sgw

commit 8684421e547f08888751ed14042b8063eadc1e45
parent 546c6d82185ea286e7365c7abe137be2ef23045d
Author: Emma Weaver <emma@waeaves.com>
Date:   Tue, 19 May 2026 15:42:26 -0400

Implemented nesting folders & line/byte counts

Diffstat:
M
TODO
20
+++++++++++++-------
M
stagit.c
861
++++++++++++++++++++++++++++++++++++++++++++++++++++---------------------------

2 files changed, 576 insertions(+), 305 deletions(-)
diff --git a/TODO b/TODO
@@ -1,26 +1,32 @@
 -STUFF I WANT TO CHANGE ABOUT STAGIT-
 (might rename it to "statigit" also.)
 
+[ ] Separate out stagit.c into clean, standalone modules.
+    - but how to handle includes/headers...
+
 [ ] Redesign options / I/O
     - Usage: stagit [-i] [-d DESTDIR] [-c CACHEDIR] REPO [REPO_2 ... REPO_N]
     - generates index page if -i
     - puts repos in their own subdirectory if -i or REPO_2
     - writes output into DESTDIR (defaults to pwd)
 
-[ ] handle missing decorations (desc/logo/favicon) better
-
 [ ] create config.h file
     - reused format strings (such as header)
 
 ---NEW FEATURES---
-[ ] add lines of code totals on files page
-[ ] add lines of code on each file's page
 [x] include binary files in static site + link to them
-[ ] add support for projects with folders in them
-    - show totals "123L + 456B" for each directory
-[ ] tweak paths that start with dots, so that they don't get hidden or 403'd
+[x] add support for projects with folders in them
+[x] backlink on subdirectory pages
+[ ] add lines of code totals on files page
+[x] add lines of code on each file's page
+[x] show line/byte totals for each directory
+[ ] tweak paths that start with dots, so that they don't get hidden or 403'd (?)
+[ ] handle missing decorations (desc/logo/favicon) better
 ---END NEW FEATURES---
 
+[ ] Remove globals
+  - RELPATH. wtf even...
+
 [ ] Redo caching, to never repeat work
   - but it needs to also not increase the complexity.
   - it's ok to duplicate storage
diff --git a/stagit.c b/stagit.c
@@ -16,6 +16,7 @@
 
 #include "compat.h"
 
+#define CPUT(f, c)  putc((c), (f))
 #define SPUTF(f, s)  fputs((s), (f))
 #define LENGTH(s)  (sizeof(s)/sizeof(*s))
 
@@ -57,6 +58,16 @@ struct ReferenceInfo {
 	struct CommitInfo *ci;
 };
 
+struct Weight {
+	size_t bytes;
+	size_t lines;
+};
+
+const char orders[] = { ' ', 'k', 'M', 'G', 'T', 'P', 'E' };
+
+/* GLOBALS */
+// TODO untangle these, pass them around instead...
+
 static git_repository *repo;
 
 /* flags + arguments */
@@ -78,7 +89,6 @@ static char *readme;
 static long long nlogcommits = -1; /* -1 indicates not used */
 
 /* cache */
-// TODO yuckyyy
 static git_oid lastoid;
 static char lastoidstr[GIT_OID_HEXSZ + 2]; /* id + newline + NUL byte */
 static FILE *rcachefp, *wcachefp;
@@ -114,6 +124,36 @@ join_path(char *buf, size_t bufsiz, const char *path, const char *path2)
 }
 
 void
+find_license()
+{
+	git_object *obj = NULL;
+	int i;
+	for (i = 0; i < LENGTH(license_files) && !license; i++) {
+		if (
+			!git_revparse_single(&obj, repo, license_files[i]) &&
+			git_object_type(obj) == GIT_OBJ_BLOB
+		)
+			license = license_files[i] + strlen("HEAD:");
+		git_object_free(obj);
+	}
+}
+
+void
+find_readme()
+{
+	git_object *obj = NULL;
+	int i;
+	for (i = 0; i < LENGTH(readme_files) && !readme; i++) {
+		if (
+			!git_revparse_single(&obj, repo, readme_files[i]) &&
+			git_object_type(obj) == GIT_OBJ_BLOB
+		)
+			readme = readme_files[i] + strlen("HEAD:");
+		git_object_free(obj);
+	}
+}
+
+void
 delta_info_free(struct DeltaInfo *di)
 {
 	if (!di) return;
@@ -375,17 +415,6 @@ err:
 	return -1;
 }
 
-FILE *
-efopen(const char *filename, const char *flags)
-{
-	FILE *fp;
-
-	if (!(fp = fopen(filename, flags)))
-		err(1, "fopen: '%s'", filename);
-
-	return fp;
-}
-
 /* Percent-encode, see RFC3986 section 2.1. */
 void
 put_percent_encoded(FILE *fp, const char *s, size_t len)
@@ -452,7 +481,7 @@ int
 mkdirp(const char *path)
 {
 	char tmp[PATH_MAX], *p;
-
+	
 	if (strlcpy(tmp, path, sizeof(tmp)) >= sizeof(tmp))
 		errx(1, "path truncated: '%s'", path);
 	for (p = tmp + (tmp[0] == '/'); *p; p++) {
@@ -463,11 +492,21 @@ mkdirp(const char *path)
 			return -1;
 		*p = '/';
 	}
-	if (mkdir(tmp, S_IRWXU | S_IRWXG | S_IRWXO) < 0 && errno != EEXIST)
-		return -1;
 	return 0;
 }
 
+FILE *
+efopen(const char *filename, const char *flags)
+{
+	FILE *fp;
+
+	mkdirp(filename);
+	if (!(fp = fopen(filename, flags)))
+		err(1, "fopen: '%s'", filename);
+
+	return fp;
+}
+
 void
 printtimez(FILE *fp, const git_time *intime)
 {
@@ -525,6 +564,12 @@ printtimeshort(FILE *fp, const git_time *intime)
 	fputs(out, fp);
 }
 
+char *
+put_absolute_path(FILE *fp, const char * path)
+{
+	fprintf(fp, "/%s%s", relpath, path);
+}
+
 void
 put_header(FILE *fp, const char *title)
 {
@@ -544,78 +589,98 @@ put_header(FILE *fp, const char *title)
 	if (description[0])
 		fputs(" - ", fp);
 	put_xml(fp, description, strlen(description));
-	fprintf(fp, "</title>\n<link rel=\"icon\" type=\"image/png\" href=\"%sfavicon.png\" />\n", relpath);
-	fputs("<link rel=\"alternate\" type=\"application/atom+xml\" title=\"", fp);
+	SPUTF(fp, "</title>" "\n");
+
+	SPUTF(fp, "<link rel=\"icon\" type=\"image/png\" href=\"");
+	put_absolute_path(fp, "favicon.png");
+	SPUTF(fp, "\" />" "\n");
+
+	SPUTF(fp, "<link rel=\"alternate\" type=\"application/atom+xml\" title=\"");
 	put_xml(fp, name, strlen(name));
-	fprintf(fp, " Atom Feed\" href=\"%satom.xml\" />\n", relpath);
-	fputs("<link rel=\"alternate\" type=\"application/atom+xml\" title=\"", fp);
+	SPUTF(fp, " Atom Feed\" href=\"");
+	put_absolute_path(fp, "atom.xml");
+	SPUTF(fp, "\" />" "\n");
+
+	SPUTF(fp, "<link rel=\"alternate\" type=\"application/atom+xml\" title=\"");
 	put_xml(fp, name, strlen(name));
-	fprintf(fp, " Atom Feed (tags)\" href=\"%stags.xml\" />\n", relpath);
-	fprintf(fp, "<link rel=\"stylesheet\" type=\"text/css\" href=\"%sstyle.css\" />\n", relpath);
-	fputs("</head>" "\n" "<body>" "\n", fp);
+	SPUTF(fp, " Atom Feed (tags)\" href=\"");
+	put_absolute_path(fp, "tags.xml");
+	SPUTF(fp, "\" />" "\n");
 
-	// TODO handle backlink using config.h
-	fprintf(fp, "<p><a href=\"%s\">home</a></p>" "\n", "/files.html");
+	SPUTF(fp, "<link rel=\"stylesheet\" type=\"text/css\" href=\"");
+	put_absolute_path(fp, "style.css");
+	SPUTF(fp, "\" />" "\n");
+	SPUTF(fp, "</head>" "\n" "<body>" "\n");
 
-	fputs("<table><tr><td>", fp);
-	fprintf(
-		fp,
-		"<a href=\"../%s\"><img src=\"%slogo.png\" alt=\"\" width=\"32\" height=\"32\" /></a>",
-		relpath,
-		relpath
-	);
-	fputs("</td><td><h1>", fp);
+	// TODO handle backlink using config.h
+	SPUTF(fp, "<p><a href=\"");
+	put_absolute_path(fp, "files.html");
+	SPUTF(fp, "\">home</a></p>" "\n");
+
+	SPUTF(fp, "<table><tr><td>");
+	fprintf(fp, "<a href=\"");
+	put_absolute_path(fp, ""); // TODO links to site root?
+	SPUTF(fp, "\"><img src=\"");
+	put_absolute_path(fp, "logo.png");
+	SPUTF(fp, "\" alt=\"\" width=\"32\" height=\"32\" /></a>");
+
+	SPUTF(fp, "</td><td><h1>");
 	put_xml(fp, stripped_name, strlen(stripped_name));
-	fputs("</h1><span class=\"desc\">", fp);
+	SPUTF(fp, "</h1><span class=\"desc\">");
 	put_xml(fp, description, strlen(description));
-	fputs("</span></td></tr>", fp);
+	SPUTF(fp, "</span></td></tr>");
 	if (cloneurl[0]) {
-		fputs("<tr class=\"url\"><td></td><td>git clone <a href=\"", fp);
+		SPUTF(fp, "<tr class=\"url\"><td></td><td>git clone <a href=\"");
 		put_xml(fp, cloneurl, strlen(cloneurl)); /* not percent-encoded */
-		fputs("\">", fp);
+		SPUTF(fp, "\">");
 		put_xml(fp, cloneurl, strlen(cloneurl));
-		fputs("</a></td></tr>", fp);
+		SPUTF(fp, "</a></td></tr>");
 	}
-	fputs("<tr><td></td><td>\n", fp);
+	SPUTF(fp, "<tr><td></td><td>" "\n");
 
-	if (readme)
-		fprintf(
-			fp,
-			"<a href=\"%sfile/%s.html\">README</a> | ",
-			relpath,
-			readme
-		);
+	if (readme) {
+		SPUTF(fp, "<a href=\"");
+		put_absolute_path(fp, "files/");
+		fprintf(fp, "%s.html", readme);
+		SPUTF(fp, "\">README</a> | ");
+	}
 	
-	fprintf(fp, "<a href=\"%sfiles.html\">Files</a> | ", relpath);
-	fprintf(fp, "<a href=\"%slog.html\">Log</a> | ", relpath);
-	fprintf(fp, "<a href=\"%srefs.html\">Refs</a>", relpath);
-
-	if (submodules)
-		fprintf(
-			fp,
-			" | <a href=\"%sfile/%s.html\">Submodules</a>",
-			relpath,
-			submodules
-		);
+	SPUTF(fp, "<a href=\"");
+	put_absolute_path(fp, "files.html");
+	SPUTF(fp, "\">Files</a> | ");
+
+	SPUTF(fp, "<a href=\"");
+	put_absolute_path(fp, "log.html");
+	SPUTF(fp, "\">Log</a> | ");
+
+	SPUTF(fp, "<a href=\"");
+	put_absolute_path(fp, "refs.html");
+	SPUTF(fp, "\">Refs</a>");
+
+	if (submodules) {
+		SPUTF(fp, " | <a href=\"");
+		put_absolute_path(fp, "files/");
+		fprintf(fp, "%s.html", submodules);
+		SPUTF(fp, "\">Submodules</a>");
+	}
 
-	if (license)
-		fprintf(
-			fp,
-			" | <a href=\"%sfile/%s.html\">LICENSE</a>",
-			relpath,
-			license
-		);
+	if (license) {
+		SPUTF(fp, " | <a href=\"");
+		put_absolute_path(fp, "files/");
+		fprintf(fp, "%s.html", license);
+		SPUTF(fp, "\">LICENSE</a>");
+	}
 
-	fputs("</td></tr></table>" "\n", fp);
-	fputs("<hr/>" "\n", fp);
+	SPUTF(fp, "</td></tr></table>" "\n");
+	SPUTF(fp, "<hr/>" "\n");
 
-	fputs("<div id=\"content\">" "\n", fp);
+	SPUTF(fp, "<div id=\"content\">" "\n");
 }
 
 void
 put_footer(FILE *fp)
 {
-	fputs("</div>" "\n" "</body>" "\n" "</html>" "\n", fp);
+	SPUTF(fp, "</div>" "\n" "</body>" "\n" "</html>" "\n");
 }
 
 size_t
@@ -655,33 +720,37 @@ put_file_html(FILE *fp, const git_blob *blob)
 void
 print_commit(FILE *fp, struct CommitInfo *ci)
 {
-	fprintf(fp, "<b>commit</b> <a href=\"%scommit/%s.html\">%s</a>\n",
-		relpath, ci->oid, ci->oid);
+	SPUTF(fp, "<b>commit</b> <a href=\"");
+	put_absolute_path(fp, "commit/");
+	fprintf(fp, "%s.html", ci->oid);
+	fprintf(fp, "\">%s</a>" "\n", ci->oid);
 
 	if (ci->parentoid[0])
-		fprintf(fp, "<b>parent</b> <a href=\"%scommit/%s.html\">%s</a>\n",
-			relpath, ci->parentoid, ci->parentoid);
+		SPUTF(fp, "<b>parent</b> <a href=\"");
+		put_absolute_path(fp, "commit/");
+		fprintf(fp, "%s.html", ci->parentoid);
+		fprintf(fp, "\">%s</a>" "\n", ci->parentoid);
 
 	if (ci->author) {
-		fputs("<b>Author:</b> ", fp);
+		SPUTF(fp, "<b>Author:</b> ");
 		put_xml(fp, ci->author->name, strlen(ci->author->name));
-		fputs(" &lt;<a href=\"mailto:", fp);
+		SPUTF(fp, " &lt;<a href=\"mailto:");
 		put_xml(fp, ci->author->email, strlen(ci->author->email)); /* not percent-encoded */
-		fputs("\">", fp);
+		SPUTF(fp, "\">");
 		put_xml(fp, ci->author->email, strlen(ci->author->email));
-		fputs("</a>&gt;\n<b>Date:</b>   ", fp);
+		SPUTF(fp, "</a>&gt;\n<b>Date:</b>   ");
 		printtime(fp, &(ci->author->when));
-		putc('\n', fp);
+		CPUT(fp, '\n');
 	}
 	if (ci->msg) {
-		putc('\n', fp);
+		CPUT(fp, '\n');
 		put_xml(fp, ci->msg, strlen(ci->msg));
-		putc('\n', fp);
+		CPUT(fp, '\n');
 	}
 }
 
 void
-print_show_file(FILE *fp, struct CommitInfo *ci)
+put_show_file(FILE *fp, struct CommitInfo *ci)
 {
 	const git_diff_delta *delta;
 	const git_diff_hunk *hunk;
@@ -771,19 +840,26 @@ print_show_file(FILE *fp, struct CommitInfo *ci)
 	for (i = 0; i < ci->ndeltas; i++) {
 		patch = ci->deltas[i]->patch;
 		delta = git_patch_get_delta(patch);
-		fprintf(fp, "<b>diff --git a/<a id=\"h%zu\" href=\"%sfile/", i, relpath);
+
+		fprintf(fp, "<b>diff --git a/<a id=\"h%zu\" href=\"", i);
+
+		put_absolute_path(fp, "files/");
 		put_percent_encoded(fp, delta->old_file.path, strlen(delta->old_file.path));
-		fputs(".html\">", fp);
+		SPUTF(fp, ".html\">");
+
 		put_xml(fp, delta->old_file.path, strlen(delta->old_file.path));
-		fprintf(fp, "</a> b/<a href=\"%sfile/", relpath);
+		SPUTF(fp, "</a> b/<a href=\"");
+
+		put_absolute_path(fp, "files/");
 		put_percent_encoded(fp, delta->new_file.path, strlen(delta->new_file.path));
-		fprintf(fp, ".html\">");
+		SPUTF(fp, ".html\">");
+
 		put_xml(fp, delta->new_file.path, strlen(delta->new_file.path));
-		fprintf(fp, "</a></b>\n");
+		SPUTF(fp, "</a></b>" "\n");
 
 		/* check binary data */
 		if (delta->flags & GIT_DIFF_FLAG_BINARY) {
-			fputs("Binary files differ.\n", fp);
+			SPUTF(fp, "Binary files differ." "\n");
 			continue;
 		}
 
@@ -794,7 +870,7 @@ print_show_file(FILE *fp, struct CommitInfo *ci)
 
 			fprintf(fp, "<a href=\"#h%zu-%zu\" id=\"h%zu-%zu\" class=\"h\">", i, j, i, j);
 			put_xml(fp, hunk->header, hunk->header_len);
-			fputs("</a>", fp);
+			SPUTF(fp, "</a>");
 
 			for (k = 0; ; k++) {
 				if (git_patch_get_line_in_hunk(&line, patch, j, k))
@@ -812,38 +888,41 @@ print_show_file(FILE *fp, struct CommitInfo *ci)
 						i, j, k, i, j, k
 					);
 				else
-					putc(' ', fp);
+					CPUT(fp, ' ');
 				put_xml_line(fp, line->content, line->content_len);
-				putc('\n', fp);
+				CPUT(fp, '\n');
 				if (line->old_lineno == -1 || line->new_lineno == -1)
-					fputs("</a>", fp);
+					SPUTF(fp, "</a>");
 			}
 		}
 	}
 }
 
 void
-write_log_line(FILE *fp, struct CommitInfo *ci)
+put_log_line(FILE *fp, struct CommitInfo *ci)
 {
 	fputs("<tr><td>", fp);
 	if (ci->author)
 		printtimeshort(fp, &(ci->author->when));
 	fputs("</td><td>", fp);
 	if (ci->summary) {
-		fprintf(fp, "<a href=\"%scommit/%s.html\">", relpath, ci->oid);
+		SPUTF(fp, "<a href=\"");
+		put_absolute_path(fp, "commit/");
+		fprintf(fp, "%s.html", ci->oid);
+		SPUTF(fp, "\">");
 		put_xml(fp, ci->summary, strlen(ci->summary));
-		fputs("</a>", fp);
+		SPUTF(fp, "</a>");
 	}
-	fputs("</td><td>", fp);
+	SPUTF(fp, "</td><td>");
 	if (ci->author)
 		put_xml(fp, ci->author->name, strlen(ci->author->name));
-	fputs("</td><td class=\"num\" align=\"right\">", fp);
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\">");
 	fprintf(fp, "%zu", ci->file_count);
-	fputs("</td><td class=\"num\" align=\"right\">", fp);
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\">");
 	fprintf(fp, "+%zu", ci->add_count);
-	fputs("</td><td class=\"num\" align=\"right\">", fp);
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\">");
 	fprintf(fp, "-%zu", ci->del_count);
-	fputs("</td></tr>\n", fp);
+	SPUTF(fp, "</td></tr>" "\n");
 }
 
 int
@@ -887,13 +966,13 @@ write_log(FILE *fp, const git_oid *oid)
 			goto err;
 
 		if (nlogcommits != 0) {
-			write_log_line(fp, ci);
+			put_log_line(fp, ci);
 			if (nlogcommits > 0)
 				nlogcommits--;
 		}
 
 		if (cachefile)
-			write_log_line(wcachefp, ci);
+			put_log_line(wcachefp, ci);
 
 		/* check if file exists if so skip it */
 		if (r) {
@@ -902,7 +981,7 @@ write_log(FILE *fp, const git_oid *oid)
 			put_header(fpfile, ci->summary);
 
 			fputs("<pre>", fpfile);
-			print_show_file(fpfile, ci);
+			put_show_file(fpfile, ci);
 			fputs("</pre>\n", fpfile);
 
 			put_footer(fpfile);
@@ -992,7 +1071,7 @@ write_log_cached(FILE *fp, const git_oid *head)
 }
 
 void
-write_log_page(FILE *fp, const git_oid *head)
+put_log_html(FILE *fp, const git_oid *head)
 {
 	relpath = "";
 	mkdir("commit", S_IRWXU | S_IRWXG | S_IRWXO);
@@ -1025,24 +1104,24 @@ print_commit_atom(FILE *fp, struct CommitInfo *ci, const char *tag)
 
 	fprintf(fp, "<id>%s</id>\n", ci->oid);
 	if (ci->author) {
-		fputs("<published>", fp);
+		SPUTF(fp, "<published>");
 		printtimez(fp, &(ci->author->when));
-		fputs("</published>\n", fp);
+		SPUTF(fp, "</published>\n");
 	}
 	if (ci->committer) {
-		fputs("<updated>", fp);
+		SPUTF(fp, "<updated>");
 		printtimez(fp, &(ci->committer->when));
-		fputs("</updated>\n", fp);
+		SPUTF(fp, "</updated>\n");
 	}
 	if (ci->summary) {
-		fputs("<title>", fp);
+		SPUTF(fp, "<title>");
 		if (tag && tag[0]) {
-			fputs("[", fp);
+			SPUTF(fp, "[");
 			put_xml(fp, tag, strlen(tag));
-			fputs("] ", fp);
+			SPUTF(fp, "] ");
 		}
 		put_xml(fp, ci->summary, strlen(ci->summary));
-		fputs("</title>\n", fp);
+		SPUTF(fp, "</title>\n");
 	}
 	fprintf(
 		fp,
@@ -1052,31 +1131,31 @@ print_commit_atom(FILE *fp, struct CommitInfo *ci, const char *tag)
 	);
 
 	if (ci->author) {
-		fputs("<author>\n<name>", fp);
+		SPUTF(fp, "<author>\n<name>");
 		put_xml(fp, ci->author->name, strlen(ci->author->name));
-		fputs("</name>\n<email>", fp);
+		SPUTF(fp, "</name>\n<email>");
 		put_xml(fp, ci->author->email, strlen(ci->author->email));
-		fputs("</email>\n</author>\n", fp);
+		SPUTF(fp, "</email>\n</author>\n");
 	}
 
-	fputs("<content>", fp);
+	SPUTF(fp, "<content>");
 	fprintf(fp, "commit %s\n", ci->oid);
 	if (ci->parentoid[0])
 		fprintf(fp, "parent %s\n", ci->parentoid);
 	if (ci->author) {
-		fputs("Author: ", fp);
+		SPUTF(fp, "Author: ");
 		put_xml(fp, ci->author->name, strlen(ci->author->name));
-		fputs(" &lt;", fp);
+		SPUTF(fp, " &lt;");
 		put_xml(fp, ci->author->email, strlen(ci->author->email));
-		fputs("&gt;\nDate:   ", fp);
+		SPUTF(fp, "&gt;\nDate:   ");
 		printtime(fp, &(ci->author->when));
-		putc('\n', fp);
+		CPUT(fp, '\n');
 	}
 	if (ci->msg) {
-		putc('\n', fp);
+		CPUT(fp, '\n');
 		put_xml(fp, ci->msg, strlen(ci->msg));
 	}
-	fputs("\n</content>\n</entry>\n", fp);
+	SPUTF(fp, "\n</content>\n</entry>\n");
 }
 
 int
@@ -1089,15 +1168,15 @@ write_atom(FILE *fp, int all)
 	git_oid id;
 	size_t i, m = 100; /* last 'm' commits */
 
-	fputs(
+	SPUTF(
+		fp,
 		"<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"
-		"<feed xmlns=\"http://www.w3.org/2005/Atom\">\n<title>",
-		fp
+		"<feed xmlns=\"http://www.w3.org/2005/Atom\">\n<title>"
 	);
 	put_xml(fp, stripped_name, strlen(stripped_name));
-	fputs(", branch HEAD</title>\n<subtitle>", fp);
+	SPUTF(fp, ", branch HEAD</title>\n<subtitle>");
 	put_xml(fp, description, strlen(description));
-	fputs("</subtitle>\n", fp);
+	SPUTF(fp, "</subtitle>\n");
 
 	/* all commits or only tags? */
 	if (all) {
@@ -1131,13 +1210,13 @@ write_atom(FILE *fp, int all)
 }
 
 void
-write_binary_file(const char *filepath, git_blob *blob) {
+write_binary_file(const char *filename, git_blob *blob) {
 	char dest[PATH_MAX] = "";
 	const char *raw = git_blob_rawcontent(blob);
 	int size = git_blob_rawsize(blob);
 
 	/* set dest */
-	join_path(dest, PATH_MAX, "file", filepath);
+	join_path(dest, PATH_MAX, "files", filename);
 
 	/* write the file! */
 	FILE *fp = efopen(dest, "w");
@@ -1146,43 +1225,61 @@ write_binary_file(const char *filepath, git_blob *blob) {
 	fclose(fp);
 }
 
+void
+put_size(FILE *fp, size_t size)
+{
+	int order = 0;
+
+	if (size == 0) {
+		fprintf(fp, "  - ");
+		return;
+	}
+
+	while (size > 1000) {
+		order++;
+		size /= 1000;
+	}
+
+	fprintf(fp, "%3zu%c", size, orders[order]);
+}
+
 size_t
-write_blob(
+put_blob(
 	git_object *obj,
-	const char *fpath, /* absolute path of page */
-	const char *entrypath, /* directory of file */
-	const char *filename, /* name of file */
-	size_t filesize
+	const char *filename,
+	const char *entrypath,
+	const char *entryname,
+	size_t bytes
 ) {
-	char tmp[PATH_MAX] = "", *d;
+	char tmp[PATH_MAX] = "";
+	char *d;
 	const char *p;
-	size_t lc = 0;
+	size_t lines = 0;
 	FILE *fp;
 	int is_binary;
 
-	if (strlcpy(tmp, fpath, sizeof(tmp)) >= sizeof(tmp))
-		errx(1, "path truncated: '%s'", fpath);
-
-	if (!(d = dirname(tmp)))
-		err(1, "dirname");
-
-	if (mkdirp(d))
-		return -1;
+	if (strlcpy(tmp, filename, sizeof(tmp)) >= sizeof(tmp))
+		errx(1, "path truncated: '%s'", filename);
 
-	for (p = fpath, tmp[0] = '\0'; *p; p++) {
+	// wtf does this do....
+	for (p = filename, tmp[0] = '\0'; *p; p++) {
 		if (*p == '/' && strlcat(tmp, "../", sizeof(tmp)) >= sizeof(tmp))
 			errx(1, "path truncated: '../%s'", tmp);
 	}
 	relpath = tmp;
 
 	is_binary = git_blob_is_binary((git_blob *)obj);
-	const char *file_url = is_binary ? filename : "";
+	const char *file_url = is_binary ? entryname : "";
 
-	fp = efopen(fpath, "w");
-	put_header(fp, filename);
+	fp = efopen(filename, "w");
+	put_header(fp, entryname);
 	fprintf(fp, "<p>");
-	put_xml(fp, filename, strlen(filename));
-	fprintf(fp, " (%zuB)</p>", filesize); // TODO also show line number (if possible)
+	put_xml(fp, entrypath, strlen(entrypath));
+	CPUT(fp, ' ');
+	put_size(fp, 0);
+	CPUT(fp, ' ');
+	put_size(fp, bytes);
+	SPUTF(fp, "B</p>");
 
 	if (is_binary) {
 		fprintf(
@@ -1196,20 +1293,20 @@ write_blob(
 		);
 		write_binary_file(entrypath, (git_blob *) obj);
 	} else {
-		lc = put_file_html(fp, (git_blob *) obj);
+		lines = put_file_html(fp, (git_blob *) obj);
 	}
 
 	put_footer(fp);
-	check_file_error(fp, fpath, 'w');
+	check_file_error(fp, filename, 'w');
 	fclose(fp);
 
 	relpath = "";
 
-	return lc;
+	return lines;
 }
 
 char
-filemode_type(git_filemode_t m)
+get_filemode_type(git_filemode_t m)
 {
 	if (S_ISREG(m)) return '-';
 	if (S_ISBLK(m)) return 'b';
@@ -1222,133 +1319,303 @@ filemode_type(git_filemode_t m)
 }
 
 const char *
-filemode(git_filemode_t m)
+get_filemode(const git_tree_entry *entry)
 {
-	static char mode[11];
-
-	memset(mode, '-', sizeof(mode) - 1);
-	mode[0] = filemode_type(m);
-	if (m & S_IRUSR) mode[1] = 'r';
-	if (m & S_IWUSR) mode[2] = 'w';
-	if (m & S_IXUSR) mode[3] = 'x';
-	if (m & S_ISUID) mode[3] = (mode[3] == 'x') ? 's' : 'S';
-	if (m & S_IRGRP) mode[4] = 'r';
-	if (m & S_IWGRP) mode[5] = 'w';
-	if (m & S_IXGRP) mode[6] = 'x';
-	if (m & S_ISGID) mode[6] = (mode[6] == 'x') ? 's' : 'S';
-	if (m & S_IROTH) mode[7] = 'r';
-	if (m & S_IWOTH) mode[8] = 'w';
-	if (m & S_IXOTH) mode[9] = 'x';
-	if (m & S_ISVTX) mode[9] = (mode[9] == 'x') ? 't' : 'T';
-
-	mode[10] = '\0';
-
-	return mode;
+	git_filemode_t mode;
+	static char ret[11];
+
+	/* used for backlink: .. */
+	if (entry == NULL)
+		return "d---------";
+
+	mode = git_tree_entry_filemode(entry);
+
+	memset(ret, '-', sizeof(ret) - 1);
+	ret[0] = get_filemode_type(mode);
+	if (mode & S_IRUSR) ret[1] = 'r';
+	if (mode & S_IWUSR) ret[2] = 'w';
+	if (mode & S_IXUSR) ret[3] = 'x';
+	if (mode & S_ISUID) ret[3] = (ret[3] == 'x') ? 's' : 'S';
+	if (mode & S_IRGRP) ret[4] = 'r';
+	if (mode & S_IWGRP) ret[5] = 'w';
+	if (mode & S_IXGRP) ret[6] = 'x';
+	if (mode & S_ISGID) ret[6] = (ret[6] == 'x') ? 's' : 'S';
+	if (mode & S_IROTH) ret[7] = 'r';
+	if (mode & S_IWOTH) ret[8] = 'w';
+	if (mode & S_IXOTH) ret[9] = 'x';
+	if (mode & S_ISVTX) ret[9] = (ret[9] == 'x') ? 't' : 'T';
+	ret[10] = '\0';
+
+	return ret;
 }
 
-int
-write_file_list(FILE *fp, git_tree *tree, const char *path)
+struct Weight
+put_filetree_file_line(
+	FILE *fp,
+	git_object *obj,
+	const git_tree_entry *entry,
+	char *filename,
+	char *entrypath,
+	const char *entryname
+) {
+	size_t bytes = git_blob_rawsize((git_blob *)obj);
+	size_t lines = put_blob(obj, filename, entrypath, entryname, bytes);
+	struct Weight ret = { bytes, lines };
+
+	SPUTF(fp, "<tr><td>");
+	SPUTF(fp, get_filemode(entry));
+	SPUTF(fp, "</td><td><a href=\"");
+	put_absolute_path(fp, "");
+	put_percent_encoded(fp, filename, strlen(filename));
+	SPUTF(fp, "\">");
+	put_xml(fp, entryname, strlen(entryname));
+	SPUTF(fp, "</a></td><td class=\"num\" align=\"right\">");
+	put_size(fp, lines);
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\">");
+	put_size(fp, bytes);
+	SPUTF(fp, "</td></tr>\n");
+	return ret;
+}
+
+void
+put_filetree_dir_line(
+	FILE *fp,
+	const git_tree_entry *entry,
+	char *destination,
+	const char *title,
+	size_t lines,
+	size_t bytes
+) {
+	SPUTF(fp, "<tr><td>");
+	SPUTF(fp, get_filemode(entry));
+	SPUTF(fp, "</td><td><a href=\"");
+	put_absolute_path(fp, "");
+	put_percent_encoded(fp, destination, strlen(destination));
+	SPUTF(fp, "\">");
+	put_xml(fp, title, strlen(title));
+	SPUTF(fp, "</a></td><td class=\"num\" align=\"right\">");
+	put_size(fp, lines);
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\">");
+	put_size(fp, bytes);
+	SPUTF(fp, "</td></tr>\n");
+}
+
+void put_filetree_backlink_line(FILE *fp, const char *path) {
+	char previous_path[PATH_MAX];
+	char previous_filename[PATH_MAX];
+	int i, r, last_slash = 0;
+	
+	/* figure out previous directory */
+	// TODO how to do this more cleanly...
+	for (i = 0; path[i] != '\0'; i++) {
+		previous_path[i] = path[i];
+		if (path[i] == '/') last_slash = i;
+	}
+	previous_path[last_slash] = '\0';
+
+	if (last_slash == 0) {
+		put_filetree_dir_line(fp, NULL, "files.html", "..", 0, 0);
+	} else {
+		r = snprintf(previous_filename, sizeof(previous_filename), "files/%s.html", previous_path);
+		if (r < 0 || (size_t) r >= sizeof(previous_filename))
+			errx(1, "path truncated: 'files/%s.html'", previous_path);
+		put_filetree_dir_line(fp, NULL, previous_filename, "..", 0, 0);
+	}
+}
+
+/* need this here to allow recursion */
+struct Weight write_filetree(
+	const char *filename,
+	const char *path,
+	git_tree *tree
+);
+
+struct Weight
+put_entry_obj(
+	FILE *fp,
+	git_object *obj,
+	const git_tree_entry *entry,
+	char *entrypath,
+	const char *entryname
+) {
+	struct Weight ret;
+	git_object_t obj_type = git_object_type(obj);
+	char filename[PATH_MAX];
+	int r;
+
+	/* filename = f("files/%s.html", entrypath) */
+	r = snprintf(filename, sizeof(filename), "files/%s.html", entrypath);
+	if (r < 0 || (size_t) r >= sizeof(filename))
+		errx(1, "path truncated: 'files/%s.html'", entrypath);
+
+	switch (obj_type) {
+	case GIT_OBJ_TREE:
+		/* NOTE: recurses */
+		ret = write_filetree(filename, entrypath, (git_tree *) obj);
+		put_filetree_dir_line(
+			fp,
+			entry,
+			filename,
+			entryname,
+			ret.lines,
+			ret.bytes
+		);
+		break;
+	case GIT_OBJ_BLOB:
+		ret = put_filetree_file_line(
+			fp,
+			obj,
+			entry,
+			filename,
+			entrypath,
+			entryname
+		);
+		break;
+	}
+
+	return ret;
+}
+
+void
+put_submodule_obj(
+	FILE *fp,
+	const git_tree_entry *entry,
+	char *entrypath
+) {
+	char oid[8];
+
+	SPUTF(fp, "<tr><td>m---------</td><td><a href=\"");
+	put_absolute_path(fp, "files/.gitmodules.html");
+	SPUTF(fp, "\">");
+	put_xml(fp, entrypath, strlen(entrypath));
+	SPUTF(fp, "</a> @ ");
+	git_oid_tostr(oid, sizeof(oid), git_tree_entry_id(entry));
+	put_xml(fp, oid, strlen(oid));
+	SPUTF(fp, "</td><td class=\"num\" align=\"right\"></td></tr>" "\n");
+}
+
+void
+put_entry()
 {
-	const git_tree_entry *entry = NULL;
-	git_object *obj = NULL;
-	const char *entryname;
-	char filepath[PATH_MAX], entrypath[PATH_MAX], oid[8];
-	size_t count, i, lc, filesize;
-	int r, ret;
+
+}
+
+struct Weight
+put_file_list(FILE *fp, git_tree *tree, const char *path)
+{
+	struct Weight ret = {};
+	size_t count;
+	size_t i;
+
+	/* add line for ".." back-link */
+	if (path[0] != '\0') put_filetree_backlink_line(fp, path);
 
 	count = git_tree_entrycount(tree);
 	for (i = 0; i < count; i++) {
+		const git_tree_entry *entry = NULL;
+		const char *entryname;
+		char entrypath[PATH_MAX];
+
 		if (
 			!(entry = git_tree_entry_byindex(tree, i)) ||
 			!(entryname = git_tree_entry_name(entry))
 		)
-			return -1;
-		join_path(entrypath, sizeof(entrypath), path, entryname);
+			return ret;
 
-		r = snprintf(filepath, sizeof(filepath), "file/%s.html", entrypath);
-		if (r < 0 || (size_t)r >= sizeof(filepath))
-			errx(1, "path truncated: 'file/%s.html'", entrypath);
+		/* entrypath = path + entryname */
+		join_path(entrypath, sizeof(entrypath), path, entryname);
 
-		if (!git_tree_entry_to_object(&obj, repo, entry)) {
-			switch (git_object_type(obj)) {
-			case GIT_OBJ_BLOB:
-				break;
-			case GIT_OBJ_TREE:
-				// TODO create page for each dir
-				// instead of dumping all files into "files.html"
-				// or, is that a good idea? hm...
-				/* NOTE: recurses */
-				ret = write_file_list(fp, (git_tree *)obj, entrypath);
-				git_object_free(obj);
-				if (ret)
-					return ret;
-				continue;
-			default:
+		{
+			git_object *obj = NULL;
+			if (!git_tree_entry_to_object(&obj, repo, entry)) {
+				struct Weight weight = put_entry_obj(
+					fp,
+					obj,
+					entry,
+					entrypath,
+					entryname
+				);
+				ret.bytes += weight.bytes;
+				ret.lines += weight.lines;
 				git_object_free(obj);
 				continue;
 			}
+		}
 
-			filesize = git_blob_rawsize((git_blob *)obj);
-			lc = write_blob(obj, filepath, entrypath, entryname, filesize);
-
-			fputs("<tr><td>", fp);
-			fputs(filemode(git_tree_entry_filemode(entry)), fp);
-			fprintf(fp, "</td><td><a href=\"%s", relpath);
-			put_percent_encoded(fp, filepath, strlen(filepath));
-			fputs("\">", fp);
-			put_xml(fp, entrypath, strlen(entrypath));
-			fputs("</a></td><td class=\"num\" align=\"right\">", fp);
-			if (lc > 0) fprintf(fp, "%zuL", lc);
-			else fprintf(fp, "%zuB", filesize);
-			fputs("</td></tr>\n", fp);
-			git_object_free(obj);
-		} else if (git_tree_entry_type(entry) == GIT_OBJ_COMMIT) {
-			/* commit object in tree is a submodule */
-			fprintf(
-				fp,
-				"<tr><td>m---------</td><td><a href=\"%sfile/.gitmodules.html\">",
-				relpath
-			);
-			put_xml(fp, entrypath, strlen(entrypath));
-			fputs("</a> @ ", fp);
-			git_oid_tostr(oid, sizeof(oid), git_tree_entry_id(entry));
-			put_xml(fp, oid, strlen(oid));
-			fputs("</td><td class=\"num\" align=\"right\"></td></tr>\n", fp);
+		if (git_tree_entry_type(entry) == GIT_OBJ_COMMIT) {
+			put_submodule_obj(fp, entry, entrypath);
+			// TODO ensure that this works with the size stuff correctly!
+			continue;
 		}
 	}
 
-	return 0;
+	return ret;
 }
 
-int
-write_files(FILE *fp, const git_oid *id)
+void
+put_filetree_dir_title(FILE *fp, const char *path)
 {
-	git_tree *tree = NULL;
-	git_commit *commit = NULL;
-	int ret = -1;
+	put_xml(fp, path, strlen(path));
+}
 
-	fputs(
-		"<table id=\"files\"><thead>\n<tr>"
+struct Weight
+write_filetree(const char *filename, const char *path, git_tree *tree)
+{
+	FILE *fp;
+	struct Weight ret;
+
+	fp = efopen(filename, "w");
+	put_header(fp, "Files");
+	
+	fprintf(fp, "<p>");
+	put_filetree_dir_title(fp, path);
+	SPUTF(fp, "</p>");
+
+	SPUTF(
+		fp,
+		"<table id=\"files\"><thead>" "\n" "<tr>"
 		"<td><b>Mode</b></td><td><b>Name</b></td>"
-		"<td class=\"num\" align=\"right\"><b>Size</b></td>"
-		"</tr>\n</thead><tbody>\n",
-		fp
+		"<td class=\"num\" align=\"right\"><b>L</b></td>"
+		"<td class=\"num\" align=\"right\"><b>B</b></td>"
+		"</tr>" "\n" "</thead><tbody>" "\n"
 	);
 
-	if (
-		!git_commit_lookup(&commit, repo, id) &&
-		!git_commit_tree(&tree, commit)
-	)
-		ret = write_file_list(fp, tree, "");
+	if (tree != NULL)
+		ret = put_file_list(fp, tree, path);
 
-	fputs("</tbody></table>", fp);
+	SPUTF(fp, "</tbody></table>");
 
-	git_commit_free(commit);
-	git_tree_free(tree);
+	SPUTF(fp, "<p>Totals: ");
+	put_size(fp, ret.lines);
+	SPUTF(fp, "L ");
+	put_size(fp, ret.bytes);
+	SPUTF(fp, "B</p>");
+
+	put_footer(fp);
+	check_file_error(fp, filename, 'w');
+	fclose(fp);
 
 	return ret;
 }
 
+void
+walk_git_tree(const git_oid *id)
+{
+	git_tree *tree = NULL;
+	git_commit *commit = NULL;
+	int has_files;
+	
+	if (!id) return;
+
+	has_files = !git_commit_lookup(&commit, repo, id) &&
+		!git_commit_tree(&tree, commit);
+
+	write_filetree("files.html", "", has_files ? tree : NULL);
+
+	git_commit_free(commit);
+	git_tree_free(tree);
+}
+
 int
 write_refs(FILE *fp)
 {
@@ -1410,6 +1677,48 @@ write_refs(FILE *fp)
 }
 
 void
+write_log_page(const git_oid *head)
+{
+	const char *filename = "log.html";
+	FILE *fp = efopen(filename, "w");
+	put_log_html(fp, head);
+	check_file_error(fp, filename, 'w');
+	fclose(fp);
+}
+
+void
+write_refs_page()
+{
+	const char *filename = "refs.html";
+	FILE *fp = efopen(filename, "w");
+	put_header(fp, "Refs");
+	write_refs(fp);
+	put_footer(fp);
+	check_file_error(fp, filename, 'w');
+	fclose(fp);
+}
+
+void
+write_feed()
+{
+	const char *filename = "atom.xml";
+	FILE *fp = efopen(filename, "w");
+	write_atom(fp, 1);
+	check_file_error(fp, filename, 'w');
+	fclose(fp);
+}
+
+void
+write_releases_feed()
+{
+	const char *filename = "tags.xml";
+	FILE *fp = efopen(filename, "w");
+	write_atom(fp, 0);
+	check_file_error(fp, filename, 'w');
+	fclose(fp);
+}
+
+void
 usage(char *argv0)
 {
 	fprintf(
@@ -1424,10 +1733,10 @@ usage(char *argv0)
 int
 main(int argc, char *argv[])
 {
-	git_object *obj = NULL;
 	const git_oid *head = NULL;
 	FILE *fp, *fpread;
 	char path[PATH_MAX], repodirabs[PATH_MAX + 1], *p;
+	git_object *obj = NULL;
 	int i;
 
 	/* parse args */
@@ -1544,25 +1853,8 @@ main(int argc, char *argv[])
 		cloneurl[strcspn(cloneurl, "\n")] = '\0';
 	}
 
-	/* check LICENSE */
-	for (i = 0; i < LENGTH(license_files) && !license; i++) {
-		if (
-			!git_revparse_single(&obj, repo, license_files[i]) &&
-			git_object_type(obj) == GIT_OBJ_BLOB
-		)
-			license = license_files[i] + strlen("HEAD:");
-		git_object_free(obj);
-	}
-
-	/* check README */
-	for (i = 0; i < LENGTH(readme_files) && !readme; i++) {
-		if (
-			!git_revparse_single(&obj, repo, readme_files[i]) &&
-			git_object_type(obj) == GIT_OBJ_BLOB
-		)
-			readme = readme_files[i] + strlen("HEAD:");
-		git_object_free(obj);
-	}
+	find_license();
+	find_readme();
 
 	if (
 		!git_revparse_single(&obj, repo, "HEAD:.gitmodules") &&
@@ -1571,39 +1863,12 @@ main(int argc, char *argv[])
 		submodules = ".gitmodules";
 	git_object_free(obj);
 
-	/* log for HEAD */
-	fp = efopen("log.html", "w");
-	write_log_page(fp, head);
-	check_file_error(fp, "log.html", 'w');
-	fclose(fp);
-
-	/* files for HEAD */
-	fp = efopen("files.html", "w");
-	put_header(fp, "Files");
-	if (head) write_files(fp, head);
-	put_footer(fp);
-	check_file_error(fp, "files.html", 'w');
-	fclose(fp);
-
-	/* summary page with branches and tags */
-	fp = efopen("refs.html", "w");
-	put_header(fp, "Refs");
-	write_refs(fp);
-	put_footer(fp);
-	check_file_error(fp, "refs.html", 'w');
-	fclose(fp);
-
-	/* Atom feed */
-	fp = efopen("atom.xml", "w");
-	write_atom(fp, 1);
-	check_file_error(fp, "atom.xml", 'w');
-	fclose(fp);
-
-	/* Atom feed for tags / releases */
-	fp = efopen("tags.xml", "w");
-	write_atom(fp, 0);
-	check_file_error(fp, "tags.xml", 'w');
-	fclose(fp);
+	/* write pages! */
+	write_log_page(head);
+	walk_git_tree(head);
+	write_refs_page();
+	write_feed();
+	write_releases_feed();
 
 	/* cleanup */
 	git_repository_free(repo);