On Wed, 13 May 2015, Brad Beckmann wrote:
On May 12, 2015, 8:47 p.m., Nilay Vaish wrote:
There are two problems I have with this patch. Firstly, it changes the
generally
accepted meaning of '*'. Secondly, it is adding another keyword to the language
which is, in my opinion, not required. Instead, the transition construct should
be upgraded so that it recognizes that the final state argument can be anything
whose type
is state. This would allow us to name the function whatever we want it to be.
What do you mean it changes the meaning of '*'? That symbol is not
currently used in SLICC. SLICC has been designed not to directly expose
pointers to the programmer, so the '*' symbol is currently unused.
I meant that '*' is used to match any number of characters in regular
expressions. In this case, it would mean any state.
Upgrading the transistion construct requires a pretty significant change
to the parser. The current patch is much cleaner and the notation is
simplier.
I am actually OK with using '*'. But I do not like that we associating a
fixed function getNextState() with '*'. I am of the opinion that adding
more keywords / fixed functions to SLICC is not right. I suggest that we
change the parser to allow for any thing that has type state be mentioned
in the transition construct.
--
Nilay
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
