On 1/19/06, Ben Skeggs <[EMAIL PROTECTED]> wrote: > Am Donnerstag, den 19.01.2006, 22:07 +0100 schrieb Jerome Glisse: > > Hi Ben, > > > > I don't know if you remember but we got a discussion > > on swizzle issue when we have to write W to a XYZ few > > month ago. You solved the issue by syncing the XYZ & W > > emit, right (emitting W with the last XYZ emit) ? > I recall the discussion. The issue at the time was a misunderstanding > on my behalf of how the XYZ and W units interact. > > > > > It seems that if you emit something like : > > srcmask(WZX) writemask(X) you have the wrong result but > > srcmask(WZX) writemask(XY or XZ or XYZ) give right result > > without syncing. I am update my swizzling table to do so. > Does this fail with the current cvs? If so, can you switch on the call > to dump_program() at the end of translate_fragment_shader() and post the > output as well as the test program? >
The test program is your, what i mean is that syncing is not necessary if you force writing to Y,Z with WZX and not only X, maybe doing so force r300 to sync internaly. Btw cvs work well, i am just testing stuff. best, Jerome Glisse ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid3432&bid#0486&dat1642 -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
