From: Adam Krolnik (krolnik@lsil.com)
Date: Tue Jun 18 2002 - 16:16:08 PDT
Precedence: bulk
Hi Steve;
>The comments seem to be based on the idea of using the library declarations
>to find the source code for modules, so they can be compiled in, like the
>Verilog-XL -y and -v options. Verilog-XL has to find and then parse the
>source code because it has no concept of a compiled library.
Yes, that was the intepretation.
>The config stuff seems to be based around the concept of compiled libraries.
Partly, but not necessarily.
>The library paths tell the parser which library to put the compiled form of
>a given source file in, based on where the source file came from. But it
>still assumes that the user specifies which source files to compile. And
If tools know where modules are (based on the library mapping files),
then
a user doesn't need to specify the source files, but instead
modules/configurations
to use. A tool could build a library but not compile all the sources,
then read
the configuration to determine what needs to be compiled and then
produce a
simulation to run.
>then the configs determine which library to get the definition of a module
>from. It is not clear to me that the parser would ever use the file paths
>to search for additional source files that were not explicitly specified
>by the user (though section 13.4.2 might be describing such a thing). If
>tools don't do this, then the issues being raised don't occur.
It could go either way - I hope that tools start with the library
mapping file
and don't require another 'list of files' to compile a simulation - its
somewhat
redundant.
Adam Krolnik
Verification Mgr.
LSI Logic Corp.
Plano TX. 75074
This archive was generated by hypermail 2.1.4
: Mon Jul 08 2002 - 12:55:37 PDT
and
sponsored by Boyd Technology, Inc.