extras/misc/authors_list.sh

   1 #!/bin/bash
   2
   3 # A script checking the git logs for commits. Final goal is updating AUTHORS.
   4 # Run it in source root
   5
   6 # To be copied and run in the git directory for having "git shortlog -sn po/" find the logs.
   7 # It will generate a subdirectory temp_update_AUTHORS
   8
   9 OLD_AUTHORS="AUTHORS_unmodified_by_No_se_script.txt"
  10 # The last version before modifying AUTHORS with this kind of script. Important, since the script shall not remove anyone.
  11 # Not even the manual editings delete anyone.
  12 # 4f696a88ec9544b98e22ee45e010869717608bd2  here did j-b start using the script
  13
  14 if [ -f $OLD_AUTHORS ]; then
  15  echo "Starting with credits from $OLD_AUTHORS ..."
  16 else
  17  echo "You need an old AUTHORS file. Only if you know what you are doing, you can use the current AUTHORS."
  18  echo "Aborting..."
  19  exit
  20 fi
  21
  22
  23 mkdir -p temp_update_AUTHORS
  24
  25 echo "Checking all git logs"
  26 git shortlog -sn -- > temp_update_AUTHORS/all_git.txt
  27
  28 echo "Checking "po only" git logs"
  29 #git shortlog -sn po/ > temp_update_AUTHORS/po_git.txt
  30 # This modified command identified more translators, leading to the removal of existing entries in AUTHORS, impossible by design.
  31 # So, start with a AUTHORS version with validated entries.
  32 # Only if the script is not modified (wrt finding translators) you can use the current AUTHORS
  33 git shortlog -sn -- po extras/package/win32/languages/ share/vlc.desktop.in  share/vlc.desktop share/applications/vlc.desktop extras/package/win32/vlc.win32.nsi.in > temp_update_AUTHORS/po_git_previous.txt
  34
  35 # Some typical (ancient) l10n files:
  36 #share/applications/vlc.desktop
  37 #extras/package/win32/languages/declaration.nsh
  38 #extras/package/win32/languages/english.nsh
  39 #extras/package/win32/languages/french.nsh
  40 #extras/package/win32/vlc.win32.nsi.in
  41
  42 sed -n '{
  43 s/.*Song Ye Wen.*//
  44 s/.*Florian Hubold.*//
  45 s/.*Sveinung Kvilhaugsvik.*//
  46 s/.*Julien Humbert.*//
  47 /^$/ !p
  48 }
  49 ' <temp_update_AUTHORS/po_git_previous.txt >temp_update_AUTHORS/po_git.txt
  50
  51 # Checking the logs, this ^^ persons did not do l10n
  52 # commited to share/vlc.desktop, but this seems to be media types, not l10n things.
  53 # Florian Hubold
  54 # share/applications/vlc.desktop
  55 # Sveinung Kvilhaugsvik
  56 # commited to extras/package/win32/vlc.win32.nsi.in   and   include/vlc_interface.h
  57 # Julien Humbert
  58 # comitted to po/POTFILES.in
  59 # Song Ye Wen
  60
  61 # TODO: add this newly discovered translators
  62 # czech translator
  63 # Radek Vybiral <radek@ns.snake.cz>
  64 # zh_TW
  65 # Thanks to Hsi-Ching Chao
  66 # Thanks to Ruei-Yuan Lu <RueiYuan.Lu@gmail.com>
  67 #extras/package/win32/languages/schinese.nsh
  68 #share/vlc.desktop
  69
  70
  71 # there are some artwork designers in git log, too. If one of them wants to be mentioned in "Programmers" also, remove here:
  72 echo "Damien Erambert"  >  temp_update_AUTHORS/artwork_git.txt
  73 echo "Daniel Dreibrodt" >> temp_update_AUTHORS/artwork_git.txt
  74 echo "Dominic Spitaler" >> temp_update_AUTHORS/artwork_git.txt
  75
  76
  77
  78
  79 echo "reading AUTHORS"
  80 sed -n '/Programming/,/^$/  s/[^-].*/&/p' < $OLD_AUTHORS | sed '1 d'  > temp_update_AUTHORS/programmers_part.txt
  81 # The part of AUTHORS between Programming and the first empty line, without the ---- line
  82
  83
  84 echo "Removing commit counts from git log"
  85 sed 's/[0-9 \t]*\(.*\)/\1/g' < temp_update_AUTHORS/all_git.txt |sort|uniq > temp_update_AUTHORS/all_git_namesonly.txt
  86 # I think "uniq" is not needed here.
  87
  88
  89 echo "Removing translators from the git log"
  90 # Remove translators. (Commiters with the same count in /po and total and hence are listed twice). Then the commit counter is removed
  91 cat temp_update_AUTHORS/all_git.txt temp_update_AUTHORS/po_git.txt|sort|uniq -u |sed 's/[0-9 \t]*\(.*\)/\1/g' | sort|uniq> temp_update_AUTHORS/coders_only.txt
  92
  93
  94 ####
  95 #For script tuning: Are there other files the translators modified? =>probably l10n files
  96 cat temp_update_AUTHORS/all_git.txt temp_update_AUTHORS/po_git.txt|sort|uniq -u| sed 's/[0-9 \t]*\(.*\)/\1/g' | sort|uniq -d> temp_update_AUTHORS/both_sides.txt
  97 ####
  98
  99 # Similar effect with second sed run:
 100 # Remove translators. I remove every line containing the name. Maybe the .* before and after the last \1 should be removed (i.e. for contributors "Firstname Secondname aka something_you_want_to_keep"
 101 #cat temp_update_AUTHORS/all_git.txt temp_update_AUTHORS/po_git.txt|sort|uniq -D|uniq|sed 's/[0-9 \t]*\(.*\)/\1/g' |sed 's:[0-9 \t]*\(.*\):s^.*\1.*^^g:' > temp_update_AUTHORS/remove_translators_gen
 102 #
 103 #sed -f temp_update_AUTHORS/remove_translators_gen < temp_update_AUTHORS/all_git_namesonly.txt > temp_update_AUTHORS/coders_only.txt
 104 # This is everyone who did code commits with git. The blank lines are the removed translators.
 105
 106
 107 # Now, I want to reduce the number of lines the human reader has to check, so we are going to kill the already listed contributors.
 108
 109 echo "Finding pre-git contributors in AUTHORS"
 110 sed 's:\(.*\):s^.*\1.*^^g:' < temp_update_AUTHORS/coders_only.txt > temp_update_AUTHORS/remove_git_commiters_gen
 111 sed -f temp_update_AUTHORS/remove_git_commiters_gen < temp_update_AUTHORS/programmers_part.txt |sort| uniq -u > temp_update_AUTHORS/pre-git.txt
 112
 113 sed 's:\(.*\):s^.*\1.*^^g:' < temp_update_AUTHORS/programmers_part.txt > temp_update_AUTHORS/remove_programmers_part_gen
 114 sed -f temp_update_AUTHORS/remove_programmers_part_gen < temp_update_AUTHORS/coders_only.txt |sort| uniq -u > temp_update_AUTHORS/new_coders_only.txt
 115
 116 # VideoLAN as a contributor can be removed, I think
 117
 118
 119
 120 sed 's/[0-9 \t]*\(.*\)/\1/g' < temp_update_AUTHORS/all_git.txt  > temp_update_AUTHORS/all_git_namesonly_ordered.txt
 121 # Just remove the tab an the # commits, keep the order. This file is going to be the sort order.
 122 # You can re-order the complete programmers part like that and simply append the pre-git commiters.
 123 # Ordering the contributors that way is not a bad idea.
 124 # The question: Is it easier/better to check the new commiters in this order?
 125 # One can find (UPPERCASE issues, middle names,..) better when listing them alphabetically.
 126
 127 # I suggest checking manually a file build like that:
 128 # alphabetically ordered, complete list of contributors/git (code) commiters, with an extra marking for new ones. Example
 129 # Old Commiter
 130 # New Commiter           ---XXX---NEW
 131 # New COMMITER           ---XXX---NEW   |same name with UPPERCASE part
 132 # Very Commiter
 133 # Very New Commiter      ---XXX---NEW   |Same person with a middle name
 134
 135 # The uppercase case can be done by script I guess, I did not look up how to make sure the intended version will be picked.
 136 # How to proceed with manually found problems? Solve them for the future (.mailmap/own script/...)
 137
 138 rm -f temp_update_AUTHORS/ordering_log.txt
 139 rm -f temp_update_AUTHORS/ordered_by_commits.txt
 140 FileName='temp_update_AUTHORS/all_git_namesonly_ordered.txt'
 141 while read LINE
 142 do
 143  if [ "$LINE" = "VideoLAN" ]; then
 144   echo "VideoLAN is not a person"
 145  else
 146 #  grep "$LINE" temp_update_AUTHORS/new_coders_only.txt >> temp_update_AUTHORS/ordering_log.txt
 147   grep "$LINE" temp_update_AUTHORS/coders_only.txt >> temp_update_AUTHORS/ordering_log.txt
 148 #  grep "$LINE" temp_update_AUTHORS/coders_only.txt >> temp_update_AUTHORS/ordering_log.txt
 149 # I want to keep the $? (it removes some broken names) but I could send the output to /dev/null
 150 # If someone's name is a prefix to some other's name, this diff will show it:
 151 # diff temp_update_AUTHORS/ordering_log.txt temp_update_AUTHORS/ordered_by_commits.txt
 152 # AFAIK this will not effect the output, since we don't use the grep output but only the git output
 153   if [ $? = "0" ]; then
 154     grep "$LINE" temp_update_AUTHORS/artwork_git.txt ||  echo "$LINE" >> temp_update_AUTHORS/ordered_by_commits.txt
 155   fi
 156  fi
 157 done < $FileName
 158
 159
 160
 161 cat temp_update_AUTHORS/all_git.txt temp_update_AUTHORS/po_git.txt|sort|uniq -D|uniq|sed 's/[0-9 \t]*\(.*\)/\1/g' > temp_update_AUTHORS/translators.txt
 162 wc -l temp_update_AUTHORS/*
 163 echo "Some contributors only commited into po. Please cross-check that with the localization part. See: temp_update_AUTHORS/translators.txt"
 164 echo "But first, please check if temp_update_AUTHORS/review.txt contains complete names and other constraints for publishing (i.e. UPPERCASE name parts, broken text, a name and it's abbreviation both present...)"
 165
 166
 167 sed 's/\(.*\)/\1               ---XXX---NEW/g' < temp_update_AUTHORS/new_coders_only.txt | cat - temp_update_AUTHORS/programmers_part.txt |sort > temp_update_AUTHORS/review.txt
 168 # This file contains VideoLAN as a contributor.
 169
 170 echo
 171 echo "For the lazy ones: Have a look at temp_update_AUTHORS/final.txt"
 172 echo "Contains all git code commiters (the translators are stored somewhere else) sorted by commits, and the pre-git commiters"
 173 echo
 174 echo 'For the lazy and brave (stupid?) ones: "cp new_AUTHORS AUTHORS" and check "git diff AUTHORS"'
 175
 176 echo "Programming" >  temp_update_AUTHORS/final.txt
 177 echo "-----------" >> temp_update_AUTHORS/final.txt
 178 cat temp_update_AUTHORS/ordered_by_commits.txt temp_update_AUTHORS/pre-git.txt >> temp_update_AUTHORS/final.txt
 179 echo
 180
 181 echo "Listing email adresses used with different names..."
 182 git shortlog -sne |sed 's/[^<]*\(.*\)/\1/g' |sort|uniq -d
 183 echo "If something was listed here you should probably modify .mailmap"
 184
 185 # This last part puts the relevant addresses into temp_twice_used_adress/check_for_this.txt
 186 # Currently, this is not needed (.mailmap is up to date)
 187
 188 #mkdir -p temp_twice_used_adress
 189 #echo "Checking all git logs"
 190 #git shortlog -sne > temp_twice_used_adress/all_shortlog_sne.txt
 191 #
 192 #echo "Removing everything but email addresses"
 193 #sed 's/[^<]*\(.*\)/\1/g' < temp_twice_used_adress/all_shortlog_sne.txt |sort|uniq -d > temp_twice_used_adress/all_git_addresses_only.txt
 194 #
 195 #FileName='temp_twice_used_adress/all_git_addresses_only.txt'
 196 #while read LINE
 197 #do
 198 #  grep "$LINE" temp_twice_used_adress/all_shortlog_sne.txt >> temp_twice_used_adress/check_for_this.txt
 199 ## I want to keep the $? (it removes some broken names) but I could send the output to /dev/null
 200 #  if [ $? = "0" ]; then
 201 #    echo "$LINE"
 202 #  fi
 203 #done < $FileName
 204
 205
 206
 207 sed -n '1,2  p' < AUTHORS   > temp_update_AUTHORS/new_AUTHORS
 208 cat temp_update_AUTHORS/final.txt >> temp_update_AUTHORS/new_AUTHORS
 209 echo >>temp_update_AUTHORS/new_AUTHORS
 210 sed -n '/Artwork/,$ p' < AUTHORS >> temp_update_AUTHORS/new_AUTHORS
 211
 212 cp temp_update_AUTHORS/new_AUTHORS .
 213 #rm -rf temp_update_AUTHORS/