pull: always extract tars when archive pulling, merging dirs

src/pull.rs

@@ -1,6 +1,7 @@
 use std::cmp::min;
 use std::error::Error;
 use std::path::Path;
+use std::path::PathBuf;
 
 use crate::package_info::PackageInfo;
 
@@ -110,23 +111,71 @@ fn copy_dir_all(src: &Path, dst: &Path) -> Result<(), Box<dyn Error>> {
     Ok(())
 }
 
-fn extract_archive(path: &Path, dest: &Path) -> Result<(), Box<dyn Error>> {
-    let file = File::open(path)?;
+/// Helper function to extract tar archive with progress tracking
+fn extract_tar_archive<D, F>(
+    file_path: &Path,
+    dest: &Path,
+    progress: ProgressCallback<'_>,
+    decoder_factory: F,
+) -> Result<Vec<String>, Box<dyn Error>>
+where
+    D: std::io::Read,
+    F: Fn(File) -> D,
+{
+    let file = File::open(file_path)?;
+    let decoder = decoder_factory(file);
+    let mut archive = Archive::new(decoder);
+
+    // Get total number of entries for progress tracking
+    let total_entries = archive.entries()?.count();
+    let mut current_entry = 0;
+
+    // Reset the archive to read entries again
+    let file = File::open(file_path)?;
+    let decoder = decoder_factory(file);
+    let mut archive = Archive::new(decoder);
+
+    let mut extracted_files = Vec::new();
+
+    for entry in archive.entries()? {
+        let mut entry = entry?;
+        let path = entry.path()?.to_path_buf();
+        let dest_path = dest.join(&path);
+
+        // Create parent directories if needed
+        if let Some(parent) = dest_path.parent() {
+            std::fs::create_dir_all(parent)?;
+        }
+
+        // Extract the file
+        entry.unpack(&dest_path)?;
+        extracted_files.push(dest_path.to_string_lossy().to_string());
+
+        current_entry += 1;
+
+        // Report progress
+        if let Some(cb) = progress {
+            cb("", "Extracting...", current_entry, total_entries);
+        }
+    }
+
+    Ok(extracted_files)
+}
+
+fn extract_archive(
+    path: &Path,
+    dest: &Path,
+    progress: ProgressCallback<'_>,
+) -> Result<Vec<String>, Box<dyn Error>> {
     let filename = path.file_name().unwrap().to_string_lossy();
 
     if filename.ends_with(".tar.gz") || filename.ends_with(".tgz") {
-        let tar = GzDecoder::new(file);
-        let mut archive = Archive::new(tar);
-        archive.unpack(dest)?;
+        extract_tar_archive(path, dest, progress, GzDecoder::new)
     } else if filename.ends_with(".tar.xz") || filename.ends_with(".txz") {
-        let tar = XzDecoder::new(file);
-        let mut archive = Archive::new(tar);
-        archive.unpack(dest)?;
+        extract_tar_archive(path, dest, progress, XzDecoder::new)
     } else {
-        return Err(format!("Unsupported archive format: {}", filename).into());
+        Err(format!("Unsupported archive format: {}", filename).into())
     }
-
-    Ok(())
 }
 
 fn checkout_pristine_tar(package_dir: &Path, filename: &str) -> Result<(), Box<dyn Error>> {
@@ -326,59 +375,45 @@ async fn fetch_archive_sources(
     for file in &info.stanza.files {
         let url = format!("{}/{}", info.archive_url, file.name);
         download_file_checksum(&url, &file.sha256, package_dir, progress).await?;
-    }
 
-    // Extract the debian tarball or diff
-    let debian_file = info
-        .stanza
-        .files
-        .iter()
-        .find(|f| f.name.contains(".debian.tar.") || f.name.contains(".diff.gz"));
-
-    if let Some(file) = debian_file {
+        // Extract all tar archives, merging extracted directories
+        if file.name.contains(".tar.") {
             let path = package_dir.join(&file.name);
             let extract_dir = package_dir.join(&info.stanza.package);
 
-        if (file.name.ends_with(".tar.xz") || file.name.ends_with(".tar.gz"))
-            && let Err(e) = extract_archive(&path, &extract_dir)
-        {
-            return Err(format!("Failed to extract {}: {}", file.name, e).into());
+            let extracted = extract_archive(&path, &extract_dir, progress)?;
+
+            // Special case: the debian tar does only contain 'debian'
+            if file.name.contains("debian.tar.") {
+                continue;
             }
 
-        // Remove archive after extraction
-        std::fs::remove_file(&path)?;
+            // List root directories extracted and use the first one as the source directory
+            debug!("Root directories extracted:");
+            let mut source_dir: Option<PathBuf> = None;
+            for file in &extracted {
+                let path = Path::new(file);
+                // Check if this is a directory and is at the archive root level
+                // (i.e., the path relative to extract_dir has no parent components)
+                if let Ok(relative_path) = path.strip_prefix(&extract_dir)
+                    && relative_path.components().count() == 1
+                    && path.is_dir()
+                {
+                    debug!("- {}", relative_path.file_name().unwrap().to_string_lossy());
+                    // Use the first directory found as the source
+                    if source_dir.is_none() {
+                        source_dir = Some(path.to_path_buf());
+                    }
+                }
             }
 
-    // Extract the orig tarball if present and not native package
-    if !info.is_native()
-        && let Some(orig_file) = info
-            .stanza
-            .files
-            .iter()
-            .find(|f| f.name.contains(".orig.tar."))
-    {
-        let path = package_dir.join(&orig_file.name);
-        let extract_dir = package_dir;
-
-        if (orig_file.name.ends_with(".tar.xz") || orig_file.name.ends_with(".tar.gz"))
-            && let Err(e) = extract_archive(&path, extract_dir)
-        {
-            return Err(format!("Failed to extract {}: {}", orig_file.name, e).into());
-        }
-
-        // Rename from 'package-origversion' to 'package', merging with existing directory
+            // Use the extracted directory as the source, assuming there is only one
+            if let Some(src_dir) = source_dir {
                 let target_dir = package_dir.join(&info.stanza.package);
-        let entries = std::fs::read_dir(package_dir)?;
-        for entry in entries {
-            let entry = entry?;
-            let entry_path = entry.path();
-            if entry_path.is_dir() {
-                let dir_name = entry.file_name().to_string_lossy().to_string();
-                if dir_name.starts_with(&format!("{}-", info.stanza.package)) {
-                    // Found a directory like 'package-version', rename it to 'package'
+
                 if target_dir.exists() {
                     // Target exists, we need to merge contents
-                        for sub_entry in std::fs::read_dir(&entry_path)? {
+                    for sub_entry in std::fs::read_dir(&src_dir)? {
                         let sub_entry = sub_entry?;
                         let sub_path = sub_entry.path();
                         let target_path = target_dir.join(sub_entry.file_name());
@@ -390,11 +425,9 @@ async fn fetch_archive_sources(
                             std::fs::copy(&sub_path, &target_path)?;
                         }
                     }
-                        std::fs::remove_dir_all(&entry_path)?;
+                    std::fs::remove_dir_all(&src_dir)?;
                 } else {
-                        std::fs::rename(&entry_path, &target_dir)?;
-                    }
-                    break;
+                    std::fs::rename(&src_dir, &target_dir)?;
                 }
             }
         }