Grep the source, Lukelh/grep

This is a PoC for grepping the blobs in the current tree.

Signed-off-by: Lars Hjemli <larsh@hjemli.net>

5 files changed, 223 insertions, 8 deletions

diff --git a/cmd.c b/cmd.c
index 5a3d157..9a040a2 100644
--- a/cmd.c
+++ b/cmd.c
@@ -127,7 +127,10 @@ static void tag_fn(struct cgit_context *ctx)
 
 static void tree_fn(struct cgit_context *ctx)
 {
-	cgit_print_tree(ctx->qry.sha1, ctx->qry.path);
+	if (ctx->qry.grep)
+		cgit_grep(ctx->qry.sha1, ctx->qry.path, ctx->qry.grep);
+	else
+		cgit_print_tree(ctx->qry.sha1, ctx->qry.path);
 }
 
 #define def_cmd(name, want_repo, want_layout, want_vpath, is_clone) \
diff --git a/ui-shared.c b/ui-shared.c
index 43166af..026c2cb 100644
--- a/ui-shared.c
+++ b/ui-shared.c
@@ -246,6 +246,61 @@ static char *repolink(const char *title, const char *class, const char *page,
 	return fmt("%s", delim);
 }
 
+static char *repolink2(const char *title, const char *class, const char *page,
+		       const char *head, const char *path, int lineno)
+{
+	char *delim = "?";
+
+	html("<a");
+	if (title) {
+		html(" title='");
+		html_attr(title);
+		html("'");
+	}
+	if (class) {
+		html(" class='");
+		html_attr(class);
+		html("'");
+	}
+	html(" href='");
+	if (ctx.cfg.virtual_root) {
+		html_url_path(ctx.cfg.virtual_root);
+		if (ctx.cfg.virtual_root[strlen(ctx.cfg.virtual_root) - 1] != '/')
+			html("/");
+		html_url_path(ctx.repo->url);
+		if (ctx.repo->url[strlen(ctx.repo->url) - 1] != '/')
+			html("/");
+		if (page) {
+			html_url_path(page);
+			html("/");
+			if (path)
+				html_url_path(path);
+		}
+	} else {
+		html(ctx.cfg.script_name);
+		html("?url=");
+		html_url_arg(ctx.repo->url);
+		if (ctx.repo->url[strlen(ctx.repo->url) - 1] != '/')
+			html("/");
+		if (page) {
+			html_url_arg(page);
+			html("/");
+			if (path)
+				html_url_arg(path);
+		}
+		delim = "&amp;";
+	}
+	if (head && strcmp(head, ctx.repo->defbranch)) {
+		html(delim);
+		html("h=");
+		html_url_arg(head);
+		delim = "&amp;";
+	}
+	if (lineno)
+		htmlf("#n%d", lineno);
+	return fmt("%s", delim);
+}
+
 static void reporevlink(const char *page, const char *name, const char *title,
 			const char *class, const char *head, const char *rev,
 			const char *path)
@@ -263,6 +318,23 @@ static void reporevlink(const char *page, const char *name, const char *title,
 	html("</a>");
 }
 
+static void reporevlink2(const char *page, const char *name, const char *title,
+			 const char *class, const char *head, const char *rev,
+			 const char *path, int lineno)
+{
+	char *delim;
+
+	delim = repolink2(title, class, page, head, path, lineno);
+	if (rev && ctx.qry.head != NULL && strcmp(rev, ctx.qry.head)) {
+		html(delim);
+		html("id=");
+		html_url_arg(rev);
+	}
+	html("'>");
+	html_txt(name);
+	html("</a>");
+}
+
 void cgit_summary_link(const char *name, const char *title, const char *class,
 		       const char *head)
 {
@@ -281,6 +353,13 @@ void cgit_tree_link(const char *name, const char *title, const char *class,
 	reporevlink("tree", name, title, class, head, rev, path);
 }
 
+void cgit_tree_link2(const char *name, const char *title, const char *class,
+		     const char *head, const char *rev, const char *path,
+		     int lineno)
+{
+	reporevlink2("tree", name, title, class, head, rev, path, lineno);
+}
+
 void cgit_plain_link(const char *name, const char *title, const char *class,
 		     const char *head, const char *rev, const char *path)
 {
@@ -875,6 +954,9 @@ static void print_header(struct cgit_context *ctx)
 
 void cgit_print_pageheader(struct cgit_context *ctx)
 {
+	int tree;
+	char *page;
+
 	html("<div id='cgit'>");
 	if (!ctx->cfg.noheader)
 		print_header(ctx);
@@ -903,16 +985,23 @@ void cgit_print_pageheader(struct cgit_context *ctx)
 				    NULL);
 		html("</td><td class='form'>");
 		html("<form class='right' method='get' action='");
-		if (ctx->cfg.virtual_root)
-			html_url_path(cgit_fileurl(ctx->qry.repo, "log",
+		if (ctx->cfg.virtual_root) {
+			tree = !strcmp(ctx->qry.page, "tree");
+			page = tree ? "tree" : "log";
+			html_url_path(cgit_fileurl(ctx->qry.repo, page,
 						   ctx->qry.vpath, NULL));
+		}
 		html("'>\n");
-		cgit_add_hidden_formfields(1, 0, "log");
+		cgit_add_hidden_formfields(1, 0, page);
 		html("<select name='qt'>\n");
-		html_option("grep", "log msg", ctx->qry.grep);
-		html_option("author", "author", ctx->qry.grep);
-		html_option("committer", "committer", ctx->qry.grep);
-		html_option("range", "range", ctx->qry.grep);
+		if (tree) {
+			html_option("grep", "source", ctx->qry.grep);
+		} else {
+			html_option("grep", "log msg", ctx->qry.grep);
+			html_option("author", "author", ctx->qry.grep);
+			html_option("committer", "committer", ctx->qry.grep);
+			html_option("range", "range", ctx->qry.grep);
+		}
 		html("</select>\n");
 		html("<input class='txt' type='text' size='10' name='q' value='");
 		html_attr(ctx->qry.search);
diff --git a/ui-shared.h b/ui-shared.h
index 87a7dac..35dca9c 100644
--- a/ui-shared.h
+++ b/ui-shared.h
@@ -20,6 +20,9 @@ extern void cgit_tag_link(const char *name, const char *title,
 extern void cgit_tree_link(const char *name, const char *title,
 			   const char *class, const char *head,
 			   const char *rev, const char *path);
+extern void cgit_tree_link2(const char *name, const char *title,
+			   const char *class, const char *head,
+			   const char *rev, const char *path, int lno);
 extern void cgit_plain_link(const char *name, const char *title,
 			    const char *class, const char *head,
 			    const char *rev, const char *path);
diff --git a/ui-tree.c b/ui-tree.c
index b1adcc7..249e99e 100644
--- a/ui-tree.c
+++ b/ui-tree.c
@@ -277,3 +277,122 @@ void cgit_print_tree(const char *rev, char *path)
 	read_tree_recursive(commit->tree, "", 0, 0, paths, walk_tree, NULL);
 	ls_tail();
 }
+
+struct grep_data {
+	regex_t re;
+	char *path;
+	int max_hits;
+};
+
+static int grep_blob(const unsigned char *sha1, const char *base, int baselen,
+		      const char *pathname, struct grep_data *grep)
+{
+	enum object_type type;
+	char *buf, *line, *lf, *fullpath;
+	unsigned long size;
+	int hits, lno, rc;
+
+	type = sha1_object_info(sha1, &size);
+	if (type == OBJ_BAD) {
+		cgit_print_error(fmt("Bad object name: %s",
+				     sha1_to_hex(sha1)));
+		return -1;
+	}
+
+	buf = read_sha1_file(sha1, &type, &size);
+	if (!buf) {
+		cgit_print_error(fmt("Error reading object %s",
+				     sha1_to_hex(sha1)));
+		return -1;
+	}
+	hits = 0;
+	line = buf;
+	lno = 1;
+	while (size) {
+		lf = strchr(line, '\n');
+		if (lf)
+			*lf = 0;
+		rc = regexec(&grep->re, line, 0, NULL, 0);
+		if (!rc) {
+			if (!hits) {
+				fullpath = fmt("%.*s%s", baselen, base,
+					       pathname);
+				htmlf("<strong>%s</strong><pre>", fullpath);
+			}
+			htmlf("%6d: ", lno);
+			while (isspace(*line))
+				line++;
+			cgit_tree_link2(line, NULL, NULL, ctx.qry.head,
+					ctx.qry.sha1, fullpath, lno);
+			html("\n");
+			++hits;
+			if (!(--grep->max_hits))
+				break;
+		}
+		if (!lf)
+			break;
+		size -= (lf - line);
+		*lf = '\n';
+		line = lf + 1;
+		lno++;
+	}
+	if (hits)
+		html("</pre>");
+	if (!grep->max_hits) {
+		cgit_print_error("Too many matches found.");
+		return -1;
+	}
+	return 0;
+}
+
+static int grep_tree(const unsigned char *sha1, const char *base, int baselen,
+		     const char *pathname, unsigned mode, int stage,
+		     void *cbdata)
+{
+	int rc;
+	char *fullpath;
+	struct grep_data *grep;
+
+	grep = cbdata;
+	fullpath = fmt("%.*s%s", baselen, base, pathname);
+	if (S_ISDIR(mode)) {
+		rc = grep->path ? prefixcmp(grep->path, fullpath) : 0;
+		if (!rc)
+			return READ_TREE_RECURSIVE;
+		return (rc > 0 ? -1 : 0);
+	} else if (S_ISREG(mode)) {
+		rc = grep->path ? prefixcmp(fullpath, grep->path) : 0;
+		if (!rc)
+			return grep_blob(sha1, base, baselen, pathname, grep);
+		return (rc < 0 ? -1 : 0);
+	}
+	return 0;
+}
+
+void cgit_grep(const char *rev, char *path, char *grep)
+{
+	unsigned char sha1[20];
+	struct commit *commit;
+	struct grep_data data;
+
+	if (!rev)
+		rev = ctx.qry.head;
+
+	curr_rev = xstrdup(rev);
+	if (get_sha1(rev, sha1)) {
+		cgit_print_error(fmt("Invalid revision name: %s", rev));
+		return;
+	}
+	commit = lookup_commit_reference(sha1);
+	if (!commit || parse_commit(commit)) {
+		cgit_print_error(fmt("Invalid commit reference: %s", rev));
+		return;
+	}
+	data.path = path;
+	data.max_hits = 100;
+	if (regcomp(&data.re, ctx.qry.search, REG_EXTENDED | REG_ICASE | REG_NOSUB)) {
+		cgit_print_error("Invalid regex");
+		return;
+	}
+	read_tree_recursive(commit->tree, "", 0, 0, NULL, grep_tree, &data);
+}
diff --git a/ui-tree.h b/ui-tree.h
index bbd34e3..2e0851f 100644
--- a/ui-tree.h
+++ b/ui-tree.h
@@ -2,5 +2,6 @@
 #define UI_TREE_H
 
 extern void cgit_print_tree(const char *rev, char *path);
+extern void cgit_grep(const char *rev, char *path, char *grep);
 
 #endif /* UI_TREE_H */