From f7676bb9cabab8d408d220051e4857c054a55efd Mon Sep 17 00:00:00 2001
From: Christian Heller <c.heller@plomlompom.de>
Date: Sat, 10 Dec 2016 23:55:21 +0100
Subject: [PATCH] Stabilize feed.xml generation, reduce bad shell scripting.
---
README.md | 5 +-
processor/feed.xml.do | 86 ++++++++++++++++++-------------
test.sh | 10 ++--
test/test_files/feed.xml.ignoring | 20 +++----
test/test_files/index.html | 2 +-
test/test_files/test.html | 2 +-
test/test_files/test.md | 2 +-
7 files changed, 72 insertions(+), 55 deletions(-)
diff --git a/README.md b/README.md
index 92047da..e3ced3c 100644
--- a/README.md
+++ b/README.md
@@ -33,9 +33,8 @@ customize the blog: ./url, ./author, ./uuid, ./title.)
bugs
----
-Due to bad shell scripting in ./processor/index.html.do and
-./processor/feed.xml.do (see the FIXME notes there), source files whose names
-contain "$" break the redo processing.
+Due to bad shell scripting in ./processor/index.html.do (see the FIXME notes
+there), source files whose names contain "$" break the redo processing.
Oh, and don't create a index.rst or index.md in the redo-managed directory,
that will also break things.
diff --git a/processor/feed.xml.do b/processor/feed.xml.do
index 1e492eb..0010240 100644
--- a/processor/feed.xml.do
+++ b/processor/feed.xml.do
@@ -1,5 +1,29 @@
#!/bin/sh
+build_entry () {
+ file="${1}"
+ uuidfile="${2}"
+ published="${3}"
+ intermediate_file="${file%.*}.intermediate"
+ htmlfile=`escape_url "${file%.*}.html"`
+ redo-ifchange "$uuid_file"
+ redo-ifchange "$intermediate_file"
+ lastmod=`stat -c%y "$file"`
+ lastmod_rfc3339=`date -u "+%Y-%m-%dT%TZ" -d "$lastmod"`
+ title=`read_and_escape_file "$intermediate_file" | head -1`
+ uuid=`read_and_escape_file "$uuid_file" | head -1`
+ body=`read_and_escape_file "$intermediate_file" | sed 1d`
+ published_rfc3339=`date -u "+%Y-%m-%dT%TZ" -d "${published}"`
+ printf "<entry>\n"
+ printf "<title type=\"html\">%s</title>\n" "$title"
+ printf "<id>urn:uuid:%s</id>\n" "$uuid"
+ printf "<updated>%s</updated>\n" "$lastmod_rfc3339"
+ printf "<published>%s</published>\n" "$published_rfc3339"
+ printf "<link href=\"%s%s\" />\n" "$basepath" "${htmlfile#\./}"
+ printf "<content type=\"html\">\n%s\n</content>\n" "$body"
+ printf "</entry>"
+}
+
# Pull in global dependencies.
. ./helpers.sh
url_file=url.meta
@@ -33,44 +57,36 @@ printf "<title type=\"html\">%s</title>\n" "$title"
printf "<author><name>%s</name></author>\n" "$author"
printf "<id>urn:uuid:%s</id>\n" "$uuid"
-# Iterate through most recent entries (go by lastmod date of source files) to
-# build feed head "updated" element, and individual entries.
-# FIXME: This ls parsing is a bad way to loop through the sorted files. Besides,
-# $'\0' is a bashism.
-first_run=0
-files=`ls -1t *.rst *.md | head -10 | tr '\n' $'\0'`
-oldIFS="$IFS"
-IFS=$'\0'
-for file in $files; do
- lastmod=`stat -c%y "$file"`
- lastmod_rfc3339=`date -u "+%Y-%m-%dT%TZ" -d "$lastmod"`
- if [ "$first_run" -lt "1" ]; then
- IFS="$oldIFS"
- printf "<updated>%s</updated>\n\n" "$lastmod_rfc3339"
- first_run=1
+# Generate feed entry snippets.
+mkdir -p feed_snippets
+for file in ./*.rst ./*.md; do
+ if [ -e "$file" ]; then
+ uuid_file="${file%.*}.uuid"
+ redo-ifchange "$uuid_file"
+ published=`stat -c%y "${uuid_file}"`
+ published_unix=$(date -u "+%s%N" -d "${published}")
+ entry=$(build_entry "${file}" "${uuid_file}" "${published}")
+ echo "${entry}" > ./feed_snippets/${published_unix}
fi
+done
- # Build some variables and dependencies.
- intermediate_file="${file%.*}.intermediate"
- htmlfile=`escape_url "${file%.*}.html"`
- uuid_file="${file%.*}.uuid"
- redo-ifchange "$intermediate_file"
- redo-ifchange "$uuid_file"
- title=`read_and_escape_file "$intermediate_file" | head -1`
- uuid=`read_and_escape_file "$uuid_file" | head -1`
- body=`read_and_escape_file "$intermediate_file" | sed 1d`
- published=`stat -c%y "$uuid_file"`
- published_rfc3339=`date -u "+%Y-%m-%dT%TZ" -d "$published"`
+# Derive feed modification date from snippets.
+mod_dates=$(grep -hE "^<updated>" ./feed_snippets/* | sed -E 's/<.?updated>//g')
+last_mod_unix=0
+for date in $mod_dates; do
+ date_unix=$(date -u "+%s" -d "${date}")
+ if [ "$date_unix" -gt "$last_mod_unix" ]; then
+ last_mod_unix=$date_unix
+ fi
+done
+lastmod_rfc3339=`date -u "+%Y-%m-%dT%TZ" -d "@${last_mod_unix}"`
+printf "<updated>%s</updated>\n\n" "$lastmod_rfc3339"
- # Write entry.
- printf "<entry>\n"
- printf "<title type=\"html\">%s</title>\n" "$title"
- printf "<id>urn:uuid:%s</id>\n" "$uuid"
- printf "<updated>%s</updated>\n" "$lastmod_rfc3339"
- printf "<published>%s</published>\n" "$published_rfc3339"
- printf "<link href=\"%s%s\" />\n" "$basepath" "$htmlfile"
- printf "<content type=\"html\">\n%s\n</content>\n" "$body"
- printf "</entry>\n\n"
+# Write feed entries.
+for file in ./feed_snippets/*; do
+ cat "${file}"
+ printf "\n"
done
+rm -rf feed_snippets
printf "</feed>"
diff --git a/test.sh b/test.sh
index e333c40..d930e66 100755
--- a/test.sh
+++ b/test.sh
@@ -32,12 +32,14 @@ generated_files_dir_escaped="test\\/test_dir"
rm -rf "$generated_files_dir"
./add_dir.sh "$generated_files_dir"
working_dir=$(pwd)
-cp "$expected_files_dir"/test.md "$generated_files_dir"/
-cp "$expected_files_dir"/foo.rst "$generated_files_dir"/
-cp "$expected_files_dir"/bar\ baz.md "$generated_files_dir"/
cd "$generated_files_dir"
+cp "$working_dir/$expected_files_dir"/test.md .
redo
-cd "$working_dir"
+cp "$working_dir/$expected_files_dir"/bar\ baz.md .
+redo
+cp "$working_dir/$expected_files_dir"/foo.rst .
+redo
+cd "$working_dir"
# Simple file comparison tests and UUID tests.
uuid_test "$generated_files_dir""/uuid.meta"
diff --git a/test/test_files/feed.xml.ignoring b/test/test_files/feed.xml.ignoring
index 039a116..0d6ea0d 100644
--- a/test/test_files/feed.xml.ignoring
+++ b/test/test_files/feed.xml.ignoring
@@ -8,36 +8,36 @@
<updated>IGNORE</updated>
<entry>
-<title type="html">foo</title>
+<title type="html">foo <em>bar</em> <strong>baz</strong></title>
<id>urn:uuid:IGNORE</id>
<updated>IGNORE</updated>
<published>IGNORE</published>
-<link href="http://example.org/bar%20baz.html" />
+<link href="http://example.org/test.html" />
<content type="html">
-<p>bar</p>
+<p>foo</p>
</content>
</entry>
<entry>
-<title type="html">a title with some nasty characters: &amp;&lt;&gt;&quot;'</title>
+<title type="html">foo</title>
<id>urn:uuid:IGNORE</id>
<updated>IGNORE</updated>
<published>IGNORE</published>
-<link href="http://example.org/foo.html" />
+<link href="http://example.org/bar%20baz.html" />
<content type="html">
-<p>this text contains some special characters: '&quot;&gt;&lt;&amp;</p>
-<p>and more than one paragraph</p>
+<p>bar</p>
</content>
</entry>
<entry>
-<title type="html">foo <em>bar</em> <strong>baz</strong></title>
+<title type="html">a title with some nasty characters: &amp;&lt;&gt;&quot;'</title>
<id>urn:uuid:IGNORE</id>
<updated>IGNORE</updated>
<published>IGNORE</published>
-<link href="http://example.org/test.html" />
+<link href="http://example.org/foo.html" />
<content type="html">
-<p>bar</p>
+<p>this text contains some special characters: '&quot;&gt;&lt;&amp;</p>
+<p>and more than one paragraph</p>
</content>
</entry>
diff --git a/test/test_files/index.html b/test/test_files/index.html
index f5b755c..3210fef 100644
--- a/test/test_files/index.html
+++ b/test/test_files/index.html
@@ -6,8 +6,8 @@
<body>
<h1>Yet another blog</h1>
<ul>
-<li><a href="bar%20baz.html" />foo</a></li>
<li><a href="foo.html" />a title with some nasty characters: &<>"'</a></li>
+<li><a href="bar%20baz.html" />foo</a></li>
<li><a href="test.html" />foo <em>bar</em> <strong>baz</strong></a></li>
</ul>
</body>
diff --git a/test/test_files/test.html b/test/test_files/test.html
index 787bf0e..565355d 100644
--- a/test/test_files/test.html
+++ b/test/test_files/test.html
@@ -6,7 +6,7 @@
<body>
<h1>foo <em>bar</em> <strong>baz</strong></h1>
<section>
-<p>bar</p>
+<p>foo</p>
</section>
</body>
</html>
\ No newline at end of file
diff --git a/test/test_files/test.md b/test/test_files/test.md
index 3dd09f7..4416db1 100644
--- a/test/test_files/test.md
+++ b/test/test_files/test.md
@@ -1,2 +1,2 @@
% foo *bar* **baz**
-bar
+foo
--
2.30.2