Tried to add refinements for redefining identifiers

[LyubomirT]

Summary of Changes

In this code modification, I have implemented the following features and improvements:

• Added support for chaining multiple [ ] (definitions) to refine a previous definition in the filecabinet and notices sections of the configuration.
• Implemented special empty [ ] to represent resetting or emptying the previous definition.
• Enhanced the code to handle additive refinements and subtractive refinements for more flexible filtering.
• Removed unnecessary whitespaces to improve readability.

Resolved Issues

This pull request addresses the following issues: #3

Testing

All changes made in this pull request have successfully passed local testing. The existing tests in the codebase have been executed to ensure that the new features and improvements do not introduce any regressions.

[ma-ilsi]

Hey @LyubomirT, thank you for your contribution to Sophie! I just did a quick look through your changes. Here are general comments on the pull request:

Ready to merge:

• Implemented special empty [ ] to represent resetting or emptying the previous definition.

This was well done, thanks! Please see the echo vs printf comment, though.

• Removed unnecessary whitespaces to improve readability.

This was an greatly needed enhancement! I deal with a very niche editing environment so didn't realize how my text files end up looking like 😅 . I think I may need to start opening up to industry-wide accepted editors.

Needs more review:

• Added support for chaining multiple [ ] (definitions) to refine a previous definition in the filecabinet and notices sections of the configuration.

• Enhanced the code to handle additive refinements and subtractive refinements for more flexible filtering.

Can you please leave a comment in the GitHub code review on where Sophie can now parse a chain on a single line? It would make it much easier for me to find since the whitespace fixes have thrown off the diff.
Example:

net_code[pattern="...";][size="+1k";]

This above, is additive refinement syntax. It increases the scope of the definition to get all files matching the pattern and all files +1kb in size. Technically speaking, the definition ends up being the set of files that fall under the union of the two filters. I think I should update #3 to be clear on the syntax.

[ma-ilsi]

Finally, if you only want to merge the whitespace organization and the special empty syntax for #3, then I can take that alone if the rest of #3 is not complete.

Regardless of how much you are contributing to refinements, whenever you are done, update your branch to main so I can run these workflows to see macOS conformance.

[LyubomirT]

Hey @LyubomirT, thank you for your contribution to Sophie! I just did a quick look through your changes. Here are general comments on the pull request:

Ready to merge:

• Implemented special empty [ ] to represent resetting or emptying the previous definition.

This was well done, thanks! Please see the echo vs printf comment, though.

• Removed unnecessary whitespaces to improve readability.

This was an greatly needed enhancement! I deal with a very niche editing environment so didn't realize how my text files end up looking like 😅 . I think I may need to start opening up to industry-wide accepted editors.

Needs more review:

• Added support for chaining multiple [ ] (definitions) to refine a previous definition in the filecabinet and notices sections of the configuration.

• Enhanced the code to handle additive refinements and subtractive refinements for more flexible filtering.

Can you please leave a comment in the GitHub code review on where Sophie can now parse a chain on a single line? It would make it much easier for me to find since the whitespace fixes have thrown off the diff. Example:

net_code[pattern="...";][size="+1k";]

This above, is additive refinement syntax. It increases the scope of the definition to get all files matching the pattern and all files +1kb in size. Technically speaking, the definition ends up being the set of files that fall under the union of the two filters. I think I should update #3 to be clear on the syntax.

Got it, I'll start working on this update right away, thanks for the information.

[LyubomirT]

I added comments ( # Handle chaining of multiple filters on a single line ) so it will be easy to find where that is done

[LyubomirT]

@ma-ilsi Updated the main branch, please let me know if you have any questions or concerns.

[ma-ilsi]

I have run the workflow for basic testing. Please have a look at the results in the checks.
Although macOS is labeled passing, unfortunately it is also failing (which reminds me I need to update the exit codes for Sophie).
I have local testing for macOS-10 which is also failing (I should probably hook that up as a runner to workflows.)

I know the output may seem very confusing because of the differences between each shell. But this is the process of making sure a shell script passes on all major operating systems and is POSIX compliant.

Suggestions:

• Make sure you are using some type of debugging method (I believe bash has a script debugger built-in).
• Try to accomplish the changes using the most widely accepted shell syntax.
• Consider testing using a Ubuntu virtual machine.

[LyubomirT]

I have run the workflow for basic testing. Please have a look at the results in the checks. Although macOS is labeled passing, unfortunately it is also failing (which reminds me I need to update the exit codes for Sophie). I have local testing for macOS-10 which is also failing (I should probably hook that up as a runner to workflows.)

I know the output may seem very confusing because of the differences between each shell. But this is the process of making sure a shell script passes on all major operating systems and is UNIX compliant.

Suggestions:

• Make sure you are using some type of debugging method (I believe bash has a script debugger built-in).
• Try to accomplish the changes using the most widely accepted shell syntax.
• Consider testing using a Ubuntu virtual machine.

I'm truly sorry for this, I've made some dumb mistakes while updating the code. Fixing them...

EDIT: Almost done with debugging, will be pushing the update soon.

[LyubomirT]

@ma-ilsi I fixed the glitches, should be working properly now.

[ma-ilsi]

Here we are supposed to match the identifier and the very first opening bracket[. But this regex is going beyond that.

[LyubomirT]

😓 My bad, I'm sorry.

Changed the line to:

cut_count=$(expr "$line" : '[^[]*\[')

I hope this is what you're looking for.

[ma-ilsi]

No worries, no need to apologize! I will see how it does locally.

[ma-ilsi]

@LyubomirT Thank you for your patience! I would like to get your whitespace changes merged but the refinements code is still not mature. Have a look at issue #8 with a fresh branch off of ma-ilsi/Sophie:main and let me know if you'd like to send in the whitespace fixes only.

The refinement changes might take a while to process. I would recommend getting the whitespace mentioned in #8 done separately and then opening a brand new pull for the refinements step-by-step. Try implementing just the empty special [ ] and then you can try one step higher after that. (The benefit of this approach is that I can read the diffs easier and bugs will be caught faster if each part is done step by step.)

[ma-ilsi]

Do you think we should add awk over sed? I think having a few rare instances of awk throws off simplicity and consistency. I was trying to stick to the most basic utils that almost every developer would know: sed, grep, cut, tr are simple to learn and enough for all the parsing needs.