Package: python-imaplib2 Version: 2.57-5 Severity: normal Tags: upstream I noticed this from trying "usecompression = yes" in offlineimap.
When fetching large messages over IMAP, it's possible to completely drain the system socket, but still have leftover data in userspace that the high-level .read methods may return. Having no data in the socket prevents the select()/poll() syscalls from waking up in the "_reader" event loop. I'm no Python hacker, however I am familiar with zlib and non-blocking sockets from C and other languages. I recommend trying to read() first, and then issue select()/poll() if read() fails with EAGAIN (or SSL_read() fails with SSL_WANT_READ/SSL_WANT_WRITE). Blindly doing select/poll with userspace buffer layers on top (OpenSSL, zlib, or both) is a common source of starvation bugs in my experience. Thank you, I'm an offlineimap user since 2003/2004 :) -- System Information: Debian Release: 10.4 APT prefers stable-updates APT policy: (500, 'stable-updates'), (500, 'stable-debug'), (500, 'oldstable-updates'), (500, 'stable'), (500, 'oldstable'), (1, 'experimental') Architecture: amd64 (x86_64) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) Versions of packages python-imaplib2 depends on: ii python 2.7.16-1 python-imaplib2 recommends no packages. python-imaplib2 suggests no packages. -- no debconf information
