This PR updates the dependencies of slicec-cs, including the latest slicec.
The only change of relevance in slicec was changing how the SliceFiles are stored in a CompilationState.
This let us perform some slight simplifications here in slicec-cs.
This PR also fixes the compile_all_test_slice test, which has probably been broken for a very long time.
1) We moved where the test files were stored from IceRpc.Tests to both IceRpc.Slice.Tests and Zeroc.Slice.Tests.
So, this test was 'passing' because it found 0 files, and so, obviously number_of_errors == 0... which is what the test checked.
2) We were doing something weird/wrong with the slice_options, because instead of creating a new one for each test-run. We were modifying an existing one in place. So instead of only compiling 1 test file, we were compiling N. So after fixing issue 1, everything failed.
I fixed the test so everything is much more isolated now. We only compile a single source file at a time.
And we assert that the parsed version of the file has the same path as the one we expect to be testing.
This PR updates the dependencies of
slicec-cs
, including the latestslicec
.The only change of relevance in
slicec
was changing how theSliceFile
s are stored in aCompilationState
. This let us perform some slight simplifications here inslicec-cs
.This PR also fixes the
compile_all_test_slice
test, which has probably been broken for a very long time. 1) We moved where the test files were stored fromIceRpc.Tests
to bothIceRpc.Slice.Tests
andZeroc.Slice.Tests
. So, this test was 'passing' because it found0
files, and so, obviouslynumber_of_errors == 0
... which is what the test checked.2) We were doing something weird/wrong with the slice_options, because instead of creating a new one for each test-run. We were modifying an existing one in place. So instead of only compiling 1 test file, we were compiling
N
. So after fixing issue 1, everything failed. I fixed the test so everything is much more isolated now. We only compile a single source file at a time. And we assert that the parsed version of the file has the same path as the one we expect to be testing.