Fixed compatibility of sep, end keyword with list/tuple containers

[anutosh491]

Fixes #1477

Some issues that were fixed are

1. sep and end keyword were assumed to be functioning correctly beforehand with data types like i32, str etc but trying out some cases showed that they aren't fully compatible . Pointed out in Variable of list[i32] type has unnecessary newline character while printing. #1477 (comment) . This has been addressed.

2. Secondly all sorts of variations of print statements involving the list/tuple containers have been fixed .

def main0():
    a: i32 = 1
    b: i32 = 2
    c: str = "abc"
    d: str = "def"
    e: list[i32] = [1, 2]
    f: list[i32] = [3, 4]
    g: tuple[i32, i32, i32] = (1, 2, 3)
    h: tuple[i32, i32] = (4, 5)
    l : list[i32] = [1,2,3]
    l2 : list[i32] = [4, 5, 6]
    s : str = "abc"
    print(a, g, c,sep = "xyz", end = "pqr")
    print(g, a, c, sep = "xyz", end = "pqr")
    print(a, e, g, sep = "xyz", end = "pqr")
    print(a, l, s)
    print(l , a, s)
    print(a, s, l)
    print(a, l, s, end = " pqr ")
    print(l, a, s, end = " pqr ")
    print(a, s, l, end = " pqr ")
    print(a, l, s, sep = "xyz")
    print(l, a, s, sep = "xyz")
    print(a, s, l, sep = "xyz")
    print(a, l, s, sep = "xyz", end = "pqr")
    print(l, a, s, sep = "xyz", end = "pqr")
    print(a, s, l, sep = "xyz", end = "pqr")
    print(e, f, a, sep = "xyz")
    print(g, e, f, end = "pqr")

main0()

(lf) anutosh491@spbhat68:~/lpython/lpython$ ./src/bin/lpython examples/expr2.py 
1xyz(1, 2, 3)xyzabcpqr(1, 2, 3)xyz1xyzabcpqr1xyz[1, 2]xyz(1, 2, 3)pqr1 [1, 2, 3] abc
[1, 2, 3] 1 abc
1 abc [1, 2, 3]
1 [1, 2, 3] abc pqr [1, 2, 3] 1 abc pqr 1 abc [1, 2, 3] pqr 1xyz[1, 2, 3]xyzabc
[1, 2, 3]xyz1xyzabc
1xyzabcxyz[1, 2, 3]
1xyz[1, 2, 3]xyzabcpqr[1, 2, 3]xyz1xyzabcpqr1xyzabcxyz[1, 2, 3]pqr[1, 2]xyz[3, 4]xyz1
(1, 2, 3) [1, 2] [3, 4]pqr(lf) anutosh491@spbhat68:~/lpython/lpython$ python examples/expr2.py 
1xyz(1, 2, 3)xyzabcpqr(1, 2, 3)xyz1xyzabcpqr1xyz[1, 2]xyz(1, 2, 3)pqr1 [1, 2, 3] abc
[1, 2, 3] 1 abc
1 abc [1, 2, 3]
1 [1, 2, 3] abc pqr [1, 2, 3] 1 abc pqr 1 abc [1, 2, 3] pqr 1xyz[1, 2, 3]xyzabc
[1, 2, 3]xyz1xyzabc
1xyzabcxyz[1, 2, 3]
1xyz[1, 2, 3]xyzabcpqr[1, 2, 3]xyz1xyzabcpqr1xyzabcxyz[1, 2, 3]pqr[1, 2]xyz[3, 4]xyz1
(1, 2, 3) [1, 2] [3, 4]pqr(lf) anutosh491@spbhat68:~/lpython/lpython$

These are like some random cases I thought and none of these work on master !

[anutosh491]

I'll pick some tests from the above shown function and add an integration file soon @Thirumalai-Shaktivel !

[Thirumalai-Shaktivel]

I think we have to add runtime tests.
@certik @czgdp1807 What do you say?

[Thirumalai-Shaktivel]

Great job, @anutosh491. Thanks for fixing this.

[czgdp1807]

I think we have to add runtime tests.

Could you please elaborate more? If you are talking about adding a file in integration_test, then I guess @anutosh491 has said he will do it soon (#1512 (comment)). Is there anything more to this? If not then I am marking it as draft.

[Thirumalai-Shaktivel]

I was talking about saving the runtime output using tests.toml: run

[czgdp1807]

I see. That's very much needed. For printing we can't use assert statements. However I guess, a better way is that print function should call str over its inputs and then just send the output of str for printing. That way we will be able to do assert with str calls and no need to store reference files for outputs.

[anutosh491]

I think we have to add runtime tests.
I was talking about saving the runtime output using tests.toml: run
I see. That's very much needed. For printing we can't use assert statements. However I guess, a better way is that print function should call str over its inputs and then just send the output of str for printing. That way we will be able to do assert with str calls and no need to store reference files for outputs.

I think I am not fully sure how do I address these comments . I get that we could use the assert keyword if we make use of str() function , but where and what sort of tests am I suppossed to add is not very clear. Any references I could look into ?

[anutosh491]

I think I am not fully sure how do I address these comments . I get that we could use the assert keyword if we make use of str() function , but where and what sort of tests am I suppossed to add is not very clear. Any references I could look into ?

cc @Thirumalai-Shaktivel

[Thirumalai-Shaktivel]

ping @czgdp1807

[czgdp1807]

I think I am not fully sure how do I address these comments . I get that we could use the assert keyword if we make use of str() function , but where and what sort of tests am I suppossed to add is not very clear. Any references I could look into ?

Nothing to worry about here. Just add normal print statements with constant strings and register only in tests.toml with asr=true, ast=true. That's all. No need to worry about printing of runtime strings.

[Thirumalai-Shaktivel]

LGTM! But, Let @czgdp1807 also review and approve this.
Also, It seems you made changes to print_arr.cpp, do you wanna add tests (in test_sep_end_kws.py) for it?

[anutosh491]

Also, It seems you made changes to print_arr.cpp

Yess I'll be updating my integration test file with tests for printing arrays . Just want to get this output confirmed (#1512 (comment)) before I go on with that ! The changes makes printing arrays compatible with other data types/containers including arrays along with the end and sep keywords !

[certik]

Does CPython and LPython print exactly the same thing, or does it differ?

[anutosh491]

1. For containers like lists/tuples and other fundamental data types, both outputs are the same while using end and sep keyword . For example (Fixed compatibility of sep, end keyword with list/tuple containers #1512 (comment)) as shown in PR description

2. For arrays though , they do differ

For something like

from ltypes import i32
from numpy import empty, int32

def main0():
    x: i32 = 2
    arr1: i32[x] = empty(x, dtype=int32)
    arr2: i32[x] = empty(x, dtype=int32)
    arr1[0] = 100
    arr1[1] = 200
    arr2[0] = 300
    arr2[1] = 400
    print(arr1, arr2)
    print(arr1, arr2, sep="abc")
    print(arr1, arr2, end="pqr\n")
    print(arr1, arr2, sep="abc", end="pqr\n")

main0()

The CPython output is

[100 200] [300 400]
[100 200]abc[300 400]
[100 200] [300 400]pqr
[100 200]abc[300 400]pqr

But the lpython output is

100 200 
300 400 
100 200abc300 400 
100 200 
300 400pqr
100 200abc300 400pqr

Differences are

1. We haven't incorporated square brackets
2. Our default separator between 2 arrays is a newline rather than enclosing the array elements with brackets and using space in between !

You coudl go through the thread (#1512 (comment)) where we have been discussing this . If you feel something more needs to be done for arrays , I can maybe open up a subsequent pr for it because I am quite confident about the initial changes which were mentioned in the issue description (#1512 (comment))

[anutosh491]

cc @certik

[certik]

I left some more comments. Otherwise I think this can be merged, and subsequent PRs can improve the rest.

[anutosh491]

I've removed the ast based reference files which wouldn't be required anymore . I think now this maybe good to go !

[certik]

I think that this is fine for now. Squashing and merging.

[anutosh491]

Thanks @Thirumalai-Shaktivel @czgdp1807 and @certik for the great reviews on this one !