BEGIN {
	# The current state of the parser:
	# -1 -> haven't read the first line of the new block
	# 0  -> reading the pre-comment
	# 1  -> reading the msgid
	# 2  -> reading the msgstr
	state=-1
	# The comment preceding the block
	precomment=""
	# The same string, but with a space after the sharp to avoid a comment
	sharpedprecomment=""
	# The msgid section, containing the string to be translated
	msgidstr=""
	# The same string, but with a sharp before every newline (commented  out)
	sharpedmsgidstr=""
	# The msgstr section, containing the destination string
	msgstrstr=""
	# The same string, but with a sharp before every newline (commented out)
	sharpedmsgstrstr=""
	# Whether the block being read should be kept
	# -1 -> should keep, overridable
	# 0  -> should not keep, overridable
	# 1  -> should keep, not overridable
	acceptable=1
	# The file location of where to put everything
	# that has been stripped out
	untranslatablefile="./untranslatable.po"
	# Wether we are still reading the licence text
	licencetext=1
}

# Detecting the end of licence
$0~/^#, fuzzy$/ {licencetext=0; next; next}

# If we are reading the licence, skip text and dont print it.
$0~// {if (licencetext==1) {next}}

# We pass on the wrong metadata
$0~/"Report-Msgid-Bugs-To:/ {next}
$0~/"PO-Revision-Date:/ {next}
$0~/"Last-Translator:/ {next}
$0~/"Language-Team:/ {next}

# This pattern matches empty lines
# The code flushes the data stored, and save
# it only if acceptable!=1
$0~/^$/ {
	if (state<=0){
		if(acceptable!=1){
			print precomment
		}else{
			#print "#  Detected untranslatable text:"
			#print sharpedprecomment
		}
		precomment=""
	}else{
		if(acceptable==1){
			print precomment
			print msgidstr
			print msgstrstr
			print ""
		}else{
			#print "#  Detected untranslatable text:"
			#print sharpedprecomment
			#print sharpedmsgidstr
			#print sharpedmsgstrstr
			print precomment>>untranslatablefile
			print msgidstr>>untranslatablefile
			print msgstrstr>>untranslatablefile
		}
		# Add the newline currently parsed
		# Re-initialisation of the variables.
		state=-1
		acceptable=-1
		precomment=""
		msgidstr=""
		msgstrstr=""
	}
}
# If the line is commented out
$0~/^#/ {
	precomment=(state==-1)?$0:precomment"\n"$0
	sharpedprecomment=(state==-1)?"#  "$0:sharpedprecomment"\n#  "$0
	state=0
}
# If the line starts with "msgid"
$0~/^msgid/ {
	state=1
	msgidstr=$0
	sharpedmsgidstr="#  "$0
}
# If the line starts with msgstr
$0~/^msgstr/ {
	state=2
	msgstrstr=$0
	sharpedmsgstrstr="#  "$0
}
# If the line starts with a '"'
$0~/^\"/ {
	if(state==1){
		msgidstr=msgidstr"\n"$0
		sharpedmsgidstr=sharpedmsgidstr"\n#  "$0
	}else{
		msgstrstr=msgstrstr"\n"$0
		sharpedmsgstrstr=sharpedmsgstrstr"\n#  "$0
	}
}
# ----------------------------------------------------------------


# This code is now the part that actually selects lines to be stripped out.

state==1 {
	if($0~/^(msgid ""|"")$/){

	}else if($0~/^(msgid "|")((:ref:`[a-zA-Z]*`)|(:obj:)|(manim.([a-z._\\]+)"$)|(((:(mod|class|func):`\~?\.[a-zA-Z0-9._]+)`| )+))/ ) {
		acceptable=(acceptable==-1)?0:acceptable
	}else{
		acceptable=1
	}
}
$0~/^msgid ":ref:`[a-zA-Z]*`"/ {
	acceptable=0
}
$0~/^msgid ":obj:/ {
	acceptable=0
}
$0~/^msgid "manim.([a-z._\\]+)"$/ {
	acceptable=0
}
$0~/^msgid "((:(mod|class|func):`~\.[a-zA-Z0-9.]+)`| )+"/ {
	acceptable=0
}
$0~/^"((:(mod|class|func):`~\.[a-zA-Z0-9.]+)`| )+"/ {
	acceptable=0
}

# When the parsing is ended, print the last missing endline
END {
print ""
}