From 024303592a4e8c0e9a0b0ab40948571529da7471 Mon Sep 17 00:00:00 2001
From: ed <s@ocv.me>
Date: Wed, 10 Jan 2024 23:59:43 +0000
Subject: [PATCH] improved logging when a client dies mid-POST;

igloo irc has an absolute time limit of 2 minutes before it just
disconnects mid-upload and that kinda looked like it had a buggy
multipart generator instead of just being funny

anticipating similar events in the future, also log the
client-selected boundary value to eyeball its yoloness
---
 copyparty/httpcli.py |  9 +++++++++
 copyparty/util.py    | 24 +++++++++++++++++-------
 2 files changed, 26 insertions(+), 7 deletions(-)

diff --git a/copyparty/httpcli.py b/copyparty/httpcli.py
index 8c27a556..bba6ff96 100644
--- a/copyparty/httpcli.py
+++ b/copyparty/httpcli.py
@@ -2349,6 +2349,7 @@ class HttpCli(object):
         files: list[tuple[int, str, str, str, str, str]] = []
         # sz, sha_hex, sha_b64, p_file, fname, abspath
         errmsg = ""
+        tabspath = ""
         dip = self.dip()
         t0 = time.time()
         try:
@@ -2446,6 +2447,8 @@ class HttpCli(object):
                     if not nullwrite:
                         atomic_move(tabspath, abspath)
 
+                    tabspath = ""
+
                     files.append(
                         (sz, sha_hex, sha_b64, p_file or "(discarded)", fname, abspath)
                     )
@@ -2491,6 +2494,12 @@ class HttpCli(object):
             errmsg = vol_san(
                 list(self.asrv.vfs.all_vols.values()), unicode(ex).encode("utf-8")
             ).decode("utf-8")
+            try:
+                got = bos.path.getsize(tabspath)
+                t = "connection lost after receiving %s of the file"
+                self.log(t % (humansize(got),), 3)
+            except:
+                pass
 
         td = max(0.1, time.time() - t0)
         sz_total = sum(x[0] for x in files)
diff --git a/copyparty/util.py b/copyparty/util.py
index b7c93568..a9697ef3 100644
--- a/copyparty/util.py
+++ b/copyparty/util.py
@@ -623,9 +623,14 @@ class _Unrecv(object):
             while nbytes > len(ret):
                 ret += self.recv(nbytes - len(ret))
         except OSError:
-            t = "client only sent {} of {} expected bytes".format(len(ret), nbytes)
-            if len(ret) <= 16:
-                t += "; got {!r}".format(ret)
+            t = "client stopped sending data; expected at least %d more bytes"
+            if not ret:
+                t = t % (nbytes,)
+            else:
+                t += ", only got %d"
+                t = t % (nbytes, len(ret))
+                if len(ret) <= 16:
+                    t += "; %r" % (ret,)
 
             if raise_on_trunc:
                 raise UnrecvEOF(5, t)
@@ -1517,15 +1522,20 @@ class MultipartParser(object):
         return ret
 
     def parse(self) -> None:
+        boundary = get_boundary(self.headers)
+        self.log("boundary=%r" % (boundary,))
+
         # spec says there might be junk before the first boundary,
         # can't have the leading \r\n if that's not the case
-        self.boundary = b"--" + get_boundary(self.headers).encode("utf-8")
+        self.boundary = b"--" + boundary.encode("utf-8")
 
         # discard junk before the first boundary
         for junk in self._read_data():
-            self.log(
-                "discarding preamble: [{}]".format(junk.decode("utf-8", "replace"))
-            )
+            if not junk:
+                continue
+
+            jtxt = junk.decode("utf-8", "replace")
+            self.log("discarding preamble |%d| %r" % (len(junk), jtxt))
 
         # nice, now make it fast
         self.boundary = b"\r\n" + self.boundary