<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;">Hi all<div><br></div><div>Moving this from LLVMDev (<a href="http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-July/075223.html">http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-July/075223.html</a>) now that i have patches i’d welcome reviews of. The following is the initial summary from that email thread.</div><div><br></div><blockquote style="margin: 0 0 0 40px; border: none; padding: 0px;"><div>"Currently tablegen parses binary literals such as 0b011 and immediately turns them in to integers internally. Due to this, 0b011 is a 2-bit value because tablegen will happily drop leading zeroes on integers.</div><div><br></div><div>I propose that we change this, and store binary literals with a size. I think this is much more natural, as when the user writes a value with 3 bits, I think they expect to get a value with 3 bits, and not a 2-bit value which is silently truncated and later silently zero extended.</div><div><br></div><div>Given this, I would then like to extend bits initializers to accept these new sized binary literals. For example, I would like to be able to write</div><div><br></div><div>bits<2> opc;</div><div>bits<8> x = { opc, 0b1100, rd{1}, rs{0} }</div><div><br></div><div>This would let us write some encodings much more concisely, instead of having to put each initializer on its own line.”</div></blockquote><div><br></div><div>For those interested in the low level details, i’ve changed binary literals such as 0b011 from creating an integer ‘3’, to creating a BitsInit ‘{ 0, 1, 1}’. As Sean suggested, this was better than introducing yet another type just for binary literals.</div><div><br></div><div>As part of this, BitsInit values now had to be able to be typed. This is because we had uses of binary literals which expected to be typed (for example list<bits<2>> [0b00, 0b01]). I’ve moved BitsInit from being a subclass of Init to a subclass of TypedInit to allow this.</div><div><br></div><div>Note that patches 2 and 3 really need to be committed together. I can squash them before commit, but just wanted to present the changes independently.</div><div><br></div><div>Patch 5 can also be committed without any of the others, but just shows that this did actually catch some useful cases where we mismatched sizes of binary expressions.</div><div><br></div><div>Thanks,</div><div>Pete</div><div><br></div><div></div></body></html>