[TVMScript] Improve printer for TIR syntax sugar

[shingjan]

This PR intends to improve the printing of three kinds of syntax sugar introduced in TVMScript from #9492, #9620 and #9634. Examples below are supposedly the output of print(some_tir_prim_func.script()).

1. For reads & writes:

Before this PR

@T.prim_func
def func(a: T.handle, b: T.handle, c: T.handle) -> None:
            ...
            T.reads([C[vi, vj], A[vi, vk], B[vj, vk]])
            T.writes([C[vi, vj], A[vi, vk]])
            ...

After this PR, note that brackets are removed

@T.prim_func
def func(a: T.handle, b: T.handle, c: T.handle) -> None:
            ...
            T.reads(C[vi, vj], A[vi, vk], B[vj, vk])
            T.writes(C[vi, vj], A[vi, vk])
            ...

2. For loops:

Before this PR

@T.prim_func
def func(a: T.handle) -> None:
    ...
    for i in T.serial(0, 128):
        for j in T.parallel(0, 128):
            for k in T.vectorized(0, 128):
                for x in T.unroll(0, 128):
                    for y in T.thread_binding(0, 128, thread="threadIdx.x"):
                        ...

After this PR, note that all loops with 0 as min will be printed out as syntax sugar

    ...
    for i in T.serial(128):
        for j in T.parallel(128):
            for k in T.vectorized(128):
                for x in T.unroll(128):
                    for y in T.thread_binding(128, thread="threadIdx.x"):
                        ...

3. For T.match_buffer():

Before this PR

@T.prim_func
def elementwise(
    a: T.handle,
    b: T.handle,
) -> None:
    A = T.match_buffer(a, (128, 128, 128, 128))
    B = T.match_buffer(b, (128, 128, 128, 128))
    # some computation

After this PR, note that we assume T.match_buffer with two or less arguments specified are syntax-sugarred.

@T.prim_func
def elementwise(
    a: T.Buffer[[128, 128, 128, 128], "float32"],
    b:T.Buffer[[128, 128, 128, 128], "float32"],
) -> None:
    # some computation

cc: @vinx13 @junrushao1994 @Hzfengsy

[vinx13]

Please fix CI error and double check if works locally

[vinx13]

from the CI error looks like there are some issues with reads/writes in previous PR

[vinx13]

a few minor issues