On Tue, Dec 18, 2018 at 11:33 PM Alex Henrie alexhenrie24@gmail.com wrote:
On Tue, Dec 18, 2018 at 11:11 PM Nikolay Sivov nsivov@codeweavers.com wrote:
This will suppress failures, but what we need is to understand how it works and adjust implementation.
The invalid codepoint produced in the Chinese locale is U+FA0F. ITextStream_Read lets it through unmodified, and ITextStream_ReadAll drops it. Do you know how to make MultiByteToWideChar do either of those things? If not, changing the test failure to a test skipped warning is still a step in the right direction, and I included comments so that the next person who wants to look into this has a good place to start.
Actually never mind, I don't know what's different about the first of the three failing tests that causes the codepoint to be dropped instead of being let through like the other two tests. The first two tests use ITextStream_ReadAll and the last one uses ITextStream_Read.
-Alex