On Thu, Oct 09, 2014 at 04:33:57AM +0200, Markus Teich wrote: > Silvan Jegen wrote: > > In this issue https://code.google.com/p/go/issues/detail?id=6007 in answer > > #30 > > they mention that they use a monotonic clock on Linux now with Go 1.3 and up > > which should solve this (this is the commit > > https://code.google.com/p/go/source/detail?r=79f855ac890d ). What platform > > and > > Go version are you using? > > Heyho, > > I'm using Go 1.3 on Gentoo Linux x86_64, so if I understand correctly, the > issue > shouldn't even occur on my machine?
I am somewhat unsure myself reading through the commit log message again. It says that this "lays the groundwork" to solve this issue and mentions that each platform and architecture's implementation of the internal runtime·nanotime() *should* be modified (with linux/x86 shown as the only modified and tested platform) to use monotonic time. The log short message says "runtime: use monotonic clock for timers (linux/386, linux/amd64)" however. Looking at the code changes, the assembler for runtime·nanotime() on amd64 has been changed as well though. Since AFAIK, if it is correctly implemented, your issue would not exist I think you should ask the golang-nuts[0] mailing list about the issue (maybe mentioning the commit and the issue 6007 as well). I can do it if you don't want to but I will be travelling (accidentally to dotGo in Paris :-) ) so it may take a while. [0] https://groups.google.com/forum/#!forum/golang-nuts
