* James Hawkins wrote:
- On Jan 21, 2008 2:14 PM, Christopher wrote:
- Dmitry Timoshkov wrote:
Also, you need to test LoadStringA, to see if it behaves similarly. It would be also interesting to test LoadStringA/W with both buffer and buffer length set to 0.
I tested LoadStringA under Windows XP, and calling it with buflen == 0 does not return a pointer to the resource. In fact LoadStringA seems to behave fairly differently from LoadStringW: in that calling with buffer == NULL causes an access violation instead of just returning 0.
That's why you need to add tests for LoadStringA to Wine's test suite.
No, he did not. The word "need" wasn't appropriate here. Christopher wasn't going to change LoadStringA. If changes of LoadStringW breaks LoadStringA then tests would show regression. If tests are too weak, then only patch commiter or previous patchers of the LoadStringA are to blame, not the casual contributor who enhances B.
And if this is some new official rule for code, then I don't find it on the site:www.winehq.org . Of course, I don't protest at the idea (as I would do the same as Dmitry wrote), but I protest against usage of this particular word -- it's too strong here. Please, be more accurate with it.
P.S.: James, you probably should start learning to cut unnecessary blocks of quoted text (esp. empty lines) and doing some block justifying in your replies using your MUA some day. Otherwise it gets too hard to read a discussion.