On Thu, 18 Apr 2019 at 22:29, Francois Gouget fgouget@codeweavers.com wrote:
On Thu, 18 Apr 2019, Henri Verbeet wrote:
This doesn't reduce the number of ok() call all that much though. It will abort earlier on failure, but successful runs are going to produce the same amount of data as before.
But then successful ok() calls don't print anything... at least if you don't set WINETEST_REPORT_SUCCESS which we don't for regular WineTest runs since that would cause the report to be just huge.
I somehow got the impression successful ok() calls still contributed to the log size, but it looks like that's just wrong.
So since these "for y" loops don't have other ok() calls or trace() calls that should be fine. Going up one loop we have one "Failed to create rasterizer state" ok() call for the bias_clamp_tests loop but my impression was this one would stay out of the check_readback_xxx() function anyway.
I was thinking of moving the ok() calls out of the y-loop, yes.