Github user mike-jumper commented on a diff in the pull request:
https://github.com/apache/guacamole-server/pull/183#discussion_r216996138
--- Diff: src/libguac/parser.c ---
@@ -222,7 +222,11 @@ int guac_parser_read(guac_parser* parser, guac_socket*
socket, int usec_timeout)
retval = guac_socket_select(socket, usec_timeout);
if (retval <= 0)
return -1;
-
+
+ /* Reset pointers if instruction buf len is less than max
instruction len */
+ if (buffer_end - unparsed_end < GUAC_INSTRUCTION_MAX_LENGTH)
+ unparsed_end = unparsed_start = parser->__instructionbuf;
--- End diff --
Here, I've added debug logging for each block written by `write()` within
`__write_all()` as well as each block read via `guac_socket_read()` within the
parser:
https://github.com/mike-jumper/guacamole-server/tree/debug-io
The logging added above should be stable in the sense that:
1. Reads and writes are written to separate files and thus the log
shouldn't become corrupt if they happen to overlap.
2. The log routine does not expect a null terminator, relying instead on a
supplied length, so there shouldn't be garbage / stale bytes in the logs that
weren't actually read/written at the time logged.
Reads are logged to `read-log.txt` within the current directory of guacd
and writes are logged to `write-log.txt` within the current directory of guacd.
The logged data is in hex dump format and will generally look like this:
```
--- 25 bytes:
00000000: 34 2E 73 69 7A 65 2C 34 2E 31 31 35 34 2C 33 2E 4.size,4.1154,3.
00000010: 36 35 34 2C 32 2E 39 36 3B 654,2.96;
--- 209 bytes:
00000000: 35 2E 61 75 64 69 6F 2C 38 2E 61 75 64 69 6F 2F 5.audio,8.audio/
00000010: 4C 38 2C 39 2E 61 75 64 69 6F 2F 4C 31 36 3B 35 L8,9.audio/L16;5
00000020: 2E 76 69 64 65 6F 3B 35 2E 69 6D 61 67 65 2C 31 .video;5.image,1
00000030: 30 2E 69 6D 61 67 65 2F 6A 70 65 67 2C 39 2E 69 0.image/jpeg,9.i
00000040: 6D 61 67 65 2F 70 6E 67 2C 31 30 2E 69 6D 61 67 mage/png,10.imag
00000050: 65 2F 77 65 62 70 3B 37 2E 63 6F 6E 6E 65 63 74 e/webp;7.connect
00000060: 2C 39 2E 6C 6F 63 61 6C 68 6F 73 74 2C 34 2E 35 ,9.localhost,4.5
00000070: 39 30 31 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 901,0.,0.,0.,0.,
00000080: 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 0.,0.,0.,0.,0.,0
00000090: 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E .,0.,0.,0.,0.,0.
000000A0: 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C ,0.,0.,0.,0.,0.,
000000B0: 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 0.,0.,0.,0.,0.,0
000000C0: 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E 2C 30 2E .,0.,0.,0.,0.,0.
000000D0: 3B ;
--- 30 bytes:
00000000: 39 2E 63 6C 69 70 62 6F 61 72 64 2C 31 2E 30 2C 9.clipboard,1.0,
00000010: 31 30 2E 74 65 78 74 2F 70 6C 61 69 6E 3B 10.text/plain;
--- 76 bytes:
00000000: 33 2E 65 6E 64 2C 31 2E 30 3B 35 2E 61 75 64 69 3.end,1.0;5.audi
00000010: 6F 2C 31 2E 30 2C 33 31 2E 61 75 64 69 6F 2F 4C o,1.0,31.audio/L
00000020: 31 36 3B 72 61 74 65 3D 34 34 31 30 30 2C 63 68 16;rate=44100,ch
00000030: 61 6E 6E 65 6C 73 3D 32 3B 34 2E 73 79 6E 63 2C annels=2;4.sync,
00000040: 39 2E 38 33 30 34 34 37 36 30 32 3B 9.830447602;
...
```
This should hopefully give us cleaner, more thorough logs.
Please try to reproduce the parse error with the above branch and provide
the resulting `write-log.txt` and `read-log.txt` for the connection that
failed. Be sure to remove the `write-log.txt` and `read-log.txt` files between
connection attempts, as the debug code just repeatedly appends to those files.
If the files exist at the beginning of a connection, the old log data will
still be there.
---
