I'm not sure exactly what you mean by trigger. There could be something like #use_file_foo_now that triggers changing the output file, or there could be file "handles" so to speak like FileGroupFoo and when you want to output something you have to pick one. Different approaches have different trade offs. Something that's a static landmark in the file ala #include would make a scanner easier, but being able to pick a file group to use ala file descriptors would be more flexible and possibly make it harder to write cryptic code where it's not clear where things are ending up.
Gabe On 02/02/11 16:56, Steve Reinhardt wrote: > > Yea, you'd think so... I guess my point is really that we should > design the best mechanism wrt the language first, then worry about > scons second, and even if the best mechanism requires us to explicitly > list dependencies in scons that's probably not sufficient reason to > reject it. And I think Nate's point is that it's not that likely either. > > Steve > > On Feb 2, 2011 4:50 PM, "nathan binkert" <[email protected] > <mailto:[email protected]>> wrote: > >> If we can't do a scanner, I don't have a huge problem with listing > output > >> files explicitly... yea, it's not as elegant, but I don't expect it to > >> change a lot either. > > > > My main question is, how does the parser determine what files it will > > generate? Isn't there some trigger in the language at least? > > > > Nate > > _______________________________________________ > > m5-dev mailing list > > [email protected] <mailto:[email protected]> > > http://m5sim.org/mailman/listinfo/m5-dev > > > _______________________________________________ > m5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/m5-dev
_______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
