On 14.11.2013 15:21, Martin Sustrik wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 14/11/13 14:57, Peter Kümmel wrote:Googling for libev and nanamsg/zeromq mostly results in hits about using zeromq _with_ libev. But I wonder if nanomsg could not be used instead of libev as library which provides a message queue.Yes, it can, but the functionality provided is somehow different (roughly speaking, libev is a lower level library than nanomsg). Thus, check the API whether it allows you to do what you want.There is already the inproc transport. What is missing for full blown message queue support? Is there a technical difference to libev which would it make impossible to use nanomsg as back-end for a message queue?It depends on what you mean by "message queue". If it's only pushing messages from one side and retrieving it from the other side, you should be OK with both libev and nanomsg.
Thanks for the clarification. With "message queue" I mean the a message buffer on with sides. Peter
Would it be faster than libev? See http://libev.schmorp.de/bench.htmlBoth libev and nanomsg are pretty fast, but you have to test your use case to get any meaningful comparison. Martin -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJShNxzAAoJENTpVjxCNN9YUvkIAK3tMCT3MMLmFJyib6DVkpJq cZtIP1YBAWIVnzu/LcUcnys/8T0vDD6tV2wVTlB7E/N8oiNx2+Ui3Y9P/6SQ5i0A LwfnwGSDwrw5uInKz1/yXWF8/xufRgYjaUno+25siyVtPJn2o4TKGkXGBzz1cQkR qjlP1wgKVLE4AKgWc5uBNkhyIJ2WTYa3O7Ci0ExSu2GqfiUZ8hZRz1/MV/Jhx8ZH 19h3tpe7Ll9kkbpB1sAI7s5F3gz3CMaN/Up7TXMEKvNesh0oB/dR/oPv3razFNau 3ErV8g8U+1xecu9aEWaD0OIbwsWoB2jIoYGCEbI3Qmi4yYXeA8gpHbCVJQs8XBY= =6WhG -----END PGP SIGNATURE-----
