From ebcb9ef2e8c137c5ed6811046a01d87b9d6d351e Mon Sep 17 00:00:00 2001
From: Joey Riches <josephriches@gmail.com>
Date: Wed, 17 Jul 2024 23:02:57 +0100
Subject: [PATCH] boulder/analysis: Add compress man handler

compresses man/info pages
---
 Cargo.lock                              |  1 +
 boulder/Cargo.toml                      |  1 +
 boulder/src/package/analysis.rs         |  1 +
 boulder/src/package/analysis/handler.rs | 78 ++++++++++++++++++++++++-
 crates/stone_recipe/src/lib.rs          |  2 +
 5 files changed, 81 insertions(+), 2 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 0140223a8..3bd4b6186 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -249,6 +249,7 @@ dependencies = [
  "tui",
  "url",
  "yaml",
+ "zstd",
 ]
 
 [[package]]
diff --git a/boulder/Cargo.toml b/boulder/Cargo.toml
index 3ae914e52..730b47460 100644
--- a/boulder/Cargo.toml
+++ b/boulder/Cargo.toml
@@ -43,3 +43,4 @@ strum.workspace = true
 thiserror.workspace = true
 tokio.workspace = true
 url.workspace = true
+zstd.workspace = true
diff --git a/boulder/src/package/analysis.rs b/boulder/src/package/analysis.rs
index 19b89cb4c..7426eed75 100644
--- a/boulder/src/package/analysis.rs
+++ b/boulder/src/package/analysis.rs
@@ -38,6 +38,7 @@ impl<'a> Chain<'a> {
                 Box::new(handler::elf),
                 Box::new(handler::pkg_config),
                 Box::new(handler::cmake),
+                Box::new(handler::compressman),
                 // Catch-all if not excluded
                 Box::new(handler::include_any),
             ],
diff --git a/boulder/src/package/analysis/handler.rs b/boulder/src/package/analysis/handler.rs
index b6d11c098..6a505c0c2 100644
--- a/boulder/src/package/analysis/handler.rs
+++ b/boulder/src/package/analysis/handler.rs
@@ -1,5 +1,13 @@
-use std::{path::PathBuf, process::Command};
-
+use std::fs::File;
+use std::io::{BufReader, BufWriter, Write};
+use std::{
+    fs,
+    os::unix::fs::symlink,
+    path::{Component, PathBuf},
+    process::Command,
+};
+
+use itertools::Itertools;
 use moss::{dependency, Dependency, Provider};
 
 use crate::package::collect::PathInfo;
@@ -134,3 +142,69 @@ pub fn cmake(bucket: &mut BucketMut, info: &mut PathInfo) -> Result<Response, Bo
 
     Ok(Decision::NextHandler.into())
 }
+
+pub fn compressman(bucket: &mut BucketMut, info: &mut PathInfo) -> Result<Response, BoxError> {
+    if !bucket.recipe.parsed.options.compressman {
+        return Ok(Decision::NextHandler.into());
+    }
+
+    let is_man_file = info.path.components().contains(&Component::Normal("man".as_ref()))
+        && info.file_name().ends_with(|c| ('1'..'9').contains(&c));
+    let is_info_file =
+        info.path.components().contains(&Component::Normal("info".as_ref())) && info.file_name().ends_with(".info");
+
+    if !(is_man_file || is_info_file) {
+        return Ok(Decision::NextHandler.into());
+    }
+
+    let mut generated_path = PathBuf::new();
+
+    /* If we have a man/info symlink update the link to the compressed file */
+    if info.path.is_symlink() {
+        let new_original = format!("{}.zst", fs::canonicalize(&info.path)?.display());
+        let new_link = format!("{}.zst", &info.path.display());
+
+        /*
+         * Depending on the order the files get analysed the new compressed file may not yet exist,
+         * compress it _now_ so the correct metadata src info is returned to the binary writer.
+         */
+        if !std::path::Path::new(&new_original).exists() {
+            let compressed_file = compress_file_zstd(fs::canonicalize(&info.path)?)?;
+            let _ = bucket.paths.install().guest.join(compressed_file);
+        }
+
+        symlink(format!("{}.zst", fs::read_link(&info.path)?.display()), &new_link)?;
+        generated_path.push(bucket.paths.install().guest.join(new_link));
+        return Ok(Decision::ReplaceFile {
+            newpath: generated_path,
+        }
+        .into());
+    }
+
+    let mut compressed_file = PathBuf::from(format!("{}.zst", info.path.display()));
+
+    /* We may have already compressed the file if we encountered a symlink to this file first */
+    if !&compressed_file.exists() {
+        compressed_file = compress_file_zstd(info.path.clone())?;
+    }
+
+    generated_path.push(bucket.paths.install().guest.join(compressed_file));
+
+    pub fn compress_file_zstd(path: PathBuf) -> Result<PathBuf, BoxError> {
+        let output_path = PathBuf::from(format!("{}.zst", path.display()));
+        let input = File::create(&output_path)?;
+        let mut reader = BufReader::new(File::open(&path)?);
+        let mut writer = BufWriter::new(input);
+
+        zstd::stream::copy_encode(&mut reader, &mut writer, 16)?;
+
+        writer.flush()?;
+
+        Ok(output_path)
+    }
+
+    Ok(Decision::ReplaceFile {
+        newpath: generated_path,
+    }
+    .into())
+}
diff --git a/crates/stone_recipe/src/lib.rs b/crates/stone_recipe/src/lib.rs
index b0565b609..d1d3cd07f 100644
--- a/crates/stone_recipe/src/lib.rs
+++ b/crates/stone_recipe/src/lib.rs
@@ -93,6 +93,8 @@ pub struct Options {
     pub strip: bool,
     #[serde(default, deserialize_with = "stringy_bool")]
     pub networking: bool,
+    #[serde(default, deserialize_with = "stringy_bool")]
+    pub compressman: bool,
 }
 
 #[derive(Debug, Clone, Deserialize)]