Update credits for 5.2.0

[rubenwardy]

The generation of the credits was automated using a script found in a comment below

Point system

An author gets:

• 8 points for a code-changing commit above 1200 changes.
• 4 points for a code-changing commit above 700 changes.
• 2 points for a code-changing commit above 100 changes.
• 1 point for any other code-changing commit.

A code-changing commit is one that changes at least one file with one of the following extensions: .cpp, .h, .lua, .cmake, .gradle, Makefile

Criteria for Active Contributors

Active contributors must have at least 3 points since 5.1.0.

This means that to appear in the credits as an "Active Contributor", since 5.1.0 you need to have either:

• Made a commit with more than 700 changes
• OR made 2 commits with more than 100 changes
• OR made at least 3 commits

Criteria for Previous Contributors

Previous contributors must have at least 21 points all-time, and not be in the active contributors section.

This means that to appear in the credits as a "Previous Contributor", you need to have either:

• 2 commit of 3000 lines, and 1 other commit.
• OR 6 commits of 700 lines.
• OR 11 commits of 100 lines.
• OR made at least 21 commits.

[rubenwardy]

I'm unsure what the requirements to be in "previous contributors" should be, and how to automate this

[Wuzzy2]

I would be careful when droping people for “only” having contributed “fixes”.
Bugfixes can still require a lot of high-skill work when it's a complex, high-profile, non-trivial bug.
Not all bugfixes are created equal.
Look at what kind of bugfixes these people contributed.

Just my 5 cents.

[Wuzzy2]

BTW: I'm still an active translator for Minetest, it's just that Weblate did not update so far (translation files on Weblate need updating, as always). :P

[rubenwardy]

I've modified the script I used to make it apply for all commits in the repository.

For previous contributors, I got to score 33, which is equal to 33 commits of less than 100 lines, 4 commits of 100 lines, 2 commits of 1000 lines, 1.1 commits of 3000 lines. This is for all-time of Minetest.

Here's a list of the contributors that have dropped off of active and previous contributors. (I've almost definitely made a mistake here)

"Diego Martínez (kaeza) <kaeza@users.sf.net>",
"neoascetic [OS X Fixes]",
"Guiseppe Bilotta (Oblomov) <guiseppe.bilotta@gmail.com> [Fixes]",
"matttpt <matttpt@gmail.com> [Fixes]",
"bigfoot547 [CSM]",
"DTA7 [Fixes, mute key]",
"Thomas-S [Disconnected, Formspecs]",
"noob3167 [Fixes]",
"adelcoding1 [Formspecs]",
"Esteban (EXio4) [Cleanups]",
"Vaughan Lapsley (vlapsley) [Carpathian mapgen]",
"CoderForTheBetter [Add set_rotation]",
"Gaël-de-Sailly [Mapgen, pitch fly]",
"zeuner [Docs, Fixes]",
"ThomasMonroe314 (tre) [Fixes]",
"Rob Blanckaert (basicer) [Fixes]",
"Jozef Behran (osjc) [Fixes]",
"Pedro Gimeno (pgimeno) [Fixes]",
"lisacvuk [Fixes]",
"Ezhh/Shara",

[rubenwardy]

#!/bin/bash

OUT=$PWD
rm $OUT/results.txt

TMP=$PWD/tmp
mkdir -p $TMP
mkdir -p $TMP/emails
rm $TMP/*

merge() {
	echo Merging $1 into $2
	cat "$TMP/$1.dat" >> "$TMP/$2.dat"
	rm "$TMP/$1.dat"
}

cd ~/dev/minetest

entries=`git log --pretty=format:"%h	%an	%ae"` # "4e4fb4db2adcd0b3abd02cd6390b350b1129cc64..HEAD"`

echo "Gathering commit data..."

IFS=$'\n'
for line in $entries
do
	hash=$(echo "$line" | cut -f1)
	author=$(echo "$line" | cut -f2 | tr -dc '[:alnum:]\n\r' | tr '[:upper:]' '[:lower:]')
	email=$(echo "$line" | cut -f3)
	email_safe=$(echo "$email" | sed 's/@/.at./g' | tr -dc '[:alnum:].' | tr '[:upper:]' '[:lower:]')

	if [ -f "$TMP/emails/$email_safe.email" ]; then
		author=$(cat $TMP/emails/$email_safe.email | sed 's/^ *//;s/ *$//')
	fi
	echo $author > $TMP/emails/$email_safe.email

	commit=$(git show $hash --numstat --pretty=format:"")

	if echo $commit | grep -iq "\.cpp\|\.h\|\.lua\|\.cmake\|\.gradle\|Makefile"; then
		commit=$(echo "$commit" | sed 's/[^0-9\t ]*//g' | sed 's/[\t ]/\n/g' )
		lines=$(echo "$commit" | awk '{s+=$1} END {print s}')
		# echo $lines
		if (( $lines > 1200 )); then
			echo "8" >> "$TMP/$author.dat"
		elif (( $lines > 700 )); then
			echo "4" >> "$TMP/$author.dat"
		elif (( $lines > 100 )); then
			echo "2" >> "$TMP/$author.dat"
		else
			echo "1" >> "$TMP/$author.dat"
		fi
	fi
done || exit $?

merge you hybriddog
merge locblot loicblot
merge dnieljuhsz juhdanad
merge vitaliy numberzero
merge andrewward rubenwardy
merge paulouellette pauloue
merge jeanpatrickguerrero kilbith
merge vanessadannenberg vanessae
merge anands clobberxd
merge aukekok sofar
merge nathanalcourant ekdohibs
rm $TMP/translations.dat
rm $TMP/updatescript.dat

echo "Summing..."

cd $TMP
for filename in *.dat; do
	sum=$(awk '{s+=$1} END {print s}' "$filename")
	echo "$sum	$filename" >> $OUT/results.txt
done || exit $?

echo "Sorting..."

sort -gr $OUT/results.txt > $OUT/results_sorted.txt

[sfan5]

it's just that Weblate did not update so far (translation files on Weblate need updating, as always)

I updated those yesterday night.
In the German translation there were just two changed strings that were marked as fuzzy.

[rubenwardy]

I would like to add a separate section for translators, to avoid lots of [Translations]. I'm not sure how to get a list, it's also past string freeze

[Wuzzy2]

Ah, so it looks like German was already moved back to 100% very fast, probably by sfan5. :D
I always assume that as long German is at 100%, Weblate was not updated already. :D But actually, the real indicator on Weblate to see whether the source strings are up-to-date is the red exclamation mark next to the resource. When it's there, it's outdated, when it's not there, it's up-to-date. It's not there anymore. Good.

Anyway, I'm still active translator because I occassionally fix typos when I run across them (rare). Otherwise, the German translation is finished and tried and tested, there's not much work to do anymore. :D I'm just saying it because I noticed my translator status was snatched away from me. xD

Although, it's not that of a big deal to me. But I wanted to give you a heads-up.

[Wuzzy2]

Hmmm, thinking about this: I think the process of dropping people from the credits entirely just because they contributed years ago isn't fair. It's also just factually wrong. If their contribution is still in the game, it still counts IMO. Dropping them would only be justified if their contribution was removed for some reason.

Maybe move those contributors that didn't make the cut into a separate text file and mention that file in the in-game credits. The section “previous contributors” is also misleading because it's not complete.

But that's a problem I noticed way too late, I guess. I'm assuming you have dropped past contributors from the credits in older releases as well? That would mean it could take a lot of work to piece together the full, TRUE credits. I'm predicting the true credits would be absolutely huge. :D

Just my 2 cents. It's not that I really care about this stuff to be honest, but since you are concerned about fairness, I wanted to share my thoughts about this.

[rubenwardy]

How I've done it now is that time doesn't matter at all as to whether you appear in previous contributors - you just need to meet a threshold based on the heuristic documented in the first post

[rubenwardy]

The much bigger problem is not crediting translators and documenters. Translations are excluded from the heuristic because they typically generate large numbers of commits. It makes sense to collect the names of translators separately to the names of other contributors for that reason, depending on how many there are they could all be credited

The heuristic should be changed to include md as a code file, that probably makes sense

[Wuzzy2]

Good point!

[sfan5]

Looks good.

[sfan5]

Crediting translators properly is something we'll look into for 5.3.0.