26 files changed, 435 insertions, 68 deletions

diff --git a/Makefile b/Makefile
index 27e966d..72c189c 100644
--- a/Makefile
+++ b/Makefile
@@ -19,7 +19,12 @@ OBJECTS = shared.o cache.o parsing.o html.o ui-shared.o ui-repolist.o \
         ui-snapshot.o ui-blob.o ui-tag.o ui-refs.o
 
 
-.PHONY: all git install clean distclean emptycache force-version get-git
+ifdef NEEDS_LIBICONV
+        EXTLIBS += -liconv
+endif
+
+
+.PHONY: all git test install clean distclean emptycache force-version get-git
 
 all: cgit git
 
@@ -49,6 +54,9 @@ git:
         cd git && $(MAKE) xdiff/lib.a
         cd git && $(MAKE) libgit.a
 
+test: all
+        $(MAKE) -C tests
+
 install: all
         mkdir -p $(DESTDIR)$(CGIT_SCRIPT_PATH)
         install cgit $(DESTDIR)$(CGIT_SCRIPT_PATH)/$(CGIT_SCRIPT_NAME)
diff --git a/cgit.css b/cgit.css
index 1b2e9d6..17c2712 100644
--- a/cgit.css
+++ b/cgit.css
@@ -8,7 +8,7 @@ body {
         font-size: 10pt;
         color: #333;
         background: white;
-        padding-left: 4px;
+        padding: 4px;
 }
 
 table {
@@ -78,29 +78,39 @@ img {
         border: none;
 }
 
-div#sidebar {
+table#layout {
+        border-collapse: collapse;
+        border: none;
+        margin: 0px;
+}
+
+td#sidebar {
         vertical-align: top;
         width: 162px;
         padding: 0px 0px 0px 0px;
-        margin: 4px;
-        float: left;
+        margin: 0px;
 }
 
-div#logo {
+td#sidebar table {
+        border-collapse: separate;
+        border-spacing: 0px;
         margin: 0px;
-        padding: 4px 0px 4px 0px;
-        text-align: center;
+        padding: 0px;
         background-color: #ccc;
+}
+
+td#sidebar table.sidebar td.sidebar {
+        padding: 4px;
         border-top: solid 1px #eee;
         border-left: solid 1px #eee;
         border-right: solid 1px #aaa;
         border-bottom: solid 1px #aaa;
 }
 
-div#sidebar div.infobox {
-        margin: 0px 0px 0px 0px;
-        padding: 0.5em;
-        text-align: left;
+div#logo {
+        margin: 0px;
+        padding: 4px 0px 4px 0px;
+        text-align: center;
         background-color: #ccc;
         border-top: solid 1px #eee;
         border-left: solid 1px #eee;
@@ -108,50 +118,44 @@ div#sidebar div.infobox {
         border-bottom: solid 1px #aaa;
 }
 
-div#sidebar div.infobox h1 {
-        font-size: 11pt;
+td#sidebar h1 {
+        font-size: 10pt;
         font-weight: bold;
-        margin: 0px;
+        margin: 8px 0px 0px 0px;
 }
 
-div#sidebar div.infobox a.menu {
+td#sidebar h1.first {
+        margin-top: 0px;
+}
+
+td#sidebar a.menu {
         display: block;
         background-color: #ccc;
         padding: 0.1em 0.5em;
         text-decoration: none;
 }
 
-div#sidebar div.infobox a.menu:hover {
+td#sidebar a.menu:hover {
         background-color: #bbb;
         text-decoration: none;
 }
 
-div#sidebar div.infobox select {
+td#sidebar select {
         width: 100%;
-        border: solid 1px #aaa;
-        background-color: #bbb;
         margin: 2px 0px 0px 0px;
-        padding: 0px;
 }
 
-td#branch-dropdown-cell {
-        width: 99%;
+td#sidebar form {
+        text-align: right;
 }
 
 input#switch-btn {
-        width: 20px;
-        border: solid 1px #aaa;
-        background-color: #bbb;
         margin: 2px 0px 0px 0px;
-        padding: 0px;
 }
 
-div#sidebar div.infobox input.txt {
+td#sidebar input.txt {
         width: 100%;
-        border: solid 1px #aaa;
-        background-color: #bbb;
         margin: 2px 0px 0px 0px;
-        padding: 0;
 }
 
 table#grid {
diff --git a/cgit.h b/cgit.h
index 163f355..ab0efeb 100644
--- a/cgit.h
+++ b/cgit.h
@@ -16,6 +16,7 @@
 #include <log-tree.h>
 #include <archive.h>
 #include <xdiff/xdiff.h>
+#include <utf8.h>
 
 
 /*
@@ -48,6 +49,11 @@
 #define TM_MONTH (TM_YEAR / 12.0)
 
 
+/*
+ * Default encoding
+ */
+#define PAGE_ENCODING "UTF-8"
+
 typedef void (*configfn)(const char *name, const char *value);
 typedef void (*filepair_fn)(struct diff_filepair *pair);
 typedef void (*linediff_fn)(char *line, int len);
@@ -90,6 +96,7 @@ struct commitinfo {
         unsigned long committer_date;
         char *subject;
         char *msg;
+        char *msg_encoding;
 };
 
 struct taginfo {
@@ -132,6 +139,7 @@ extern char *cgit_virtual_root;
 extern char *cgit_script_name;
 extern char *cgit_cache_root;
 extern char *cgit_repo_group;
+extern char *cgit_robots;
 
 extern int cgit_nocache;
 extern int cgit_snapshots;
diff --git a/cgit.png b/cgit.png
index ee48197..22f7e95 100644
--- a/cgit.png
+++ b/cgit.png
diff --git a/cgitrc b/cgitrc
index 6363c9c..8c616e0 100644
--- a/cgitrc
+++ b/cgitrc
@@ -8,6 +8,11 @@
 #nocache=0
 
 
+## This variable can be used to override the default value for "robots"
+## meta-tag. If unset, the meta-tag isn't generated.
+#robots=index, nofollow
+
+
 ## Set allowed snapshot types by default. Can be overridden per repo
 # can be any combination of zip/tar.gz/tar.bz2/tar
 #snapshots=0
diff --git a/html.c b/html.c
index 6c9cc8b..eb163d9 100644
--- a/html.c
+++ b/html.c
@@ -132,7 +132,7 @@ void html_option(char *value, char *text, char *selected_value)
         html_attr(value);
         html("'");
         if (selected_value && !strcmp(selected_value, value))
-                html(" selected");
+                html(" selected='selected'");
         html(">");
         html_txt(text);
         html("</option>\n");
diff --git a/parsing.c b/parsing.c
index 55a485d..5093b8b 100644
--- a/parsing.c
+++ b/parsing.c
@@ -199,6 +199,7 @@ struct commitinfo *cgit_parse_commit(struct commit *commit)
         ret->committer_email = NULL;
         ret->subject = NULL;
         ret->msg = NULL;
+        ret->msg_encoding = NULL;
 
         if (p == NULL)
                 return ret;
@@ -233,6 +234,14 @@ struct commitinfo *cgit_parse_commit(struct commit *commit)
                 p = strchr(t, '\n') + 1;
         }
 
+        if (!strncmp(p, "encoding ", 9)) {
+                p += 9;
+                t = strchr(p, '\n') + 1;
+                ret->msg_encoding = substr(p, t);
+                p = t;
+        } else
+                ret->msg_encoding = xstrdup(PAGE_ENCODING);
+
         while (*p && (*p != '\n'))
                 p = strchr(p, '\n') + 1; // skip unknown header fields
 
@@ -253,6 +262,22 @@ struct commitinfo *cgit_parse_commit(struct commit *commit)
         } else
                 ret->subject = substr(p, p+strlen(p));
 
+        if(strcmp(ret->msg_encoding, PAGE_ENCODING)) {
+                t = reencode_string(ret->subject, PAGE_ENCODING,
+                                    ret->msg_encoding);
+                if(t) {
+                        free(ret->subject);
+                        ret->subject = t;
+                }
+
+                t = reencode_string(ret->msg, PAGE_ENCODING,
+                                    ret->msg_encoding);
+                if(t) {
+                        free(ret->msg);
+                        ret->msg = t;
+                }
+        }
+
         return ret;
 }
 
diff --git a/shared.c b/shared.c
index 6117f5c..fd8b1e3 100644
--- a/shared.c
+++ b/shared.c
@@ -26,6 +26,7 @@ char *cgit_virtual_root = NULL;
 char *cgit_script_name  = CGIT_SCRIPT_NAME;
 char *cgit_cache_root   = CGIT_CACHE_ROOT;
 char *cgit_repo_group   = NULL;
+char *cgit_robots       = "index, nofollow";
 
 int cgit_nocache               =  0;
 int cgit_snapshots             =  0;
@@ -197,6 +198,8 @@ void cgit_global_config_cb(const char *name, const char *value)
                 cgit_agefile = xstrdup(value);
         else if (!strcmp(name, "renamelimit"))
                 cgit_renamelimit = atoi(value);
+        else if (!strcmp(name, "robots"))
+                cgit_robots = xstrdup(value);
         else if (!strcmp(name, "repo.group"))
                 cgit_repo_group = xstrdup(value);
         else if (!strcmp(name, "repo.url"))
@@ -267,6 +270,8 @@ void *cgit_free_commitinfo(struct commitinfo *info)
         free(info->committer);
         free(info->committer_email);
         free(info->subject);
+        free(info->msg);
+        free(info->msg_encoding);
         free(info);
         return NULL;
 }
diff --git a/tests/.gitignore b/tests/.gitignore
new file mode 100644
index 0000000..c1c1c0b
--- /dev/null
+++ b/tests/.gitignore
@@ -0,0 +1,2 @@
+trash
+test-output.log
diff --git a/tests/Makefile b/tests/Makefile
new file mode 100644
index 0000000..697e5a1
--- /dev/null
+++ b/tests/Makefile
@@ -0,0 +1,13 @@
+
+
+T = $(wildcard t[0-9][0-9][0-9][0-9]-*.sh)
+
+all: $(T)
+
+$(T):
+        @$@
+
+clean:
+        $(RM) -rf trash
+
+.PHONY: $(T) clean
diff --git a/tests/setup.sh b/tests/setup.sh
new file mode 100755
index 0000000..51d5a75
--- /dev/null
+++ b/tests/setup.sh
@@ -0,0 +1,108 @@
+# This file should be sourced by all test-scripts
+#
+# Main functions:
+#   prepare_tests(description) - setup for testing, i.e. create repos+config
+#   run_test(description, script) - run one test, i.e. eval script
+#
+# Helper functions
+#   cgit_query(querystring) - call cgit with the specified querystring
+#   cgit_url(url) - call cgit with the specified virtual url
+#
+# Example script:
+#
+# . setup.sh
+# prepare_tests "html validation"
+# run_test 'repo index' 'cgit_url "/" | tidy -e'
+# run_test 'repo summary' 'cgit_url "/foo" | tidy -e'
+
+
+mkrepo() {
+        name=$1
+        count=$2
+        dir=$PWD
+        test -d $name && return
+        printf "Creating testrepo %s\n" $name
+        mkdir -p $name
+        cd $name
+        git init
+        for ((n=1; n<=count; n++))
+        do
+                echo $n >file-$n
+                git add file-$n
+                git commit -m "commit $n"
+        done
+        cd $dir
+}
+
+setup_repos()
+{
+        rm -rf trash/cache
+        mkdir -p trash/cache
+        mkrepo trash/repos/foo 5 >/dev/null
+        mkrepo trash/repos/bar 50 >/dev/null
+        cat >trash/cgitrc <<EOF
+virtual-root=/
+cache-root=$PWD/trash/cache
+
+nocache=0
+snapshots=tar.gz tar.bz zip
+enable-log-filecount=1
+enable-log-linecount=1
+summary-log=5
+summary-branches=5
+summary-tags=5
+
+repo.url=foo
+repo.path=$PWD/trash/repos/foo/.git
+repo.desc=the foo repo
+
+repo.url=bar
+repo.path=$PWD/trash/repos/bar/.git
+repo.desc=the bar repo
+EOF
+}
+
+prepare_tests()
+{
+        setup_repos
+        test_count=0
+        test_failed=0
+        echo "$@" "($0)"
+}
+
+tests_done()
+{
+        printf "\n"
+        if test $test_failed -gt 0
+        then
+                printf "[%s of %s tests failed]\n" $test_failed $test_count
+                false
+        fi
+}
+
+run_test()
+{
+        desc=$1
+        script=$2
+        ((test_count++))
+        eval "$2" >test-output.log
+        res=$?
+        if test $res = 0
+        then
+                printf "  %s: ok - %s\n" $test_count "$desc"
+        else
+                ((test_failed++))
+                printf "  %s: fail - %s\n" $test_count "$desc"
+        fi
+}
+
+cgit_query()
+{
+        CGIT_CONFIG="$PWD/trash/cgitrc" QUERY_STRING="$1" "$PWD/../cgit"
+}
+
+cgit_url()
+{
+        CGIT_CONFIG="$PWD/trash/cgitrc" QUERY_STRING="url=$1" "$PWD/../cgit"
+}
+
diff --git a/tests/t0010-validate-html.sh b/tests/t0010-validate-html.sh
new file mode 100755
index 0000000..907a415
--- /dev/null
+++ b/tests/t0010-validate-html.sh
@@ -0,0 +1,31 @@
+#!/bin/sh
+
+. ./setup.sh
+
+
+test_url()
+{
+        tidy_opt="-eq"
+        test -z "$NO_TIDY_WARNINGS" || tidy_opt+=" --show-warnings no"
+        cgit_url "$1" | sed -e "1,4d" >trash/tidy-$test_count
+        tidy $tidy_opt trash/tidy-$test_count
+        rc=$?
+        if test $rc = 2
+        then
+                false
+        else
+                :
+        fi
+}
+
+prepare_tests 'Validate html with tidy'
+
+run_test 'index page' 'test_url ""'
+run_test 'foo' 'test_url "foo"'
+run_test 'foo/log' 'test_url "foo/log"'
+run_test 'foo/tree' 'test_url "foo/tree"'
+run_test 'foo/tree/file-1' 'test_url "foo/tree/file-1"'
+run_test 'foo/commit' 'test_url "foo/commit"'
+run_test 'foo/diff' 'test_url "foo/diff"'
+
+tests_done
diff --git a/tests/t0101-index.sh b/tests/t0101-index.sh
new file mode 100755
index 0000000..12ed00c
--- /dev/null
+++ b/tests/t0101-index.sh
@@ -0,0 +1,13 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on index page"
+
+run_test 'generate index page' 'cgit_url "" >trash/tmp'
+run_test 'find foo repo' 'grep -e "foo" trash/tmp'
+run_test 'find bar repo' 'grep -e "bar" trash/tmp'
+run_test 'no tree-link' 'grep -ve "foo/tree" trash/tmp'
+run_test 'no log-link' 'grep -ve "foo/log" trash/tmp'
+
+tests_done
diff --git a/tests/t0102-summary.sh b/tests/t0102-summary.sh
new file mode 100755
index 0000000..7edd675
--- /dev/null
+++ b/tests/t0102-summary.sh
@@ -0,0 +1,20 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on summary page"
+
+run_test 'generate foo summary' 'cgit_url "foo" >trash/tmp'
+run_test 'find commit 1' 'grep -e "commit 1" trash/tmp'
+run_test 'find commit 5' 'grep -e "commit 5" trash/tmp'
+run_test 'find branch master' 'grep -e "master" trash/tmp'
+run_test 'no tags' 'grep -ve "tags" trash/tmp'
+
+run_test 'generate bar summary' 'cgit_url "bar" >trash/tmp'
+run_test 'no commit 45' 'grep -ve "commit 45" trash/tmp'
+run_test 'find commit 46' 'grep -e "commit 46" trash/tmp'
+run_test 'find commit 50' 'grep -e "commit 50" trash/tmp'
+run_test 'find branch master' 'grep -e "master" trash/tmp'
+run_test 'no tags' 'grep -ve "tags" trash/tmp'
+
+tests_done
diff --git a/tests/t0103-log.sh b/tests/t0103-log.sh
new file mode 100755
index 0000000..b08cd29
--- /dev/null
+++ b/tests/t0103-log.sh
@@ -0,0 +1,15 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on log page"
+
+run_test 'generate foo/log' 'cgit_url "foo/log" >trash/tmp'
+run_test 'find commit 1' 'grep -e "commit 1" trash/tmp'
+run_test 'find commit 5' 'grep -e "commit 5" trash/tmp'
+
+run_test 'generate bar/log' 'cgit_url "bar/log" >trash/tmp'
+run_test 'find commit 1' 'grep -e "commit 1" trash/tmp'
+run_test 'find commit 50' 'grep -e "commit 50" trash/tmp'
+
+tests_done
diff --git a/tests/t0104-tree.sh b/tests/t0104-tree.sh
new file mode 100755
index 0000000..2516c72
--- /dev/null
+++ b/tests/t0104-tree.sh
@@ -0,0 +1,21 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on tree page"
+
+run_test 'generate bar/tree' 'cgit_url "bar/tree" >trash/tmp'
+run_test 'find file-1' 'grep -e "file-1" trash/tmp'
+run_test 'find file-50' 'grep -e "file-50" trash/tmp'
+
+run_test 'generate bar/tree/file-50' 'cgit_url "bar/tree/file-50" >trash/tmp'
+
+run_test 'find line 1' '
+        grep -e "<a id=.n1. name=.n1. href=.#n1.>1</a>" trash/tmp
+'
+
+run_test 'no line 2' '
+        grep -e "<a id=.n2. name=.n2. href=.#n2.>2</a>" trash/tmp
+'
+
+tests_done
diff --git a/tests/t0105-commit.sh b/tests/t0105-commit.sh
new file mode 100755
index 0000000..aa2bf33
--- /dev/null
+++ b/tests/t0105-commit.sh
@@ -0,0 +1,22 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on commit page"
+
+run_test 'generate foo/commit' 'cgit_url "foo/commit" >trash/tmp'
+run_test 'find tree link' 'grep -e "<a href=./foo/tree/.>" trash/tmp'
+run_test 'find parent link' 'grep -E "<a href=./foo/commit/\?id=.+>" trash/tmp'
+
+run_test 'find commit subject' '
+        grep -e "<div class=.commit-subject.>commit 5</div>" trash/tmp
+'
+
+run_test 'find commit msg' 'grep -e "<div class=.commit-msg.></div>" trash/tmp'
+run_test 'find diffstat' 'grep -e "<table summary=.diffstat. class=.diffstat.>" trash/tmp'
+
+run_test 'find diff summary' '
+         grep -e "1 files changed, 1 insertions, 0 deletions" trash/tmp
+'
+
+tests_done
diff --git a/tests/t0106-diff.sh b/tests/t0106-diff.sh
new file mode 100755
index 0000000..e140bcc
--- /dev/null
+++ b/tests/t0106-diff.sh
@@ -0,0 +1,20 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Check content on diff page"
+
+run_test 'generate foo/diff' 'cgit_url "foo/diff" >trash/tmp'
+run_test 'find diff header' 'grep -e "a/file-5 b/file-5" trash/tmp'
+run_test 'find blob link' 'grep -e "<a href=./foo/tree/file-5?id=" trash/tmp'
+run_test 'find added file' 'grep -e "new file mode 100644" trash/tmp'
+
+run_test 'find hunk header' '
+        grep -e "<div class=.hunk.>@@ -0,0 +1 @@</div>" trash/tmp
+'
+
+run_test 'find added line' '
+        grep -e "<div class=.add.>+5</div>" trash/tmp
+'
+
+tests_done
diff --git a/tests/t0107-snapshot.sh b/tests/t0107-snapshot.sh
new file mode 100755
index 0000000..8e90e10
--- /dev/null
+++ b/tests/t0107-snapshot.sh
@@ -0,0 +1,36 @@
+#!/bin/sh
+
+. ./setup.sh
+
+prepare_tests "Verify snapshot"
+
+run_test 'get foo/snapshot/test.tar.gz' '
+        cgit_url "foo/snapshot/test.tar.gz" >trash/tmp
+'
+
+run_test 'check html headers' '
+        head -n 1 trash/tmp |
+             grep -e "Content-Type: application/x-tar" &&
+
+        head -n 2 trash/tmp |
+             grep -e "Content-Disposition: inline; filename=.test.tar.gz."
+'
+
+run_test 'strip off the header lines' '
+         tail -n +6 trash/tmp > trash/test.tar.gz
+'
+
+run_test 'verify gzip format' 'gunzip --test trash/test.tar.gz'
+run_test 'untar' 'tar -xf trash/test.tar.gz -C trash'
+
+run_test 'count files' '
+        c=$(ls -1 trash/foo/ | wc -l) &&
+        test $c = 5
+'
+
+run_test 'verify untarred file-5' '
+         grep -e "^5$" trash/foo/file-5 &&
+         test $(cat trash/foo/file-5 | wc -l) = 1
+'
+
+tests_done
diff --git a/ui-commit.c b/ui-commit.c
index 4ac8955..bd55a33 100644
--- a/ui-commit.c
+++ b/ui-commit.c
@@ -84,7 +84,7 @@ void print_fileinfo(struct fileinfo *info)
         html("</td><td class='right'>");
         htmlf("%d", info->added + info->removed);
         html("</td><td class='graph'>");
-        htmlf("<table width='%d%%'><tr>", (max_changes > 100 ? 100 : max_changes));
+        htmlf("<table summary='file diffstat' width='%d%%'><tr>", (max_changes > 100 ? 100 : max_changes));
         htmlf("<td class='add' style='width: %.1f%%;'/>",
               info->added * 100.0 / max_changes);
         htmlf("<td class='rem' style='width: %.1f%%;'/>",
@@ -157,7 +157,7 @@ void cgit_print_commit(char *hex)
         }
         info = cgit_parse_commit(commit);
 
-        html("<table class='commit-info'>\n");
+        html("<table summary='commit info' class='commit-info'>\n");
         html("<tr><th>author</th><td>");
         html_txt(info->author);
         html(" ");
@@ -209,7 +209,7 @@ void cgit_print_commit(char *hex)
         html("</div>");
         if (!(commit->parents && commit->parents->next && commit->parents->next->next)) {
                 html("<div class='diffstat-header'>Diffstat</div>");
-                html("<table class='diffstat'>");
+                html("<table summary='diffstat' class='diffstat'>");
                 max_changes = 0;
                 cgit_diff_commit(commit, inspect_filepair);
                 for(i = 0; i<files; i++)
diff --git a/ui-diff.c b/ui-diff.c
index ac9a3fa..4fcf852 100644
--- a/ui-diff.c
+++ b/ui-diff.c
@@ -141,7 +141,7 @@ void cgit_print_diff(const char *new_rev, const char *old_rev, const char *prefi
                 if (!commit2 || parse_commit(commit2))
                         cgit_print_error(fmt("Bad commit: %s", sha1_to_hex(old_rev_sha1)));
         }
-        html("<table class='diff'>");
+        html("<table summary='diff' class='diff'>");
         html("<tr><td>");
         cgit_diff_tree(old_rev_sha1, new_rev_sha1, filepair_cb, prefix);
         html("</td></tr>");
diff --git a/ui-log.c b/ui-log.c
index 9f5fdf6..a41d2b2 100644
--- a/