On Wed, Dec 4, 2019 at 2:04 PM Nikolay Sivov nsivov@codeweavers.com wrote:
I don't think it makes sense to test for null explicitly. Any invalid char for given context would be a syntax error.
Hi Nikolay,
I agree that it should work that way. However, the tests show that it does not. My belief is that the stream is being treated as a string, so when a null character encountered, it is treating as marking the end of the stream, even though it may actually be a part of the stream. My patch makes sure that condition is being covered, at least for the cases that are being tested.
Thanks, Jeff