index : cgit

/*
 * The plain page, which hands over a repository's own bytes rather than
 * rendering a view of them. A file is written out whole under a content type
 * guessed from its name, though a repository that has not enabled html serving
 * keeps only the types a browser will not act on. A directory, or a request
 * carrying no path, is answered with a bare document of links to the entries
 * below it rather than with one of cgit's themed pages.
 */

#define USE_THE_REPOSITORY_VARIABLE

#include "cgit.h"
#include "html.h"
#include "shared.h"
#include "ui-plain.h"
#include "ui-shared.h"

// A listing is opened by the entry that matched and closed only once the walk
// is over, so the end of the page has to tell the three cases apart.
enum response {
	RESPONSE_NONE,
	RESPONSE_BLOB,
	RESPONSE_LISTING
};

struct walk_tree_context {
	// Length of the directory part of the requested path, slash included,
	// and -1 when no path was requested so that no base length can equal
	// it.
	int dir_len;
	enum response response;
};

/*
 * Everything below text/ and application/ can carry markup or script that a
 * browser would run against the site, so only PDF is let back through.
 */
static int is_unsafe_type(const char *mimetype)
{
	return (starts_with(mimetype, "text/") ||
		starts_with(mimetype, "application/")) &&
		strcmp(mimetype, "application/pdf");
}

/*
 * Writes the response for the object, error pages included, so the walk
 * does not go on to report the path as missing.
 */
static void print_object(const struct object_id *oid, const char *path)
{
	enum object_type type;
	char *buf, *mimetype;
	size_t size;

	type = odb_read_object_info(the_repository->objects, oid, &size);
	if (type == OBJ_BAD) {
		cgit_print_error_page(404, "Not Found", "Not found");
		return;
	}

	if (cgit_blob_too_large(size)) {
		cgit_print_blob_too_large(size);
		return;
	}

	buf = odb_read_object(the_repository->objects, oid, &type, &size);
	if (!buf) {
		cgit_print_error_page(404, "Not Found", "Not found");
		return;
	}

	mimetype = cgit_get_mimetype_for_filename(path);
	ctx.page.mimetype = mimetype;

	if (!ctx.repo->enable_html_serving) {
		ctx.page.untrusted = 1;
		if (mimetype && is_unsafe_type(mimetype))
			ctx.page.mimetype = NULL;
	}

	if (!ctx.page.mimetype)
		ctx.page.mimetype = buffer_is_binary(buf, size) ? "application/octet-stream" : "text/plain";
	ctx.page.filename = path;
	ctx.page.size = size;
	cgit_print_http_headers();
	html_raw(buf, size);
	// ctx.page.mimetype may still point into mimetype, and the headers
	// are read again by whatever renders after this, so it stays alive
	// for the rest of the request.
	free(buf);
}

static char *build_path(const char *base, int baselen, const char *path)
{
	if (path[0])
		return xstrfmt("%.*s%s/", baselen, base, path);
	else
		return xstrfmt("%.*s/", baselen, base);
}

static void print_dir(const char *base, int baselen, const char *path)
{
	char *fullpath;
	const char *leading_slash;
	size_t len;

	fullpath = build_path(base, baselen, path);
	leading_slash = (fullpath[0] == '/' ? "" : "/");
	cgit_print_http_headers();
	// A full document of its own, and without the doctype and charset
	// the browser would parse it in quirks mode.
	html("<!DOCTYPE html>\n<html lang='en'>\n<head>\n");
	html("<meta charset='UTF-8'>\n");
	htmlf("<title>%s", leading_slash);
	html_txt(fullpath);
	htmlf("</title>\n</head>\n<body>\n<h2>%s", leading_slash);
	html_txt(fullpath);
	html("</h2>\n<ul>\n");
	len = strlen(fullpath);
	if (len > 1) {
		char *slash;

		// Nothing left to drop means the parent is the root, which
		// cgit_plain_link is asked for with a null path.
		fullpath[len - 1] = 0;
		slash = strrchr(fullpath, '/');
		if (slash) {
			*(slash + 1) = 0;
		} else {
			free(fullpath);
			fullpath = NULL;
		}
		html("<li>");
		cgit_plain_link("../", NULL, NULL, ctx.qry.head, ctx.qry.oid, fullpath);
		html("</li>\n");
	}
	free(fullpath);
}

static void print_dir_entry(const struct object_id *oid, const char *base,
	int baselen, const char *path, unsigned mode)
{
	char *fullpath;

	fullpath = build_path(base, baselen, path);
	if (!S_ISDIR(mode) && !S_ISGITLINK(mode))
		fullpath[strlen(fullpath) - 1] = 0;
	html("<li>");
	if (S_ISGITLINK(mode))
		cgit_submodule_link(NULL, fullpath, oid_to_hex(oid));
	else
		cgit_plain_link(path, NULL, NULL, ctx.qry.head, ctx.qry.oid, fullpath);
	html("</li>\n");
	free(fullpath);
}

static void print_dir_tail(void)
{
	html("</ul>\n</body>\n</html>\n");
}

static int walk_tree(const struct object_id *oid, struct strbuf *base,
	const char *pathname, unsigned mode, void *context)
{
	struct walk_tree_context *walk = context;

	if (walk->dir_len >= 0 && base->len == (size_t)walk->dir_len) {
		// A tree can hold two entries of one name, and only the first
		// answers: a second response would follow the first in the same
		// body.
		if (walk->response != RESPONSE_NONE)
			return 0;
		if (S_ISREG(mode) || S_ISLNK(mode)) {
			print_object(oid, pathname);
			walk->response = RESPONSE_BLOB;
		} else if (S_ISDIR(mode)) {
			print_dir(base->buf, base->len, pathname);
			walk->response = RESPONSE_LISTING;
			return READ_TREE_RECURSIVE;
		}
	} else if (base->len < INT_MAX && (int)base->len > walk->dir_len) {
		print_dir_entry(oid, base->buf, base->len, pathname, mode);
		walk->response = RESPONSE_LISTING;
	} else if (S_ISDIR(mode)) {
		return READ_TREE_RECURSIVE;
	}

	return 0;
}

static int dir_prefix_len(const char *path)
{
	const char *slash = strrchr(path, '/');

	if (slash)
		return slash - path + 1;
	return 0;
}

void cgit_print_plain(void)
{
	const char *rev = ctx.qry.oid;
	struct object_id oid;
	struct commit *commit;
	struct pathspec_item path_item;
	struct pathspec paths = { 0 };
	struct walk_tree_context walk = {
		.response = RESPONSE_NONE
	};

	if (!rev)
		rev = ctx.qry.head;

	if (repo_get_oid(the_repository, rev, &oid)) {
		cgit_print_error_page(404, "Not Found", "Not found");
		return;
	}
	commit = lookup_commit_reference(the_repository, &oid);
	if (!commit || repo_parse_commit(the_repository, commit)) {
		cgit_print_error_page(404, "Not Found", "Not found");
		return;
	}
	// The path is taken literally, since as a glob every entry a pattern
	// like * matches would be answered with its own HTTP headers inside
	// the body of the first.
	cgit_literal_pathspec(&paths, &path_item, ctx.qry.path);
	if (!ctx.qry.path) {
		// The walk is never handed an entry for the top of the tree
		// itself, so the listing it would have opened is opened here.
		walk.dir_len = -1;
		print_dir("", 0, "");
		walk.response = RESPONSE_LISTING;
	} else {
		walk.dir_len = dir_prefix_len(ctx.qry.path);
	}
	read_tree(the_repository, repo_get_commit_tree(the_repository, commit), &paths, walk_tree, &walk);
	if (walk.response == RESPONSE_NONE)
		cgit_print_error_page(404, "Not Found", "Not found");
	else if (walk.response == RESPONSE_LISTING)
		print_dir_tail();
}