[Xcb] X Test Suite results, again
jamey at minilop.net
Sun Dec 18 04:02:45 PST 2005
Here are the differences between XTS5 results from stock Debian X.org
6.8.2 Xlib vs. Xlib/XCB. Now I understand the odd code that some of you
watched me rip out of Xlib today.
Xlib13: XSetModifierMapping-8 was failing on XCB because the
XSetModifierMapping function assumes that the first 32 bytes of the
response, *even if it's an error*, will be copied into the caller's
reply buffer. MappingFailed and BadValue happen to both be equal to 2.
Cool! So that's why I wrote that odd code in my version of _XReply. I've
put in a better version of that code now: it's an obvious single memcpy
instead of the complex thing I had before.
Xlib12: XSync-2 doesn't fail on XCB even though it mishandles X errors
on stock Xlib. (I've thought very carefully about the order that errors
get handled in, and I could sort of believe that the stock Xlib code
didn't get the same amount of thought, so I feel kind of OK about this.)
On the other hand two assertions are untestable on XCB, so they return
UNTESTED instead of PASS.
Xlib17: XCreatePixmapFromBitmapData-2 fails on XCB with "Got
BadDrawable, Expecting BadValue", and I have no idea why yet.
So, I lost a FAIL one place and gained an unexplained one elsewhere.
We're breaking even: ship it!
Also, I believe my fix to XCloseDisplay's use of tet_fork broke that
test, as both Xlibs FAIL on it now, and stock Xlib used to PASS it.
Phooey. I've just fiddled with it some and haven't found a correct
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: Digital signature
Url : http://lists.freedesktop.org/archives/xcb/attachments/20051218/5d8e0ad9/attachment.pgp
More information about the Xcb