Support .gitattributes export-ignore in distribution archives

CHANGES.md

@@ -2,6 +2,8 @@
 
 ### Added
 
+- Support `.gitattributes` `export-ignore` attribute to exclude files from
+  distribution archives (#<PR_NUMBER>, @jberdine)
 - Add `dune-release delegate-info version` to show the current version as infered
   by the tool (#495, @samoht)
 - Add `--dev-repo` to `dune-release` and `dune-release publish` to overwrite

README.md

@@ -110,6 +110,31 @@ The full documentation of this command is available with
 dune-release help distrib
 ```
 
+#### Excluding files with .gitattributes
+
+Files marked with the `export-ignore` attribute in `.gitattributes` will be excluded from the distribution archive. This can be used to exclude development files like `dune-workspace` that should not be included in releases.
+
+Example `.gitattributes`:
+```
+dune-workspace export-ignore
+.github/** export-ignore
+```
+
+**Supported patterns:**
+- Exact filenames: `dune-workspace`
+- Directory patterns: `.github/**` (matches all files under `.github/`)
+- Glob patterns: `*.log`, `test_*`, `file?.txt`
+- Double star in path: `**/build`, `src/**/test.ml`
+- Path normalization: handles `./` and `../` in paths
+
+**Not supported:**
+- Negation patterns (`!pattern`)
+- Escaped patterns (`\!` for literal `!`)
+- Quoted patterns (`"a b"` for patterns with spaces)
+- Character classes (`[abc]`)
+- Case insensitivity (`core.ignorecase`)
+- `.gitattributes` files in subdirectories
+
 
 ### Publish the distribution online
 

lib/archive.ml

@@ -117,33 +117,45 @@ module Tar = struct
     String.concat (List.rev (end_of_file :: t))
 end
 
-let path_set_of_dir dir ~exclude_paths =
-  let not_excluded p = Ok (not (Fpath.Set.mem (Fpath.base p) exclude_paths)) in
+let path_set_of_dir dir ~exclude_paths ~export_ignore =
+  let not_excluded p =
+    if Fpath.Set.mem (Fpath.base p) exclude_paths then Ok false
+    else
+      match Fpath.rem_prefix dir p with
+      | None -> Ok true
+      | Some rel_path ->
+          Ok (not (List.exists (Gitattributes.matches rel_path) export_ignore))
+  in
   let traverse = `Sat not_excluded in
   let elements = `Sat not_excluded in
   let err _ e = e in
   OS.Dir.contents ~dotfiles:true dir
   >>= OS.Path.fold ~dotfiles:true ~err ~elements ~traverse Fpath.Set.add
         Fpath.Set.empty
 
-let tar dir ~exclude_paths ~root ~mtime =
+let tar dir ~exclude_paths ~export_ignore ~root ~mtime =
   let tar_add file tar =
     let fname =
       match Fpath.rem_prefix dir file with
       | None -> assert false
       | Some file -> Fpath.(root // file)
     in
-    Logs.info (fun m -> m "Archiving %a" Fpath.pp fname);
     tar >>= fun tar ->
     OS.Dir.exists file >>= function
-    | true -> Tar.add tar fname ~mode:0o775 ~mtime `Dir
+    | true ->
+        (* Skip directories - they will be created implicitly when their
+           contents are added. This ensures that directories whose contents
+           are excluded via export-ignore patterns don't appear as empty
+           directories in the archive. *)
+        Ok tar
     | false ->
+        Logs.info (fun m -> m "Archiving %a" Fpath.pp fname);
         OS.Path.Mode.get file >>= fun mode ->
         OS.File.read file >>= fun contents ->
         let mode = if 0o100 land mode > 0 then 0o775 else 0o664 in
         Tar.add tar fname ~mode ~mtime (`File contents)
   in
-  path_set_of_dir dir ~exclude_paths >>= fun fset ->
+  path_set_of_dir dir ~exclude_paths ~export_ignore >>= fun fset ->
   Fpath.Set.fold tar_add fset (Ok Tar.empty) >>| fun tar -> Tar.to_string tar
 
 (* Bzip2 compression and unarchiving *)

lib/archive.mli

@@ -13,16 +13,19 @@ open Bos_setup
 val tar :
   Fpath.t ->
   exclude_paths:Fpath.set ->
+  export_ignore:Gitattributes.pattern list ->
   root:Fpath.t ->
   mtime:int64 ->
   (string, R.msg) result
-(** [tar dir ~exclude_paths ~root ~mtime] is a (us)tar archive that contains the
-    file hierarchy [dir] except the relative hierarchies present in
-    [exclude_paths]. In the archive, members of [dir] are rerooted at [root] and
-    sorted according to {!Fpath.compare}. They have their modification time set
-    to [mtime] and their file permissions are [0o775] for directories and files
-    executable by the user and [0o664] for other files. No other file metadata
-    is preserved.
+(** [tar dir ~exclude_paths ~export_ignore ~root ~mtime] is a (us)tar archive
+    that contains the file hierarchy [dir] except:
+    - relative hierarchies present in [exclude_paths] (basename matching)
+    - files matching patterns in [export_ignore] (from [.gitattributes])
+
+    In the archive, members of [dir] are rerooted at [root] and sorted according
+    to {!Fpath.compare}. They have their modification time set to [mtime] and
+    their file permissions are [0o775] for directories and files executable by
+    the user and [0o664] for other files. No other file metadata is preserved.
 
     {b Note.} This is a pure OCaml implementation, no [tar] tool is needed. *)
 

lib/gitattributes.ml

@@ -0,0 +1,132 @@
+open Bos_setup
+
+type pattern =
+  | Exact of string  (** Exact match against basename or full path. *)
+  | Prefix of string
+      (** Pattern like [dir/**] that matches everything under a directory, but
+          not the directory itself. *)
+  | Glob of Re.re  (** Compiled glob pattern. *)
+
+(** [glob_to_re pattern] is a compiled regex for glob [pattern]. Supports [*]
+    (any chars except /), [?] (single char except /), and [**] (any path
+    segments, but only when adjacent to /). *)
+let glob_to_re pattern =
+  let buf = Buffer.create (String.length pattern * 2) in
+  Buffer.add_char buf '^';
+  let len = String.length pattern in
+  let rec loop i =
+    if i >= len then ()
+    else
+      let c = pattern.[i] in
+      match c with
+      | '*' ->
+          if i + 1 < len && pattern.[i + 1] = '*' then
+            (* ** only crosses path separators when adjacent to / *)
+            let preceded_by_slash = i > 0 && pattern.[i - 1] = '/' in
+            let at_start = i = 0 in
+            if i + 2 < len && pattern.[i + 2] = '/' then (
+              (* **/ matches zero or more directories *)
+              Buffer.add_string buf "(.*/)?";
+              loop (i + 3))
+            else if i + 2 >= len && (preceded_by_slash || at_start) then (
+              (* /** at end or just ** alone - matches anything *)
+              Buffer.add_string buf ".*";
+              loop (i + 2))
+            else (
+              (* ** not adjacent to / - acts like * *)
+              Buffer.add_string buf "[^/]*";
+              loop (i + 2))
+          else (
+            (* * matches anything except path separator *)
+            Buffer.add_string buf "[^/]*";
+            loop (i + 1))
+      | '?' ->
+          (* ? matches any single character except path separator *)
+          Buffer.add_string buf "[^/]";
+          loop (i + 1)
+      | '.' | '+' | '^' | '$' | '(' | ')' | '[' | ']' | '{' | '}' | '|' | '\\'
+        ->
+          (* Escape regex metacharacters *)
+          Buffer.add_char buf '\\';
+          Buffer.add_char buf c;
+          loop (i + 1)
+      | _ ->
+          (* Literal character *)
+          Buffer.add_char buf c;
+          loop (i + 1)
+  in
+  loop 0;
+  Buffer.add_char buf '$';
+  Re.Pcre.regexp (Buffer.contents buf)
+
+let parse_pattern s =
+  let s = String.trim s in
+  (* Remove leading slash if present - we always match relative paths *)
+  let s =
+    if String.is_prefix ~affix:"/" s then
+      String.Sub.to_string (String.sub ~start:1 s)
+    else s
+  in
+  let has_wildcard s = String.exists (fun c -> c = '*' || c = '?') s in
+  if String.is_suffix ~affix:"/**" s then
+    (* Directory pattern: match everything under the directory *)
+    let prefix =
+      String.Sub.to_string (String.sub ~stop:(String.length s - 3) s)
+    in
+    if has_wildcard prefix then
+      (* Prefix contains wildcards, treat whole pattern as glob *)
+      Glob (glob_to_re s)
+    else Prefix prefix
+  else if has_wildcard s then
+    (* Has wildcards - compile as glob *)
+    Glob (glob_to_re s)
+  else
+    (* Exact match *)
+    Exact s
+
+let matches path pattern =
+  let path = Fpath.normalize path in
+  let path_str = Fpath.to_string path in
+  let basename = Fpath.basename path in
+  match pattern with
+  | Exact s ->
+      (* Match against basename or full relative path *)
+      String.equal s basename || String.equal s path_str
+  | Prefix prefix ->
+      (* Match everything under the directory, but not the directory itself *)
+      String.is_prefix ~affix:(prefix ^ "/") path_str
+  | Glob re ->
+      (* Match against full path or basename for patterns like *.log *)
+      Re.execp re path_str || Re.execp re basename
+
+let parse_export_ignore content =
+  (* Strip UTF-8 BOM if present at start of file *)
+  let content =
+    if String.is_prefix ~affix:"\xef\xbb\xbf" content then
+      String.Sub.to_string (String.sub ~start:3 content)
+    else content
+  in
+  content |> String.cuts ~sep:"\n"
+  |> List.filter_map (fun line ->
+         let line = String.trim line in
+         (* Skip empty lines and comments *)
+         if String.length line = 0 || String.is_prefix ~affix:"#" line then None
+         else
+           (* Format: <pattern> <attr1> <attr2> ...
+              Attributes can be separated by spaces or tabs *)
+           let parts =
+             String.fields ~empty:false
+               ~is_sep:(fun c -> c = ' ' || c = '\t')
+               line
+           in
+           match parts with
+           | pattern :: attrs
+             when List.exists (String.equal "export-ignore") attrs ->
+               Some (parse_pattern pattern)
+           | _ -> None)
+
+let read_export_ignore dir =
+  let file = Fpath.(dir / ".gitattributes") in
+  OS.File.exists file >>= function
+  | false -> Ok []
+  | true -> OS.File.read file >>| parse_export_ignore

lib/gitattributes.mli

@@ -0,0 +1,32 @@
+(** Gitattributes parsing for export-ignore.
+
+    Parses [.gitattributes] files and extracts patterns marked with the
+    [export-ignore] attribute. These patterns can be used to exclude files from
+    distribution archives. *)
+
+open Bos_setup
+
+(** {1 Patterns} *)
+
+type pattern
+(** The type for gitattributes patterns. *)
+
+val parse_pattern : string -> pattern
+(** [parse_pattern s] is the pattern parsed from string [s]. Supports:
+    - Exact matches: [filename]
+    - Directory patterns: [dir/**]
+    - Glob patterns: [*.ext], [prefix*] *)
+
+val matches : Fpath.t -> pattern -> bool
+(** [matches path pattern] holds if [path] matches [pattern]. [path] should be
+    relative to the repository root. *)
+
+(** {1 Parsing .gitattributes} *)
+
+val parse_export_ignore : string -> pattern list
+(** [parse_export_ignore content] is the list of patterns marked with
+    [export-ignore] in [.gitattributes] file [content]. *)
+
+val read_export_ignore : Fpath.t -> (pattern list, R.msg) result
+(** [read_export_ignore dir] is the list of patterns marked with [export-ignore]
+    in [dir/.gitattributes], or the empty list if the file doesn't exist. *)

lib/pkg.ml

@@ -465,7 +465,9 @@ let distrib_archive ~dry_run ~keep_dir ~include_submodules p =
   >>= fun () ->
   distrib_prepare ~dry_run ~dist_build_dir ~version >>= fun () ->
   let exclude_paths = Fpath.Set.of_list Distrib.exclude_paths in
-  Archive.tar dist_build_dir ~exclude_paths ~root ~mtime >>= fun tar ->
+  Gitattributes.read_export_ignore dist_build_dir >>= fun export_ignore ->
+  Archive.tar dist_build_dir ~exclude_paths ~export_ignore ~root ~mtime
+  >>= fun tar ->
   distrib_archive_path p >>= fun archive ->
   Archive.bzip2 ~dry_run ~force:true ~dst:archive tar >>= fun () ->
   (if keep_dir then Ok () else Sos.delete_dir ~dry_run dist_build_dir)