Skip test involving fork() on windows

[fingolfin]

This otherwise is far too slow on 64 bit windows.

Fixes #3506

[olexandr-konovalov]

Thanks, but this is not enough. As #3506 says, even the first iteration hangs it. You need also to tweak parameters in

children := List([1..20], x -> runChild(Random([1..2000]), Random([false,true])));;

What worked for me was simple runChild(1,false) and runChild(1,true).

[coveralls]

Coverage remained the same at 85.26% when pulling 6967c23 on fingolfin:mh/children into f7f36d7 on gap-system:master.

[codecov]

Codecov Report

Merging #3507 into master will increase coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #3507      +/-   ##
==========================================
+ Coverage   85.42%   85.43%   +0.01%     
==========================================
  Files         699      699              
  Lines      346778   346778              
==========================================
+ Hits       296229   296270      +41     
+ Misses      50549    50508      -41

Impacted Files	Coverage Δ
src/iostream.c	65.15% <0%> (-0.76%)	⬇️
src/stats.c	94.97% <0%> (-0.21%)	⬇️
src/opers.c	94.87% <0%> (+0.06%)	⬆️
src/listfunc.c	97.47% <0%> (+0.19%)	⬆️
src/weakptr.c	97.96% <0%> (+0.4%)	⬆️
src/stringobj.c	94.38% <0%> (+0.59%)	⬆️
src/streams.c	76.62% <0%> (+1.39%)	⬆️
src/sysfiles.c	41.39% <0%> (+1.86%)	⬆️

[fingolfin]

I've decided to simply don't run this part of the test file at all on Windows. It's IMHO just not worth the hassle to try to just launch 1-2 child processes: we already have other tests which do that, so we don't gain anything from it.

[olexandr-konovalov]

Ok! I guess I still can test #3513 where this test is not modified?

[fingolfin]

@alex-konovalov does your approval of this PR mean that you tested it on Windows / Cygwin and it resolved the issue there?

[olexandr-konovalov]

No - it's pretty obvious that this test should be excluded from Windows, so my approval is just based on inspecting the diff.

[fingolfin]

Argh, I confused the PRs, my bad :-(

[olexandr-konovalov]

Backported to stable-4.10 in 2dbf56e