Coincidentally we had a failure in the same test on Linux today (after running it hundreds of times before without issue).
My colleague had the bright idea of running that test many times so that we could reliably reproduce the error on Linux. We then found the bug, which was in fact a bug in our code (not properly initialising a struct being passed to C). This is almost certainly what was causing the issue on Windows. So I think the problem is likely solved. Bill. On Sunday, 24 July 2016 23:35:42 UTC+2, Yichao Yu wrote: > > On Sun, Jul 24, 2016 at 5:22 PM, 'Bill Hart' via julia-users > <julia...@googlegroups.com <javascript:>> wrote: > > I built the dlls we make use of in our Nemo package a slightly odd way, > but > > everything worked, all tests passed. > > > > I decided not to be lazy and built the dlls the correct way, and all of > a > > sudden I get a ReadOnlyMemoryError() whilst running our test code. > > > > This is with either Julia 0.4.0 or 0.4.6 on Windows 64. > > > > depends22 shows no problems with the dlls and that they don't depend on > the > > msys dll. I'm quite sure I've built them correctly. > > > > Julia seems to think the dlls are valid, I just get that error. > > > > Does anyone have any insight into what this error is and what is causing > it? > > The error means there's some memory error. It's impossible to tell > what's causing it without more detail. > > > > > Bill. >
