On 19 September 2011 10:59, Ian Romanick <span dir="ltr"><<a href="mailto:idr@freedesktop.org">idr@freedesktop.org</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"> -----BEGIN PGP SIGNED MESSAGE-----<br> Hash: SHA1<br> <div class="im"><br> On 09/15/2011 04:40 PM, Paul Berry wrote:<br> > This patch fixes a bug in ir_hirearchical_visitor: when traversing<br> > an exec_list representing the formal or actual parameters of a<br> > function, it modified base_ir to point to each parameter in turn,<br> > rather than leaving it as a pointer to the enclosing statement.<br> > This was a problem, since base_ir is used by visitor classes to<br> > locate the statement containing the node being visited (usually so<br> > that additional statements can be inserted before or after it).<br> > Without this fix, visitors might attempt to insert statements into<br> > parameter lists.<br> <br> </div>This sure sounds like a bug, but I'm really surprised that this hasn't<br> cause catastrophic failures before. Do you have a shader in mind that<br> fails due to this issue?<br></blockquote><div><br>No, but if this bug isn't fixed, then the patch that follows in this patch series ("glsl: Implement a lowering pass for gl_ClipDistance.") fails pretty catastrophically.<br> </div></div>