Run file_data_filter on deleted files

The `file_data_filter` method should be called when files are deleted.
In this case the `data` and `file_ctx` keys map to None. This is so
that a filter which modifies file names can apply the same name
transformations before files are deleted.

README.md

@@ -179,7 +179,7 @@ values in the dictionary after filters have been run are used to create the git
 commit.
 
 ```
-file_data = {'filename':filename,'file_ctx':file_ctx,'d':d}
+file_data = {'filename':filename,'file_ctx':file_ctx,'data':file_contents}
 
 def file_data_filter(self,file_data):
 ```
@@ -189,6 +189,11 @@ can be modified by any filter. `file_ctx` is the filecontext from the
 mercurial python library.  After all filters have been run, the values
 are used to add the file to the git commit.
 
+The `file_data_filter` method is also called when files are deleted,
+but in this case the `data` and `file_ctx` keys map to None. This is
+so that a filter which modifies file names can apply the same name
+transformations when files are deleted.
+
 Submodules
 ----------
 See README-SUBMODULES.md for how to convert subrepositories into git

hg-fast-export.py

@@ -311,9 +311,18 @@ def export_commit(ui,repo,revision,old_marks,max,count,authors,
     % (branch, type.encode(), revision + 1, max, len(modified), len(removed))
   )
 
-  for filename in removed:
+  for file in removed:
     if fn_encoding:
-      filename=filename.decode(fn_encoding).encode('utf8')
+      filename=file.decode(fn_encoding).encode('utf8')
+    else:
+      filename=file
+
+    if plugins and plugins['file_data_filters']:
+      file_data = {'filename':filename, 'file_ctx':None, 'data':None}
+      for filter in plugins['file_data_filters']:
+        filter(file_data)
+      filename=file_data['filename']
+
     filename=strip_leading_slash(filename)
     if filename==b'.hgsub':
       remove_gitmodules(ctx)

t/file_data_filter.expected

@@ -0,0 +1,27 @@
+blob
+mark :1
+data 7
+a_file
+
+blob
+mark :2
+data 17
+a_file_to_rename
+
+reset refs/heads/master
+commit refs/heads/master
+mark :3
+author Grevious Bodily Harmsworth <gbh@example.com> 1679014800 +0000
+committer Grevious Bodily Harmsworth <gbh@example.com> 1679014800 +0000
+data 2
+r0M 100644 :1 a.txt
+M 100644 :2 c.txt
+
+commit refs/heads/master
+mark :4
+author Grevious Bodily Harmsworth <gbh@example.com> 1679018400 +0000
+committer Grevious Bodily Harmsworth <gbh@example.com> 1679018400 +0000
+data 2
+r1from :3
+D c.txt
+

t/file_data_filter.t

@@ -0,0 +1,81 @@
+#!/bin/bash
+#
+# Copyright (c) 2023 Felipe Contreras
+# Copyright (c) 2023 Frej Drejhammar
+#
+# Check that the file_data_filter is called for removed files.
+#
+
+test_description='Smoke test'
+
+. "${SHARNESS_TEST_SRCDIR-$(dirname "$0")/sharness}"/sharness.sh || exit 1
+
+check() {
+	echo "$3" > expected &&
+	git -C "$1" show -q --format='%s' "$2" > actual &&
+	test_cmp expected actual
+}
+
+git_create() {
+	git init -q "$1"
+}
+
+git_convert() {
+	(
+	cd "$2" &&
+	hg-fast-export.sh --repo "../$1" \
+			  -s --hgtags -n \
+			  --plugin ../../plugins/rename_file_test_plugin
+	)
+}
+
+setup() {
+	cat > "$HOME"/.hgrc <<-EOF
+	[ui]
+	username = Grevious Bodily Harmsworth <gbh@example.com>
+	EOF
+}
+
+commit0() {
+	(
+	cd hgrepo &&
+	echo "a_file" > a.txt &&
+	echo "a_file_to_rename" > b.txt &&
+	hg add a.txt b.txt &&
+	hg commit -d "2023-03-17 01:00Z" -m "r0"
+	)
+}
+
+commit1() {
+	(
+	cd hgrepo &&
+	hg remove b.txt &&
+	hg commit -d "2023-03-17 02:00Z" -m "r1"
+	)
+}
+make-branch() {
+    hg branch "$1"
+    FILE=$(echo "$1" | sha1sum | cut -d " " -f 1)
+    echo "$1" > $FILE
+    hg add $FILE
+    hg commit -d "2023-03-17 $2:00Z" -m "Added file in branch $1"
+}
+
+setup
+
+test_expect_success 'all in one' '
+	test_when_finished "rm -rf hgrepo gitrepo" &&
+
+	(
+	hg init hgrepo &&
+	commit0 &&
+	commit1
+	) &&
+	git_create gitrepo &&
+	git_convert hgrepo gitrepo &&
+	git -C gitrepo fast-export --all > actual &&
+
+ 	test_cmp "$SHARNESS_TEST_DIRECTORY"/file_data_filter.expected actual
+'
+
+test_done

t/plugins/rename_file_test_plugin/__init__.py

@@ -0,0 +1,15 @@
+import subprocess
+import shlex
+import sys
+from mercurial import node
+
+def build_filter(args):
+    return Filter(args)
+
+class Filter:
+    def __init__(self, args):
+        self.filter_contents = shlex.split(args)
+
+    def file_data_filter(self,file_data):
+        if file_data['filename'] == b'b.txt':
+            file_data['filename'] = b'c.txt'