chgserver: discard buffered output before restoring fds (issue6207)

On Python 3, flush() appears not discarding buffered data on EPIPE, and
the buffered data will be carried over to the restored stdout.

mercurial/chgserver.py

@@ -434,8 +434,11 b' class chgcmdserver(commandserver.server)'
             self._oldios.append((ch, fp, fd))
 
     def _restoreio(self):
+        if not self._oldios:
+            return
+        nullfd = os.open(os.devnull, os.O_WRONLY)
         ui = self.ui
-        for (ch, fp, fd), (cn, fn, _mode) in zip(self._oldios, _iochannels):
+        for (ch, fp, fd), (cn, fn, mode) in zip(self._oldios, _iochannels):
             newfp = getattr(ui, fn)
             # close newfp while it's associated with client; otherwise it
             # would be closed when newfp is deleted
@@ -443,6 +446,12 b' class chgcmdserver(commandserver.server)'
                 newfp.close()
             # restore original fd: fp is open again
             try:
+                if newfp is fp and 'w' in mode:
+                    # Discard buffered data which couldn't be flushed because
+                    # of EPIPE. The data should belong to the current session
+                    # and should never persist.
+                    os.dup2(nullfd, fp.fileno())
+                    fp.flush()
                 os.dup2(fd, fp.fileno())
             except OSError as err:
                 # According to issue6330, running chg on heavy loaded systems
@@ -459,6 +468,7 b' class chgcmdserver(commandserver.server)'
             os.close(fd)
             setattr(self, cn, ch)
             setattr(ui, fn, fp)
+        os.close(nullfd)
         del self._oldios[:]
 
     def validate(self):

tests/test-chg.t

@@ -152,6 +152,49 b' chg waits for pager if runcommand raises'
   crash-pager: going to crash
   [255]
 
+no stdout data should be printed after pager quits, and the buffered data
+should never persist (issue6207)
+
+"killed!" may be printed if terminated by SIGPIPE, which isn't important
+in this test.
+
+  $ cat > $TESTTMP/bulkwrite.py <<'EOF'
+  > import time
+  > from mercurial import error, registrar
+  > cmdtable = {}
+  > command = registrar.command(cmdtable)
+  > @command(b'bulkwrite')
+  > def bulkwrite(ui, repo, *pats, **opts):
+  >     ui.write(b'going to write massive data\n')
+  >     ui.flush()
+  >     t = time.time()
+  >     while time.time() - t < 2:
+  >         ui.write(b'x' * 1023 + b'\n')  # will be interrupted by SIGPIPE
+  >     raise error.Abort(b"write() doesn't block")
+  > EOF
+
+  $ cat > $TESTTMP/fakepager.py <<'EOF'
+  > import sys
+  > import time
+  > sys.stdout.write('paged! %r\n' % sys.stdin.readline())
+  > time.sleep(1)  # new data will be written
+  > EOF
+
+  $ cat >> .hg/hgrc <<EOF
+  > [extensions]
+  > bulkwrite = $TESTTMP/bulkwrite.py
+  > EOF
+
+  $ chg bulkwrite --pager=on --color no --config ui.formatted=True
+  paged! 'going to write massive data\n'
+  killed! (?)
+  [255]
+
+  $ chg bulkwrite --pager=on --color no --config ui.formatted=True
+  paged! 'going to write massive data\n'
+  killed! (?)
+  [255]
+
   $ cd ..
 
 server lifecycle