Fixes when working with utf-8 content

When using set_payload(), we should pass along the charset as well,
otherwise we run into trouble when converting to/from bytes again.

Signed-off-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>

2 files changed, 10 insertions, 4 deletions

diff --git a/b4/__init__.py b/b4/__init__.py
index b5f2aa6..6af2acd 100644
--- a/b4/__init__.py
+++ b/b4/__init__.py
@@ -2431,10 +2431,12 @@ def git_range_to_patches(gitdir: Optional[str], start: str, end: str,
     if not commits:
         raise RuntimeError(f'Could not run rev-list {start}..{end}')
     for commit in commits:
-        ecode, out = git_run_command(gitdir, ['show', '--format=email', commit], decode=False)
+        ecode, out = git_run_command(gitdir, ['show', '--format=email', '--encoding=utf-8', commit], decode=False)
         if ecode > 0:
             raise RuntimeError(f'Could not get a patch out of {commit}')
         msg = email.message_from_bytes(out)
+        msg.set_charset('utf-8')
+        msg.replace_header('Content-Transfer-Encoding', '8bit')
         logger.debug('  %s', msg.get('Subject'))
 
         patches.append((commit, msg))
@@ -2462,6 +2464,7 @@ def git_range_to_patches(gitdir: Optional[str], start: str, end: str,
             # Move the original From and Date into the body
             origfrom = msg.get('From')
             if origfrom:
+                origfrom = LoreMessage.clean_header(origfrom)
                 origpair = email.utils.parseaddr(origfrom)
                 if origpair[1] != mailfrom[1]:
                     msg.replace_header('From', format_addrs([mailfrom]))
@@ -2484,7 +2487,7 @@ def git_range_to_patches(gitdir: Optional[str], start: str, end: str,
             payload = '\n'.join(inbodyhdrs) + '\n\n' + payload
         if not payload.find('\n-- \n') > 0:
             payload += f'\n-- \nb4 {__VERSION__}\n'
-        msg.set_payload(payload)
+        msg.set_payload(payload, charset='utf-8')
 
         if extrahdrs is None:
             extrahdrs = list()
diff --git a/b4/submit.py b/b4/submit.py
index b82b022..0834ff8 100644
--- a/b4/submit.py
+++ b/b4/submit.py
@@ -325,12 +325,15 @@ def update_trailers(cover_commit: str, cmdargs: argparse.Namespace) -> None:
     logger.info('Calculating patch-ids from %s commits', len(patches)-1)
     msg_map = dict()
     commit_map = dict()
+    updates = dict()
     # Ignore the cover letter
     for commit, msg in patches[1:]:
         body = msg.get_payload()
         patchid = b4.LoreMessage.get_patch_id(body)
         msg_map[patchid] = msg
         commit_map[patchid] = commit
+        if signoff and f'{signoff[0]}: <{signoff[1]}>' not in body:
+            updates[patchid] = list()
 
     if cmdargs.thread_msgid:
         cmdargs.msgid = cmdargs.thread_msgid
@@ -348,7 +351,6 @@ def update_trailers(cover_commit: str, cmdargs: argparse.Namespace) -> None:
     for list_msg in list_msgs:
         bbox.add_message(list_msg)
 
-    updates = dict()
     lser = bbox.get_series(sloppytrailers=cmdargs.sloppytrailers)
     mismatches = list(lser.trailer_mismatches)
     for lmsg in lser.patches[1:]:
@@ -476,7 +478,7 @@ def send(cover_commit: str, cmdargs: argparse.Namespace) -> None:
     }
     body = Template(cover_template.lstrip()).safe_substitute(tptvals)
     cmsg = email.message.EmailMessage()
-    cmsg.set_payload(body)
+    cmsg.set_payload(body, charset='utf-8')
     cmsg.add_header('Subject', csubject)
     if cmdargs.prefixes:
         prefixes = list(cmdargs.prefixes)
@@ -721,6 +723,7 @@ def main(cmdargs: argparse.Namespace) -> None:
 
     if cmdargs.new_series_name:
         start_new_series(cmdargs)
+        return
 
     if not check_our_branch():
         return