Some system source reformatting/refactoring

[rgchris]

If nothing but for my own understanding and sanity (but really as a recommendation), I've reformatted/refactored sys-load.reb and sys-base.reb.

Methodology

• Break up lines containing multiple expressions (where possible)
• Consistent syntax spacing (editor/folding friendly)
• Expand TLWs (three-letter words)
• Intentional naming (words reflect intent, not vague value types)
• Reformatting and refactoring are contained in separate commits to demonstrate changes
• Remain drop-in compatible with current source

Purpose

From Rebol 3 Alpha, system source code is needlessly dense and cryptic. These edits are intended to better elucidate some of Rebol's critical internal operations in order to inform changes going forward (in particular, the module system)

[Oldes]

That is quite a lot changes :-)

[Oldes]

I may merge it, if it will be mergeable - #153

[Oldes]

Unfortunately there is an issue when importing extensions :/

** Script error: -apply- does not allow #(none!) for its name argument
** Where: apply case import do either either if -apply-
** Near: apply :load-module [
    options/modules/:file
    version 

[rgchris]

It should have been mergeable!—I did try not to change any logic. I'll take a look—do you have a sample that works before that doesn't now?

[rgchris]

I found it—refactoring error, it should work now

[Oldes]

Code like this one fails with your modification (which is the most common way to import something for me).

[Oldes]

Ok.. trying the fix.

[Oldes]

Btw... is really more readable when there is so many empty lines? To be honest, I prefer to see as much code on the page as possible. Now I must scroll up and down so much to read it.

[rgchris]

Two issues that were problems for me:

• The condensed style makes editor code folding less effective. Particularly on CASE or SWITCH code where folding can help 'zoom out'* and view higher level code flows. It is a different way of seeing as much code on the page as possible.

• Run-on code segments where different operations kind of blend together. Empty lines aren't just aesthetic, they add logic to groupings of code and help define how to organize code, identify problems, and ultimately aid in optimization (the more forcibly I spread code out in this way, the more I can find ways to shrink it back down again based on reasoning)

I guess ultimately coding style comes down to personal preference. My coding style used to be similar to the condensed style, but I'd have problems going back to my own code and trying to decipher it. I developed this style in part to address this, and to take advantage of folding editors. For me has led to a better approach to coding in Rebol (and in general)—it's kind of infectious when you get used to it.

Some other benefits,

• Breaking expressions into separate lines helps to identify problems by line number,

• More room for whole words (vs. having to use compressed words—like HDR or MOD—to make them fit)

• I'd say there are aesthetic benefits too, it just looks better. Yes, condensed style all fits, but it's not pretty:

  

  I'd go as far as to say this makes Rebol look bad to those that aren't familiar with it.

---

* Here's an example of empty lines on a CASE statement (for LOAD-HEADER)

Here's the zoomed-out look at each decision point in the CASE statement

Then when you unfold a branch, it's delimited by those empty lines:

[Oldes]

It was merged #153

[Oldes]

Just there is not your name in the commit, because you did not started the pull request. I wonder if it is possible from Gitlab to GitHub.

[rgchris]

I'm not bothered :)

Unfortunately because I'd contributed to Ren-C in the past, I can't clone another branch of rebol/rebol on GitHub (perhaps that rule has changed/will change?)

[Oldes]

You don't have to clone. Instead you just add another remote repository and create a new branch from it. But it must be done locally, not from Github's web.

[rgchris]

I found another way—to fork into a separate organization. I don't know if that's better or worse, but it seems I was able to fork and submit a pull request 🤞