summaryrefslogtreecommitdiff
path: root/cad
diff options
context:
space:
mode:
authordmcmahill <dmcmahill>2000-04-09 23:11:49 +0000
committerdmcmahill <dmcmahill>2000-04-09 23:11:49 +0000
commit309b9a5ccbc19426f4fbb4143b1cb1eb3d19ead0 (patch)
tree2d2b0f3175ee99a9facc1be1899a44d03bc5d68a /cad
parent3dd4fab6801e59f44aada93cdee40f8ad8de2abe (diff)
downloadpkgsrc-309b9a5ccbc19426f4fbb4143b1cb1eb3d19ead0.tar.gz
update to verilog-current-20000409.
changes since the last packaged snapshot include: Icarus Verilog 20000326 Snapshot: -------------------------------- The VVM backend rewrite continues. More templates are gone, and the bit functions have been pretty much rewritten. The vvm library now handles bit values with strengths, and most of the devices to the right things with those strengths. The most obvious implication of this is that you can write multiple drivers to a net and expect the values to be properly resolved, and in particular the HiZ value works as it should. So I am well on the way to completing strength modeling support. What is still missing is support for strength specifications in the Verilog source. Although the parser supports the strength related keywords, they are not passed on to elaboration, or used to generate drivers with the proper strengths. So that's in the works. While doing all this VVM rewrite, I've made the generated code considerably smaller. And of course fewer templates are used. The upshot of this is that compiles of larger designs should go a whole lot faster. This is important because people are using Icarus Verilog for increasingly larger designs. On some larger examples, I've achieved more then 3X compile time improvement. Icarus Verilog 20000409 Snapshot: -------------------------------- Named events now work! Event object declarations and trigger statements are fully supported, and blocking on a single event also works. I'm not up to named events in event lists because I'm in the midst of redesigning the way events on nets and regs are implemented. However, the common case works fine, so there you are. I've also added support for some more arithmetic operators. Division and Modulus now work in many contexts, and are not far from working everywhere. Also, comparison operators work in places they used to not. There was a compile error in memory objects that managed to slip through a couple snapshots, that I finally cured. The problem was pretty gross, but somehow not quite tickled by my tests. Oh well. I've integrated some VCD improvements from Anthony Bybell. Some of you recognize the name as the author of GTKWave, so if he says VCD works like so, then that's how VCD works:-) Anyhow, he fixed the VCD output to be more portable, and also a bit smaller when vectors are involved. He also fixed some bugs with multiple calls to $dumpvars.
Diffstat (limited to 'cad')
-rw-r--r--cad/verilog-current/Makefile6
-rw-r--r--cad/verilog-current/files/md54
2 files changed, 5 insertions, 5 deletions
diff --git a/cad/verilog-current/Makefile b/cad/verilog-current/Makefile
index c5131f53f03..91a0b68399b 100644
--- a/cad/verilog-current/Makefile
+++ b/cad/verilog-current/Makefile
@@ -1,8 +1,8 @@
-# $NetBSD: Makefile,v 1.2 2000/03/25 21:09:16 dmcmahill Exp $
+# $NetBSD: Makefile,v 1.3 2000/04/09 23:11:49 dmcmahill Exp $
#
-DISTNAME= verilog-20000318
-PKGNAME= verilog-current-20000318
+DISTNAME= verilog-20000409
+PKGNAME= verilog-current-20000409
CATEGORIES= cad
MASTER_SITES= ftp://icarus.com/pub/eda/verilog/snapshots/
diff --git a/cad/verilog-current/files/md5 b/cad/verilog-current/files/md5
index 1a92d697005..f1c1e531537 100644
--- a/cad/verilog-current/files/md5
+++ b/cad/verilog-current/files/md5
@@ -1,3 +1,3 @@
-$NetBSD: md5,v 1.2 2000/03/25 21:09:16 dmcmahill Exp $
+$NetBSD: md5,v 1.3 2000/04/09 23:11:50 dmcmahill Exp $
-MD5 (verilog-20000318.tar.gz) = 8d03f502d46986cc8724341e640dfa3b
+MD5 (verilog-20000409.tar.gz) = 7b2fc0db2afb7953e133dc6f928a538c