Nope, that unfortunately doesn't help - I get the same error. // T
On Tuesday, September 15, 2015 at 12:28:49 PM UTC+2, Tim Holy wrote: > > Not sure if this will fix it, but you seem to be delimiting the rows of > the > matrix with carriage-returns; try using a semicolon. > > --Tim > > On Tuesday, September 15, 2015 12:59:07 AM Tomas Lycken wrote: > > I’ve been getting test failures on Contour.jl > > <https://github.com/tlycken/Contour.jl> for a while now, but since they > > only happened on 0.4 they’ve been a bit down on my priority list. Now > that > > 0.4 is about to be released, it’s time to fix them. > > > > I’m having some trouble understanding what the problem is though; The > REPL > > borks with the same error as the test: > > > > julia> include("testdata.jl") > > ERROR: LoadError: OverflowError() > > in include at boot.jl:260 > > in include_from_node1 at loading.jl:271 > > while loading C:\Users\Tomas > > Lycken\AppData\Local\Julia-0.4.0-rc1\testdata.jl, in expression starting > on > > line 1 > > > > I’ve attached the file testdata.jl - it is simply an expression on the > form > > > > z = [#200 space-separated floats > > #200 space-separated floats > > #... > > ] > > > > defining a 165x200 Matrix{Float64}. The odd thing is that if I comment > out > > every line, and then uncomment one line at a time, the error invariably > > returns when the row count becomes larger than 40 (i.e. 40x200 works, > > 41x200 doesn’t). This is *regardless of which line is the 41st to be > > uncommented.* > > > > The file loads without problem on 0.3.11. > > > > Is this a regression in 0.4, or some changed behavior that I need to > adapt > > to? If the latter, how do I adapt? > > > > Thanks, > > > > Tomas > > > >