Sort-of-pseudo-automated the Legacy list

[DandelionSprout]

As per DandelionSprout/adfilt#7 (comment) and some comments prior and afterwards.

Notes

I have zero idea how to make line = re.sub select text from multiple lines, which means that I am not able to remove !#if env_chromium- or !#if !env_firefox-dependant entries.

I also have zero idea how to make Python import local text files instead of URL sources, meaning that the script must be run twice to achieve the correct results. This is detailed in the README.md file.

The README.md file's instructions are based on downloading the file and running it locally with a file explorer and a terminal. If you guys were to figure out a way to do it in Git or Azure or whatever you guys use, you can write instructions for those methods beneath the existing instructions.

[mapx-]

for nostif, nosiif etc see https://github.com/gorhill/uBlock/wiki/Resources-Library

[DandelionSprout]

Okay, now I think I've got these things right in general.

[JustOff]

I truly appreciate the effort to resolve the issue with legacy filters, but I'd like to warn against using the proposed solution in its current state, since I see at least two significant gaps:

• The resulting legacy.txt contains invalid lines, such as, but not limited to:

 `google-analytics.com` in Peter Lowe's list
 Dan Pollock's, hpHosts, MVPS, Peter Lowe's
 `statcounter.com` in Peter Lowe's, MVPS, Dan Pollock's and hpHosts
 `2mdn.net` in Peter Lowe's
 `s0.2mdn.net` in MVPS, hpHosts, Dan Pollock's
 `adm.fwmrm.net` from hpHosts, MVPS, Dan Pollock's
 `mopub.com` in Peter Lowe's
 `safelinking.net` in hpHosts
 `outbrain.com` in hpHosts
 `flurry.com` in Peter Lowe's, hpHosts, MVPS
 `2mdn.net` in Peter Lowe's, `s0.2mdn.net` in hpHosts
 Peter Lowe's, hpHosts
 `yui.yahooapis.com` in hpHosts

• The resulting legacy.txt always includes filters converted from all built-in uBlock filter lists (Annoyances, Badware risks, Privacy, Resource abuse, Unbreak), regardless of whether the user has enabled them or not.

[THEtomaso]

The resulting legacy.txt contains invalid lines

So, there a few bugs to weed out, but why would you try to convert anything from those hosts files?
They only contain network filters, none of which needs convertion!

--

The resulting legacy.txt always includes filters converted from all built-in uBlock filter lists

Oh, okiehsch specifically pointed out only to include rules from uBlock filters, and only the ones that doesn't work with uBO Legacy anymore:
https://github.com/DandelionSprout/adfilt/issues/7/#issuecomment-572953302

For the other built-in uBlock filter lists, additional Legacy versions are needed, like I suggested here:
https://github.com/DandelionSprout/adfilt/issues/7/#issuecomment-573006225
This will require a uBO Legacy extension update too, of course.

[DandelionSprout]

I am aware that there's a handful of remnant comment lines that somehow haven't been properly !-ed, and I know of a fix for it, but I can't add it myself before 23:30 CET tonight due to handball spectating commitments.

As for which file(s) to use, that feels like something that I'm not the right guy to debate. I tried most of all to have it fit well into the existing legacy list's infrastructure.

[THEtomaso]

As for which file(s) to use, that feels like something that I'm not the right guy to debate.

What okiehsch ment is that uBlock filters - Legacy should only include rules converted from uBO's main filter list, and none from the specialized ones:

[JustOff]

I'd suggest instead to consider the option to convert new shorthands back to legacy syntax directly in uBlock when retrieving filter lists.

Here is a simple hack doing so, based on rules taken from LegacyJSconverter.py.

[JustOff]

Since the converter was added today to firefox-legacy-1.16.4.14, this PR has become obsolete.

Thanks again to @DandelionSprout for immediately offering help in resolving the legacy filter issue, as well as for the LegacyJSconverter.py that I used to create the converter embedded into uBlock.