thanks; for the record i found an open issue - https://github.com/JuliaLang/julia/issues/5222
On Saturday, 29 March 2014 11:46:50 UTC-3, Avik Sengupta wrote: > > I think the Pkg.clone(pwd()) is correct, it copies the current directory > (where travis has cloned your git repo) to the julia package directory. > Which is why your second travis run fails, since CRC now does not exist > within the package dir. > > The first run is failing when trying to spawn a julia-basic executable to > run the tests (this happens after a test/runtests.jl is successfully > detected). The path to julia has been defined as const JULIA = > joinpath(JULIA_HOME, "julia-basic"). > > Not sure why that wouldn't work on travis, but maybe its to do with how > the ubuntu package is installed? > > Regards > - > Avik > > On Saturday, 29 March 2014 12:36:35 UTC, andrew cooke wrote: >> >> >> [this is NOT about Pkg.test() being new - I am ONLY using julianighlies >> in the matrix] >> >> i recently started a new project using the auto-generated files. when i >> enable it on travis i see an error like this: >> >> https://travis-ci.org/andrewcooke/CRC.jl/builds/21822323 >> >> if i remove the Pkg.clone(pwd()) - which i don't understand the need for >> - i get >> >> https://travis-ci.org/andrewcooke/CRC.jl/builds/21822544 >> >> neither runs the tests, which exist and work just fine from the command >> line on my laptop (if i run julia -e 'Pkg.test("CRC")') >> >> the repo is at https://github.com/andrewcooke/CRC.jl and the travis.yml >> is >> >> language: cpp >> compiler: >> - clang >> notifications: >> email: false >> env: >> matrix: >> - JULIAVERSION="julianightlies" >> before_install: >> - sudo add-apt-repository ppa:staticfloat/julia-deps -y >> - sudo add-apt-repository ppa:staticfloat/${JULIAVERSION} -y >> - sudo apt-get update -qq -y >> - sudo apt-get install libpcre3-dev julia -y >> script: >> - julia -e 'Pkg.init(); Pkg.clone(pwd()); Pkg.test("CRC")' >> >> it seems like this should "just work". has anyone else seen this? i'd >> raise an issue, but i suspect i've messed up somehow.... >> >> thanks, >> andrew >> >> >> >> >> >> >> >> >> >> >>
