Add support for --scan-tree=<path> option to cgit

This option makes cgit scan a directory tree looking for git repositories,
generating suitable definitions for a cgitrc file on stdout.

Signed-off-by: Lars Hjemli <hjemli@gmail.com>

4 files changed, 184 insertions, 3 deletions

diff --git a/Makefile b/Makefile
index a1caaf4..302ff4c 100644
--- a/Makefile
+++ b/Makefile
@@ -54,6 +54,7 @@ OBJECTS += cmd.o
 OBJECTS += configfile.o
 OBJECTS += html.o
 OBJECTS += parsing.o
+OBJECTS += scan-tree.o
 OBJECTS += shared.o
 OBJECTS += ui-atom.o
 OBJECTS += ui-blob.o
diff --git a/cgit.c b/cgit.c
index a47cad0..5a93fcd 100644
--- a/cgit.c
+++ b/cgit.c
@@ -12,6 +12,7 @@
 #include "configfile.h"
 #include "html.h"
 #include "ui-shared.h"
+#include "scan-tree.h"
 
 const char *cgit_version = CGIT_VERSION;
 
@@ -320,9 +321,39 @@ static void process_request(void *cbdata)
 		cgit_print_docend();
 }
 
+int cmp_repos(const void *a, const void *b)
+{
+	const struct cgit_repo *ra = a, *rb = b;
+	return strcmp(ra->url, rb->url);
+}
+
+void print_repo(struct cgit_repo *repo)
+{
+	printf("repo.url=%s\n", repo->url);
+	printf("repo.name=%s\n", repo->name);
+	printf("repo.path=%s\n", repo->path);
+	if (repo->owner)
+		printf("repo.owner=%s\n", repo->owner);
+	if (repo->desc)
+		printf("repo.desc=%s\n", repo->desc);
+	if (repo->readme)
+		printf("repo.readme=%s\n", repo->readme);
+	printf("\n");
+}
+
+void print_repolist(struct cgit_repolist *list)
+{
+	int i;
+
+	for(i = 0; i < list->count; i++)
+		print_repo(&list->repos[i]);
+}
+
+
 static void cgit_parse_args(int argc, const char **argv)
 {
 	int i;
+	int scan = 0;
 
 	for (i = 1; i < argc; i++) {
 		if (!strncmp(argv[i], "--cache=", 8)) {
@@ -351,6 +382,16 @@ static void cgit_parse_args(int argc, const char **argv)
 		if (!strncmp(argv[i], "--ofs=", 6)) {
 			ctx.qry.ofs = atoi(argv[i]+6);
 		}
+		if (!strncmp(argv[i], "--scan-tree=", 12)) {
+			scan++;
+			scan_tree(argv[i] + 12);
+		}
+	}
+	if (scan) {
+		qsort(cgit_repolist.repos, cgit_repolist.count,
+			sizeof(struct cgit_repo), cmp_repos);
+		print_repolist(&cgit_repolist);
+		exit(0);
 	}
 }
 
@@ -383,14 +424,14 @@ int main(int argc, const char **argv)
 	cgit_repolist.count = 0;
 	cgit_repolist.repos = NULL;
 
-	parse_configfile(cgit_config_env ? cgit_config_env : CGIT_CONFIG,
-			 config_cb);
-	ctx.repo = NULL;
 	if (getenv("SCRIPT_NAME"))
 		ctx.cfg.script_name = xstrdup(getenv("SCRIPT_NAME"));
 	if (getenv("QUERY_STRING"))
 		ctx.qry.raw = xstrdup(getenv("QUERY_STRING"));
 	cgit_parse_args(argc, argv);
+	parse_configfile(cgit_config_env ? cgit_config_env : CGIT_CONFIG,
+			 config_cb);
+	ctx.repo = NULL;
 	http_parse_querystring(ctx.qry.raw, querystring_cb);
 
 	/* If virtual-root isn't specified in cgitrc and no url
diff --git a/scan-tree.c b/scan-tree.c
new file mode 100644
index 0000000..cdafb02
--- /dev/null
+++ b/scan-tree.c
@@ -0,0 +1,136 @@
+#include "cgit.h"
+#include "html.h"
+
+#define MAX_PATH 4096
+
+/* return 1 if path contains a objects/ directory and a HEAD file */
+static int is_git_dir(const char *path)
+{
+	struct stat st;
+	static char buf[MAX_PATH];
+
+	if (snprintf(buf, MAX_PATH, "%s/objects", path) >= MAX_PATH) {
+		fprintf(stderr, "Insanely long path: %s\n", path);
+		return 0;
+	}
+	if (stat(buf, &st)) {
+		if (errno != ENOENT)
+			fprintf(stderr, "Error checking path %s: %s (%d)\n",
+				path, strerror(errno), errno);
+		return 0;
+	}
+	if (!S_ISDIR(st.st_mode))
+		return 0;
+
+	sprintf(buf, "%s/HEAD", path);
+	if (stat(buf, &st)) {
+		if (errno != ENOENT)
+			fprintf(stderr, "Error checking path %s: %s (%d)\n",
+				path, strerror(errno), errno);
+		return 0;
+	}
+	if (!S_ISREG(st.st_mode))
+		return 0;
+
+	return 1;
+}
+
+char *readfile(const char *path)
+{
+	FILE *f;
+	static char buf[MAX_PATH];
+
+	if (!(f = fopen(path, "r")))
+		return NULL;
+	fgets(buf, MAX_PATH, f);
+	fclose(f);
+	return buf;
+}
+
+static void add_repo(const char *base, const char *path)
+{
+	struct cgit_repo *repo;
+	struct stat st;
+	struct passwd *pwd;
+	char *p;
+
+	if (stat(path, &st)) {
+		fprintf(stderr, "Error accessing %s: %s (%d)\n",
+			path, strerror(errno), errno);
+		return;
+	}
+	if ((pwd = getpwuid(st.st_uid)) == NULL) {
+		fprintf(stderr, "Error reading owner-info for %s: %s (%d)\n",
+			path, strerror(errno), errno);
+		return;
+	}
+	if (base == path)
+		p = fmt("%s", path);
+	else
+		p = fmt("%s", path + strlen(base) + 1);
+
+	if (!strcmp(p + strlen(p) - 5, "/.git"))
+		p[strlen(p) - 5] = '\0';
+
+	repo = cgit_add_repo(xstrdup(p));
+	repo->name = repo->url;
+	repo->path = xstrdup(path);
+	repo->owner = (pwd ? xstrdup(pwd->pw_gecos ? pwd->pw_gecos : pwd->pw_name) : "");
+
+	p = fmt("%s/description", path);
+	if (!stat(p, &st))
+		repo->desc = xstrdup(readfile(p));
+
+	p = fmt("%s/README.html", path);
+	if (!stat(p, &st))
+		repo->readme = "README.html";
+}
+
+static void scan_path(const char *base, const char *path)
+{
+	DIR *dir;
+	struct dirent *ent;
+	char *buf;
+	struct stat st;
+
+	if (is_git_dir(path)) {
+		add_repo(base, path);
+		return;
+	}
+	dir = opendir(path);
+	if (!dir) {
+		fprintf(stderr, "Error opening directory %s: %s (%d)\n",
+			path, strerror(errno), errno);
+		return;
+	}
+	while((ent = readdir(dir)) != NULL) {
+		if (ent->d_name[0] == '.') {
+			if (ent->d_name[1] == '\0')
+				continue;
+			if (ent->d_name[1] == '.' && ent->d_name[2] == '\0')
+				continue;
+		}
+		buf = malloc(strlen(path) + strlen(ent->d_name) + 2);
+		if (!buf) {
+			fprintf(stderr, "Alloc error on %s: %s (%d)\n",
+				path, strerror(errno), errno);
+			exit(1);
+		}
+		sprintf(buf, "%s/%s", path, ent->d_name);
+		if (stat(buf, &st)) {
+			fprintf(stderr, "Error checking path %s: %s (%d)\n",
+				buf, strerror(errno), errno);
+			free(buf);
+			continue;
+		}
+		if (S_ISDIR(st.st_mode))
+			scan_path(base, buf);
+		free(buf);
+	}
+	closedir(dir);
+}
+
+void scan_tree(const char *path)
+{
+	scan_path(path, path);
+}
diff --git a/scan-tree.h b/scan-tree.h
new file mode 100644
index 0000000..b103b16
--- /dev/null
+++ b/scan-tree.h
@@ -0,0 +1,3 @@
+
+
+extern void scan_tree(const char *path);