On 15/07/2014 7:53 PM, [email protected] wrote: > Am 15.07.2014 00:50, schrieb Ryan Kelly: >> On 14/07/2014 7:08 PM, [email protected] wrote: >>> No, this helped not - nothing changed. When I added the environment >>> variable PIP_DEFAULT_TIMEOUT of 60 seconds the timeout changed from 15 >>> seconds to about 20 seconds. So it's obvious that the timeout parameter >>> to pip has it's effect, but there must be another timeout of 20 seconds >>> (from which component?) that overrules the 60 seconds of pip. >>> >>> The proxy server is successfully used, I see that in the logilfe. >>> >>> Which component could overrule the 60 second timeout of pip with these >>> 20 seconds? >> I can't think of anything off the top of my head. Please attach the >> full output from the `make build` command and I'll dig through to see >> what I can find. > If done some more testing and found, that timeouts from 1 second up to > 20.9 seconds are working as expected. > > [..snip..] > > In most cases it fails on packet zope.interface, but sometimes on others. > > In a single case all downloads where successful but make failed because > of dependencies of g++. > > Is there an option to keep alreay downloaded files rsp. is there an > option that I must not call "make clean" in front of "make build"? > > If I don't do "make clean" the "make build" states "nothing to do" even > if "make build" failed before.
I've updated the makefile to allow this. If you try a `make build` and it fails, a subsequent `make build` should now pick up from where it left off. Hopefully this will help get you through the timeout issues. I can't think of another approach at the moment. Ryan _______________________________________________ Sync-dev mailing list [email protected] https://mail.mozilla.org/listinfo/sync-dev
