generator: Remove generated and unused files from previous runs of the generator.

If you go back in time in git (eg. git reset, git bisect) then you can
end up in a situation where a file that was generated by a later
version is left around unused in the earlier version.

This isn't a problem for most things, but gobject documentation
generation picks up any file in a directory, even unreferenced ones,
and breaks.  So the correct thing to do is to remove these files.

3 files changed, 44 insertions, 1 deletions

diff --git a/generator/main.ml b/generator/main.ml
index 74419f10..4cb3c7c0 100644
--- a/generator/main.ml
+++ b/generator/main.ml
@@ -126,8 +126,11 @@ Run it from the top source directory using the command
     fun (typ, jtyp) ->
       let cols = cols_of_struct typ in
       let filename = sprintf "java/com/redhat/et/libguestfs/%s.java" jtyp in
-      output_to filename (generate_java_struct jtyp cols);
+      output_to filename (generate_java_struct jtyp cols)
   ) camel_structs;
+  delete_except_generated
+    ~skip:["java/com/redhat/et/libguestfs/LibGuestFSException.java"]
+    "java/com/redhat/et/libguestfs/*.java";
 
   output_to "java/Makefile.inc" generate_java_makefile_inc;
   output_to "java/com_redhat_et_libguestfs_GuestFS.c" generate_java_c;
@@ -156,6 +159,8 @@ Run it from the top source directory using the command
       let filename = sprintf "gobject/src/%s.c" short in
       output_to filename (generate_gobject_struct_source short typ cols)
   ) structs;
+  delete_except_generated "gobject/include/guestfs-gobject/struct-*.h";
+  delete_except_generated "gobject/src/struct-*.c";
 
   List.iter (
     function
@@ -170,6 +175,8 @@ Run it from the top source directory using the command
         (generate_gobject_optargs_source short name optargs f)
     | { style = _, _, [] } -> ()
   ) all_functions;
+  delete_except_generated "gobject/include/guestfs-gobject/optargs-*.h";
+  delete_except_generated "gobject/src/optargs-*.c";
 
   output_to "gobject/include/guestfs-gobject/tristate.h"
     generate_gobject_tristate_header;
diff --git a/generator/pr.ml b/generator/pr.ml
index fd927206..1c3812a1 100644
--- a/generator/pr.ml
+++ b/generator/pr.ml
@@ -31,6 +31,7 @@ let lines = ref 0
 
 (* Name of each file generated. *)
 let files = ref []
+let fileshash = Hashtbl.create 13
 
 (* Print-to-current-output function, used everywhere.  It has
  * printf-like semantics.
@@ -45,6 +46,7 @@ let pr fs =
 
 let output_to ?(perm = 0o444) filename k =
   files := filename :: !files;
+  Hashtbl.add fileshash filename ();
 
   let filename_new = filename ^ ".new" in
   chan := open_out filename_new;
@@ -63,6 +65,36 @@ let output_to ?(perm = 0o444) filename k =
     printf "written %s\n%!" filename;
   )
 
+let delete_except_generated ?(skip = []) glob =
+  let cmd = sprintf "ls -1 %s" glob in
+  let chan = open_process_in cmd in
+  let lines = ref [] in
+  let rec loop () =
+    lines := input_line chan :: !lines;
+    loop ()
+  in
+  let lines = try loop () with End_of_file -> List.rev !lines in
+  (match close_process_in chan with
+  | WEXITED 0 -> ()
+  | WEXITED i ->
+    failwithf "command exited with non-zero status (%d)" i
+  | WSIGNALED i | WSTOPPED i ->
+    failwithf "command signalled or stopped with non-zero status (%d)" i
+  );
+
+  (* Build the final skip hash. *)
+  let skiphash = Hashtbl.copy fileshash in
+  List.iter (fun filename -> Hashtbl.add skiphash filename ()) skip;
+
+  (* Remove the files. *)
+  List.iter (
+    fun filename ->
+      if not (Hashtbl.mem skiphash filename) then (
+        unlink filename;
+        printf "deleted %s\n%!" filename
+      )
+  ) lines
+
 let get_lines_generated () =
   !lines
 
diff --git a/generator/pr.mli b/generator/pr.mli
index 693c6813..e0a8227c 100644
--- a/generator/pr.mli
+++ b/generator/pr.mli
@@ -28,6 +28,10 @@ val output_to : ?perm:int -> string -> (unit -> unit) -> unit
     [filename] is only updated if the output is different from what
     is in the file already. *)
 
+val delete_except_generated : ?skip: string list -> string -> unit
+(** Remove files matching [glob], unless those files have been
+    generated (so far), OR match a name in the [~skip] list. *)
+
 val get_lines_generated : unit -> int
 (** Return number of lines of code generated. *)