|
From: Gustaf N. <ne...@wu...> - 2016-12-05 14:33:16
|
Am 05.12.16 um 14:32 schrieb Lucio Andrés Illanes Albornoz: >> By using such a library, there is a hope to reduce the number >> of separate source code variants and ifdefs in the Tcl code base >> and reduce maintenance costs. > Wishful thinking. well, my point is: do we have to re-invent event library abstractions in Tcl, or can we use something, that does this already (e.g. libevent, libev, libuv). If our answer is yes, i am totally fine with this. But we should consider it. >> Addressing the shortcomings of select() is certainly worthwhile, >> but the question is whether we want to test all these variants >> on all platforms, with sometimes broken implementations > All? You mean FreeBSD, DragonFlyBSD, NetBSD, OpenBSD, Mac OS X, and Linux. > I'm open for suggestions. Not to mention Solaris, AIX and WIN32... and maybe Android. The event interfaces tend to be buggy, at least when it comes to details. If we can "outsource" some of these problems and the ongoing maintenance to the implementers of event libraries, we don't have to deal with the platform dependencies in the same degree within Tcl. My alarm shrilled when you mentioned that you added kqueue(2) support for Mac OS without the ability to test it. Read [1], and you will know why i wrote this mail. -g [1] http://pod.tst.eu/http://cvs.schmorp.de/libev/ev.pod#PORTABILITY_NOTES |