[slang] Fix PLA task concatenation ascending order

[likeamahoney]

Added support for elaborating correctness of examples like the one from LRM section 20.16:

module async_array(a1,a2,a3,a4,a5,a6,a7,b1,b2,b3);
  input a1, a2, a3, a4, a5, a6, a7 ;
  output b1, b2, b3;
  logic [1:7] mem[1:3]; // memory declaration for array personality
  logic b1, b2, b3;
  initial begin
      // set up the personality from the file array.dat
     $readmemb("array.dat", mem);
     // set up an asynchronous logic array with the input
     // and output terms expressed as concatenations
     $async$and$array(mem,{a1,a2,a3,a4,a5,a6,a7},{b1,b2,b3});
  end
endmodule

By default, slang sets all concatenation expression bitwidth ranges in descending order. I added an ASTFlag to change this behavior when concatenation is an argument of the PLA task.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.71%. Comparing base (4221f2f) to head (3b40ce6).
Report is 4 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1047   +/-   ##
=======================================
  Coverage   94.71%   94.71%           
=======================================
  Files         191      191           
  Lines       47656    47667   +11     
=======================================
+ Hits        45136    45147   +11     
  Misses       2520     2520           

Files	Coverage Δ
source/ast/builtins/SystemTasks.cpp	91.13% <100.00%> (+0.03%)	⬆️

... and 2 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4221f2f...3b40ce6. Read the comment docs.

[MikePopoloski]

This seems like an overly intrusive way of fixing this -- I'd prefer to avoid adding a specialized ASTFlag just for this one case. It seems like it should be fixable in the PLATask handler itself by changing checkArguments to see that it's a concat expression and just not issuing an error in that case.

[MikePopoloski]

Adding SLANG_EXPORT here doesn't do anything; this class is not externally visible to downstream users of the library.

[likeamahoney]

removed

[likeamahoney]

This seems like an overly intrusive way of fixing this -- I'd prefer to avoid adding a specialized ASTFlag just for this one case. It seems like it should be fixable in the PLATask handler itself by changing checkArguments to see that it's a concat expression and just not issuing an error in that case.

fixed