There are 3(logchar type) x 4(charset type) x 2(unichar?) x 2(wchar_t?) x 2(prefer_boost?) x 2(qt_support?) =192 combinations of build options log4cxx provides of which only one is being tested.
What should we do with those? On Fri, Oct 28, 2022 at 12:23 PM Robert Middleton <rmiddle...@apache.org> wrote: > Upon further investigation, it appears as though it was the cache > functionality that broke. Updating from v2 to v3 seems to have fixed > the problem. It looks like when you updated vcpkg to the latest > version that fixed it for one build, probably because the yml file > changed or something? > > Anyway, it's fixed now. Are there any other configurations that we > should test for while I'm making updates? > > -Robert Middleton > > On Thu, Oct 27, 2022 at 1:37 AM Stephen Webb <swebb2...@gmail.com> wrote: > > > > It works using a more recent version of vcpkg - not sure what the issue > > with the old version is. > > > > On Thu, Oct 27, 2022 at 1:41 PM Robert Middleton <rmiddle...@apache.org> > > wrote: > > > > > I've been adding some more builds for log4cxx(to test more > > > combinations of features and stuff) but I seem to have broken the > > > windows build - cmake can't find APR with the .pc file. It seems that > > > whatever we had before cached had the .pc files in it, but trying to > > > rebuild clean(no cache) does not have .pc files in it to find the > > > library. > > > > > > I'm thinking that this may have had something to do with the recent > > > changes to how we find APR, but I'm not much of a Windows guy so it > > > could take me a while to figure out what exactly is broken. I know > > > that Steven Webb uses vcpkg, would you be able to take a look at it > > > and figure out why it might be failing? > > > > > > The branch in question: > > > https://github.com/apache/logging-log4cxx/tree/LOGCXX-562 > > > > > > -Robert Middleton > > > >