From c543f42fe50d9b10175059ee5ed9186f30445f40 Mon Sep 17 00:00:00 2001
From: Lars-Dominik Braun <lars@6xq.net>
Date: Thu, 14 Nov 2019 17:17:56 +0100
Subject: doc: Process HTML template with m4

Seems to be more reliable than C preprocessor, dependency tracking never
really worked anyway.
---
 doc/index.html | 16 ++++++++--------
 gen.sh         | 23 ++++++++++++++++-------
 2 files changed, 24 insertions(+), 15 deletions(-)

diff --git a/doc/index.html b/doc/index.html
index e930892..b0fa3a1 100644
--- a/doc/index.html
+++ b/doc/index.html
@@ -119,7 +119,7 @@
 		The corpus used for the following analysis consists of
 		</p>
 
-		#include "corpus.html"
+		include(`corpus.html')
 
 		<p>
 		The plot below shows <bdo dir="ltr" lang="ar">ا ل ي م و ن</bdo> can be
@@ -163,7 +163,7 @@
 <figure id="ar-lulua-heat">
 <div class="lbox" lang="en">
 	<img src="ar-lulua-heat.svg">
-	#include "ar-lulua-fingerhand.html"
+	include(`ar-lulua-fingerhand.html')
 </div>
 </figure>
 
@@ -181,7 +181,7 @@
 	<figure id="ar-asmo663">
 	<div class="lbox">
 		<img src="ar-asmo663-heat.svg">
-		#include "ar-asmo663-fingerhand.html"
+		include(`ar-asmo663-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
@@ -206,7 +206,7 @@
 	<figure id="ar-linux">
 	<div class="lbox">
 		<img src="ar-linux-heat.svg">
-		#include "ar-linux-fingerhand.html"
+		include(`ar-linux-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
@@ -231,7 +231,7 @@
 	<figure id="ar-malas">
 	<div class="lbox">
 		<img src="ar-malas-heat.svg">
-		#include "ar-malas-fingerhand.html"
+		include(`ar-malas-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
@@ -262,7 +262,7 @@
 	<figure id="ar-osman">
 	<div class="lbox">
 		<img src="ar-osman-heat.svg">
-		#include "ar-osman-fingerhand.html"
+		include(`ar-osman-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
@@ -295,7 +295,7 @@
 	<figure>
 	<div class="lbox">
 		<img src="ar-khorshid-heat.svg">
-		#include "ar-khorshid-fingerhand.html"
+		include(`ar-khorshid-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
@@ -325,7 +325,7 @@
 	<figure>
 	<div class="lbox">
 		<img src="ar-phonetic-heat.svg">
-		#include "ar-phonetic-fingerhand.html"
+		include(`ar-phonetic-fingerhand.html')
 	</div>
 	<figcaption class="pure-g flexreverse">
 		<div class="pure-u-1 pure-u-md-1-2">
diff --git a/gen.sh b/gen.sh
index de1fed0..589035a 100755
--- a/gen.sh
+++ b/gen.sh
@@ -90,10 +90,8 @@ rule analyze-corpushtml
 rule wordlist
     command = lulua-analyze -l ar-lulua latinime < \$in > \$out
 
-rule cpp
-    command = gcc -E -x c -nostdinc -MMD -MF \$out.d -C -P -I \$docdir/_temp \$in -o \$out
-    depfile = \$out.d
-    deps = gcc
+rule html
+    command = m4 -I \$docdir/_temp \$template > \$out
 
 rule cp
     command = cp \$in \$out
@@ -105,7 +103,6 @@ rule gz
 build \$docdir/_build: mkdir
 build \$docdir/_build/fonts: mkdir
 build \$docdir/_temp: mkdir
-build \$docdir/_build/index.html: cpp \$docdir/index.html || \$docdir/_build
 build \$docdir/_build/letterfreq.json: letterfreq \$statsdir/ar-lulua/all.pickle || \$docdir/_build
 build \$docdir/_build/style.css: cp \$docdir/style.css || \$docdir/_build
 build \$docdir/_build/lulua-logo.svg: cp \$docdir/lulua-logo.svg || \$docdir/_build
@@ -116,6 +113,7 @@ build \$docdir/_build/lulua.combined.gz: gz \$docdir/_temp/lulua.combined || \$d
 
 build \$docdir/_build/fonts/IBMPlexArabic-Regular.woff2: cp \$fontdir/IBMPlexArabic-Regular.woff2 || \$docdir/_build/fonts
 build \$docdir/_build/fonts/IBMPlexArabic-Thin.woff2: cp \$fontdir/IBMPlexArabic-Thin.woff2 || \$docdir/_build/fonts
+
 EOF
 
 # targets for every layout
@@ -138,10 +136,10 @@ build \$statsdir/${l}/tanzil-quaran.pickle: write-tanzil \$corpusdir/tanzil-quar
 build \$statsdir/${l}/arwiki.pickle: write-arwiki \$corpusdir/arwiki/arwiki-20190701-pages-articles.xml.bz2 || \$statsdir/${l}
     layout = ${l}
 
-build \$statsdir/${l}/un-v1.0-tei.pickle: write-tei2 \$corpusdir/un-v1.0-tei || \$statsdir/${l}
+build \$statsdir/${l}/un-v1.0-tei.pickle: write-tei2 \$corpusdir/un-v1.0-tei/raw || \$statsdir/${l}
     layout = ${l}
 
-build \$statsdir/${l}/opensubtitles-2018.pickle: write-opensubtitles \$corpusdir/opensubtitles-2018 || \$statsdir/${l}
+build \$statsdir/${l}/opensubtitles-2018.pickle: write-opensubtitles \$corpusdir/opensubtitles-2018/raw || \$statsdir/${l}
     layout = ${l}
 
 build \$statsdir/${l}/all.pickle: combine \$statsdir/${l}/bbcarabic.pickle \$statsdir/${l}/aljazeera.pickle \$statsdir/${l}/tanzil-quaran.pickle \$statsdir/${l}/arwiki.pickle \$statsdir/${l}/hindawi.pickle \$statsdir/${l}/un-v1.0-tei.pickle \$statsdir/${l}/opensubtitles-2018.pickle || \$statsdir/${l}
@@ -159,6 +157,8 @@ build \$docdir/_temp/${l}-fingerhand.html: analyze-fingerhand \$statsdir/${l}/al
     layout = ${l}
 
 EOF
+# included by index.html and thus must be its dependencies
+fingerhandfiles+=" \$docdir/_temp/${l}-fingerhand.html"
 done
 
 # layouts with xmodmap support
@@ -177,11 +177,20 @@ cat <<EOF
 build \$docdir/_temp/metadata-$c.yaml: analyze-corpusstats \$statsdir/ar-lulua/$c.pickle \$corpusdir/$c/metadata.yaml || \$docdir/_temp \$corpusdir/$c/metadata.yaml
     metadata = \$corpusdir/$c/metadata.yaml
     stats = \$statsdir/ar-lulua/$c.pickle
+
 EOF
 outfiles+=" \$docdir/_temp/metadata-$c.yaml"
 done
 
 cat <<EOF
 build \$docdir/_temp/corpus.html: analyze-corpushtml $outfiles || \$docdir/_temp
+
+EOF
+
+# html, which depends on several other files generated above
+cat <<EOF
+build \$docdir/_build/index.html: html \$docdir/index.html \$docdir/_temp/corpus.html $fingerhandfiles || \$docdir/_build
+    template = \$docdir/index.html
+
 EOF
 
-- 
cgit v1.2.3