search

Minres / CoreDSL

Xtext project to parse CoreDSL files
Apache License 2.0
16 stars 3 forks source link

Inconsistencies in Programmer's Manual #31

Closed wysiwyng closed 2 years ago

wysiwyng commented 2 years ago

In https://github.com/Minres/CoreDSL/wiki/Structure-and-concepts, a sign is applied to an encoding entry: encoding: imm[31:12]s :: ...;, which I would love to see implemented again, but is currently incorrect.

Declarations of pointer types seems to be no longer allowed, but dereferencing them still is: Dereferencing: https://github.com/Minres/CoreDSL/blob/0be747b6012d63ce97264e6e43b0b2181d5aefd5/com.minres.coredsl/src/com/minres/coredsl/CoreDsl.xtext#L269-L276 Declaration: https://github.com/Minres/CoreDSL/blob/0be747b6012d63ce97264e6e43b0b2181d5aefd5/com.minres.coredsl/src/com/minres/coredsl/CoreDsl.xtext#L185-L189 Documentation: https://github.com/Minres/CoreDSL/wiki/Types#derived-types,

jopperm commented 2 years ago

Thanks for your feedback!

In https://github.com/Minres/CoreDSL/wiki/Structure-and-concepts, a sign is applied to an encoding entry: encoding: imm[31:12]s :: ...;, which I would love to see implemented again, but is currently incorrect.

Removed that in the spec. While undeniably convenient, we discussed this earlier and decided against having special syntax only for the encoding fields.

Declarations of pointer types seems to be no longer allowed, but dereferencing them still is: Dereferencing:

Yeah, the pointer types are currently in bit in limbo. AFAIK, there is no use-case that needs support for them in CoreDSL, but I also don't have compelling arguments to forbid them. To reach a consistent state, I'd suggest that @AtomCrafty drops the & and * operators (this wouldn't affect the alias-syntax, correct?) as well from the grammar, and I'll update the spec accordingly. @eyck, WDYT?

eyck commented 2 years ago

No objections. As you remember we discussed that briefly and if any use case utilizing pointers come up we can extend the syntax accordingly.

jopperm commented 2 years ago

Ok, I modified the spec. @AtomCrafty, please remove the remaining pointer support from the grammar.

AtomCrafty commented 2 years ago

The & and * prefix operators have been removed in the recent rewrite of the expression grammar.

jopperm commented 2 years ago

Thanks, then I believe we can close this issue.