needed.
+        let mut state = self.0.hasher().build_hasher();
+        data.hash(&mut state);
+        let hash = state.finish();
+
+        match self.0.raw_entry_mut().from_hash(hash, |k| k == &data) {
+            RawEntryMut::Vacant(entry) => {
+                // Allocate a new `EntryData` as the JSON/pid combination is
+                // not present in the map.
+                let owned = EntryData::try_from(data)?;
+                entry.insert(owned, meta);
+            }
+            RawEntryMut::Occupied(mut entry) => {
+                let existing = entry.get_mut();
+                existing.merge(&meta);
+            }
         }
+
+        Ok(())
     }
 
     /// Parse metrics data from a `.db` file and store in the `EntryMap`.
@@ -135,9 +153,9 @@ impl EntryMap {
             }
 
             let meta = EntryMetadata::new(&raw_entry, &file_info)?;
-            let data = EntryData::new(&raw_entry, &file_info, meta.is_pid_significant())?;
+            let data = BorrowedData::new(&raw_entry, &file_info, meta.is_pid_significant())?;
 
-            self.merge_or_store(data, meta);
+            self.merge_or_store(data, meta)?;
 
             pos += raw_entry.total_len();
         }
@@ -155,6 +173,16 @@ mod test {
     use crate::file_entry::FileEntry;
     use crate::testhelper::{self, TestFile};
 
+    impl EntryData {
+        /// A helper function for tests to convert owned data to references.
+        fn as_borrowed(&self) -> BorrowedData {
+            BorrowedData {
+                json: &self.json,
+                pid: self.pid.as_deref(),
+            }
+        }
+    }
+
     #[test]
     fn test_into_sorted() {
         let _cleanup = unsafe { magnus::embed::init() };
@@ -311,7 +339,9 @@ mod test {
         map.0
             .insert(starting_entry.data.clone(), starting_entry.meta.clone());
 
-        map.merge_or_store(matching_entry.data, matching_entry.meta);
+        let matching_borrowed = matching_entry.data.as_borrowed();
+        map.merge_or_store(matching_borrowed, matching_entry.meta)
+            .unwrap();
 
         assert_eq!(
             5.0,
@@ -320,10 +350,12 @@ mod test {
         );
         assert_eq!(1, map.0.len(), "no entry added");
 
+        let same_key_different_worker_borrowed = same_key_different_worker.data.as_borrowed();
         map.merge_or_store(
-            same_key_different_worker.data,
+            same_key_different_worker_borrowed,
             same_key_different_worker.meta,
-        );
+        )
+        .unwrap();
 
         assert_eq!(
             5.0,
@@ -333,7 +365,9 @@ mod test {
 
         assert_eq!(2, map.0.len(), "additional entry added");
 
-        map.merge_or_store(unmatched_entry.data, unmatched_entry.meta);
+        let unmatched_entry_borrowed = unmatched_entry.data.as_borrowed();
+        map.merge_or_store(unmatched_entry_borrowed, unmatched_entry.meta)
+            .unwrap();
 
         assert_eq!(
             5.0,

data/lib/prometheus/client/version.rb

@@ -1,5 +1,5 @@
 module Prometheus
   module Client
-    VERSION = '0.20.0'.freeze
+    VERSION = '0.20.2'.freeze
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: prometheus-client-mmap
 version: !ruby/object:Gem::Version
-  version: 0.20.0
+  version: 0.20.2
 platform: ruby
 authors:
 - Tobias Schmidt
@@ -9,7 +9,7 @@ authors:
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-04-13 00:00:00.000000000 Z
+date: 2023-04-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rb_sys
@@ -130,7 +130,6 @@ files:
 - ext/fast_mmaped_file/value_access.c
 - ext/fast_mmaped_file/value_access.h
 - ext/fast_mmaped_file_rs/.cargo/config.toml
-- ext/fast_mmaped_file_rs/.tool-versions
 - ext/fast_mmaped_file_rs/Cargo.lock
 - ext/fast_mmaped_file_rs/Cargo.toml
 - ext/fast_mmaped_file_rs/README.md
@@ -213,7 +212,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.10
+rubygems_version: 3.4.12
 signing_key:
 specification_version: 4
 summary: A suite of instrumentation metric primitivesthat can be exposed through a

data/ext/fast_mmaped_file_rs/.tool-versions

@@ -1 +0,0 @@
-rust 1.65.0