The bit slicing API was introduced in #291 , but it is confusing when it allows reversed slicing, since the ranges in Python are all left-closed right-open, while the reversed slicing breaks this rule. To avoid confusion, I think we should prevent using reversed slicing and force the begin of the slice smaller than the end of the slice. This is also consistent with other hardware-oriented Python DSLs like PyMTL3 (See Page 8 of the documentation).
For reversing the sliced bits, we need to introduce a new API .reverse(). The grammar and behavior is like the one described in the VHLS documentation. Therefore, we have the following example:
>>> a = 0xabcd0123
>>> a[28:32] # containing the bit of 28, 29, 30, 31
0xa
>>> a[4:24]
0xcd012
>>> a[28:32].reverse()
0x5
And we no longer support grammar like a[32:28] which leads to confusion.
The bit slicing API was introduced in #291 , but it is confusing when it allows reversed slicing, since the ranges in Python are all left-closed right-open, while the reversed slicing breaks this rule. To avoid confusion, I think we should prevent using reversed slicing and force the begin of the slice smaller than the end of the slice. This is also consistent with other hardware-oriented Python DSLs like PyMTL3 (See Page 8 of the documentation).
For reversing the sliced bits, we need to introduce a new API
.reverse(). The grammar and behavior is like the one described in the VHLS documentation. Therefore, we have the following example:And we no longer support grammar like
a[32:28]which leads to confusion.