Making extension texts more uniform

I finally made some time to deal with #361 and such. But it needs some preparations.
As i said in #361 - it makes sense to first un-consider irrelevant extension. And obviously, i want to at least somewhat automate this process.
I already have programs to data-scrap most of relevant here info, but they work in a lot of hacky ways, with a lot of ways for me to miss something. I think i can redo it in a lot better way.

The main problem with data-scraping texts of extensions is that they aren't uniform. For instance:

1. In the end of `EXT_packed_depth_stencil.txt`:
```
Revision History

#12 September 26, 2005
- Fix copy/paste error in the example.
```
In most of places tabulation is kept in a way, where only section headers have 0 indentation before them. And so simplest algorithm would see `#12 September 26, 2005` as name of new section. And it's not as easy as treating lines starting from `#` as non-section-starter's, i'ev only shown simplest example here.

2. You can say "then make whitelist-like rules, what can be considered section name". But they aren't uniform too. For some reason, everyone loves calling `New Procedures and Functions` their own way. There is even random line break in the middle of it's name in one of extensions - `SGIX_slim.txt`.

---

So, i can either make pre-parsing text transformation rules, or go through extension texts a few times with small programs, making 1 thing more uniform at a time (for instance first fix identation's, then separately make section names more uniform).
Second option only makes sense if i'm also going to submit fixed version of extension files to this repository.

**TL;DR**: Does pull have a chance of being merged, if it wouldn't affect the meaning of extension texts, but would make them look better?
And if yes - are there things i absolutely should not touch?