Revival string list

[aman-godara]

This PR is based upon #311 by Arjen. Since 311 is very long, I decided to break it into parts and making it easier for others to review and for me to concentrate on one thing at a time.
How to Start Reviewing by gareth.

Previous discussions (in order of time):
Linked Issue: #268
string list: #269
string list new: #311

Time Line:

• 1. Arjen's kickstart
• 2. Corrected a bug in append operator(//)
• 3. Refactored code to use stdlib's string_type, Provided support for allocatable character (character(len=:), allocatable)
• 4. Used concept of forward backward indexes
• 5. added copy function
• 6. Blocked integer indexes to stringlist use only stringlist_index_type indexes
• 7. to_current_idxn (get) vs to_future_at_idxn (insert_at)
• 8. added == and /= operator
• 9. Improve test cases for insert_at, append and prepend
• 10. Add Documentation

Inviting: @zoziha, @St-Maxwell, others are welcomed as well.

[aman-godara]

Explanation of forward and backward indexes:
forward index: fidx
backward index: bidx

After inserting an element at fidx( x ), the added element will appear at index fidx( x ) i.e. at xth position when counted from left to right.
After inserting an element at bidx( y ), the added element will appear at index bidx( y ) i.e. yth position when counted from right to left.
inserting at fidx( 1 ) is equivalent to inserting at bidx( len + 1 )
inserting at bidx( 1 ) is equivalent to inserting at fidx( len + 1 )

Inserting an element at HEAD is like inserting at fidx( 1 ) and inserting an element at TAIL is like inserting at bidx( 1 )

backward index is like the forward index of the reversed form of the list.

---

index type	...	---	A	B	..	Y	Z	---	...
integer	...	---	1	2	..	len - 1	len	---	...
forward	...	fidx( 0 )	fidx( 1 )	fidx( 2 )	..	fidx( len - 1 )	fidx( len )	fidx( len + 1 )	...
backward	...	bidx( len + 1 )	bidx( len )	bidx( len - 1 )	..	bidx( 2 )	bidx( 1 )	bidx( 0 )	...

---

If you want to insert an element new_element before integer index q use the forward index of integer index q. If you want to insert an element new_element after integer index q use the backward index of integer index q.

Currently there is NO function in the PR which converts integer index to forward index or backward index [not to be confused with fidx and bidx of PR]

[aman-godara]

Apart from reviewing code, you can also leave your opinion and feedback on the design. Coming up with a good design is the major challenge in this PR.

The design that I have proposed has problems (I agree) one of them is: no one would prefer to use these forward and backward indexes over integer indexes. Having these stringlist_index_type indexes just makes code more verbose by calling fidx and bidx function again and again (I wish I had a better solution to propose).

I could have used positive integer indexes for forward indexes and negative integer indexes for backward indexes but the problem was with the zero. Take help of this comment to understand it better. While reviewing #311 I got the idea of using stringlist_index_type indexes: forward and backward. Now, I need community's help in figuring out other problems. So feel free to leave your opinion.

[zoziha]

It seems that CMakeLists.txt and Makefile.manual have not been set up to compile your code, so maybe I should focus on understanding the issues you want to discuss, not grammatical typos.

[aman-godara]

I have 2 questions:

1). Should the module be renamed to stringlist_type to be consistent with string_type?
2). Should to_current_idxn be made public?

[aman-godara]

1). Should the module be renamed to stringlist_type to be consistent with string_type?
2). Should to_current_idxn be made public?

Let me rename module from stringlist to stringlist_type and keep to_current_idxn as private.

[aman-godara]

All changes have been made.

[everythingfunctional]

I could probably nitpick some more, but honestly this is a pretty good first implementation. I also don't want to jeopardize you completing GSoC, because IMO, this is more than sufficient to pass. Thanks for the hard work you put in.

[jvdp1]

nice job. Thank you!
I approved it, pending some minor comments.

[jvdp1]

Why is there not a constructor with array of type string_type?

[aman-godara]

After size component was discarded it was causing ambiguous interfaces to provide support for string arrays. Which prompted me to remove it but I have added it again. Please have look at the commit 4b69e3d

[jvdp1]

Well, I don't completely agree with you. But if you want to not mention a result, then I would suggest to remove the section `Result value`, as done in https://stdlib.fortran-lang.org/page/specs/stdlib_sorting.html#specifications-of-the-stdlib_sorting-procedures . So there will be no confusion. Le ven. 20 août 2021 à 15:53, Aman Godara ***@***.***> a écrit :

…

***@***.**** commented on this pull request. ------------------------------ In doc/specs/stdlib_stringlist_type.md <#470 (comment)>: > + +#### Class + +Pure subroutine. + +#### Argument + +- `idx`: [[stdlib_stringlist_type(module):stringlist_index_type(type)]]. + This argument is intent(in). + +- `string`: Character scalar or [[stdlib_string_type(module):string_type(type)]]. + This argument is intent(in). + +#### Result value + +No result. If we mention this can it likely cause a reader to misunderstand this subroutine as a function instead which is returning a new stringlist, and a user might end using a syntax like this: list = list%insert_at( fidx(10), "10th element") — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#470 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AD5RO7EKSSMDRRMWZBZ3GL3T5ZM57ANCNFSM5AYNZL5A> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email> .

[jvdp1]

Thank you @aman-godara for this work. Really useful.

[aman-godara]

Looks like to_string has been deleted or something from stdlib_ascii. It is one of the dependencies for stringlist.

[awvwgk]

to_string has been moved stdlib_strings with #444.

[milancurcic]

Thanks @aman-godara and reviewers! Unless there are objections let's merge this tomorrow.