Implement a new caching system to make repeated runs much much faster.

Makefile

@@ -2,13 +2,12 @@ include config.mk
 
 # flags and incs
 PKGS = $(GITLIB)
-CFLAGS = -DVERSION=\"$(VERSION)\" -Wall -pedantic -O3 $(GIT) $(ERR) $(EXPLAIN)
+CFLAGS = -DVERSION=\"$(VERSION)\" -Wall -pedantic -O3 $(GIT) $(GITHASH) $(ERR) $(EXPLAIN)
 LIBS = `$(PKG_CONFIG) --libs --cflags $(PKGS)`
 
 all: XD
-XD: XD.o
+XD: XD.o hash.o helpers.o
 	$(CC) *.o $(CFLAGS) $(LIBS) -o $@
-XD.o: XD.c
 
 clean:
 	rm -f XD *.o

XD.1

@@ -12,9 +12,12 @@
 .Sh DESCRIPTION
 .Nm
 Displays information using a smiley face like so: :)
-to interpret it refer to the following tables:
-Displays information using a smiley face.
-to interpret XD's output refer to the following tables (or
+.Nm
+does some special caching to run as fast as possible but there's only so much
+that can be done without relying on an external daemon. Therefore when new
+changes are made in very large repositories it may take a few seconds to
+determine if there are any changes. To interpret XD's output refer to the
+following tables (or
 .Nm
 \fB-e\fR):
 .Ss Eyes

XD.c

@@ -1,3 +1,4 @@
+#include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -16,42 +17,11 @@
 #include <git2.h>
 #endif
 
+#include "helpers.h"
+#include "hash.h"
+
 #define P(X) fwrite(X, 1, 1, stdout)
 
-#if defined(ERR) || defined(EXPLAIN)
-void
-l(const char *fmt, ...)
-{
-  va_list ap;
-
-  va_start(ap, fmt);
-  vfprintf(stderr, fmt, ap);
-  va_end(ap);
-
-  if (fmt[0] && fmt[strlen(fmt) - 1] == ':') {
-    fputc(' ', stderr);
-    perror(NULL);
-  } else {
-    fputc('\n', stderr);
-  }
-}
-#endif
-
-#ifdef ERR
-#define L(...) l(__VA_ARGS__)
-#else
-#define L(...)
-#endif
-
-#ifdef EXPLAIN
-static int explain = 0;
-#define E(...) if (explain) {    \
-                 l(__VA_ARGS__); \
-               } else
-#else
-#define E(...)
-#endif
-
 #ifdef GIT
 /**
  * @brief search all parent directories for a git repo
@@ -194,9 +164,20 @@ has_untracked(git_repository *repo)
 {
   git_status_options opts = GIT_STATUS_OPTIONS_INIT;
   git_status_list *list = NULL;
+  repohash *storedhash;
   int r = 0;
 
-  /* FIXME: this is really slow in large git repos :( */
+  #ifdef GITHASH
+  if ((storedhash = read_hash(repo))
+    && storedhash->hash == generate_hash(repo)) {
+    r = storedhash->changes;
+    free(storedhash);
+    return r;
+  }
+  #endif
+
+  /* if we need to regen the hash then we need to do a hard check on the real
+   * git repository */
   opts.show = GIT_STATUS_SHOW_INDEX_AND_WORKDIR;
   opts.flags = GIT_STATUS_OPT_INCLUDE_UNTRACKED |
     GIT_STATUS_OPT_RENAMES_HEAD_TO_INDEX |
@@ -209,9 +190,10 @@ has_untracked(git_repository *repo)
   }
 
   /* if any changes are found return 1 */
-  if (git_status_list_entrycount(list) > 0) {
-    r = 1;
-  }
+  r = git_status_list_entrycount(list) > 0;
+  #ifdef GITHASH
+  write_hash(repo, (repohash){ .hash = generate_hash(repo), .changes = r });
+  #endif
 
   git_status_list_free(list);
   return r;

config.mk

@@ -7,9 +7,11 @@ PREFIX = /usr/local
 MANDIR = $(PREFIX)/share/man
 
 GIT =
+GITHASH =
 GITLIB =
 # comment to disable git support
 GIT = -DGIT
+GITHASH = -DGITHASH
 GITLIB = libgit2
 
 ERR =
@@ -24,6 +26,11 @@ EXPLAIN = -DEXPLAIN
 ifneq ($(GIT),)
 	VERSION := $(VERSION)"\\nlibgit2 "`$(PKG_CONFIG) --modversion $(GITLIB)`
 endif
+ifeq ($(GITHASH),)
+	VERSION := $(VERSION)"\\ngit hashing disabled"
+else
+	VERSION := $(VERSION)"\\ngit hashing enabled"
+endif
 ifeq ($(ERR),)
 	VERSION := $(VERSION)"\\nerrors disabled"
 else

hash.c

@@ -0,0 +1,132 @@
+#include <git2/repository.h>
+#include <git2/types.h>
+#include <linux/limits.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <stdint.h>
+#include <errno.h>
+
+#include "hash.h"
+#include "helpers.h"
+
+#ifdef GITHASH
+static uint64_t
+murmur64(uint64_t k)
+{
+  k ^= k >> 33;
+  k *= 0xff51afd7ed558ccdLLU;
+  k ^= k >> 33;
+  k *= 0xc4ceb9fe1a85ec53LLU;
+  k ^= k >> 33;
+
+  return k;
+}
+
+uint64_t
+generate_hash(git_repository *repo)
+{
+  struct stat dir, index;
+  char path[PATH_MAX] = { 0 };
+
+  const char *gitpath = git_repository_path(repo);
+  if (strlen(gitpath) + strlen("/..") > PATH_MAX - 1) {
+    L("strlen: %s", strerror(errno));
+    return -1;
+  }
+
+  strcat(path, gitpath);
+  if (stat(path, &index) < 0) {
+    return -1;
+  }
+  strcat(path, "/..");
+  if (stat(path, &dir) < 0) {
+    return -1;
+  }
+  return murmur64(dir.st_mtim.tv_nsec ^ index.st_mtim.tv_nsec);
+}
+
+repohash
+*read_hash(git_repository *repo)
+{
+  FILE *f;
+  uint64_t data;
+  uint8_t changes;
+  repohash *hash = malloc(sizeof(repohash));
+  char path[PATH_MAX] = { 0 };
+
+  const char *gitpath = git_repository_path(repo);
+  if (strlen(gitpath) + strlen(XD_HASH_PATH) > PATH_MAX - 1) {
+    L("strlen: %s", strerror(errno));
+    return NULL;
+  }
+  strcat(path, gitpath);
+  strcat(path, XD_HASH_PATH);
+
+  f = fopen(path, "r");
+  if (!f) {
+    L("fopen: %s", strerror(errno));
+    return NULL;
+  }
+
+  if (fread(&data, sizeof(uint64_t), 1, f) != 1) {
+    L("fread: %s", strerror(errno));
+    return NULL;
+  }
+
+  if (fseek(f, sizeof(uint64_t), SEEK_SET) < 0) {
+    L("fseek: %s", strerror(errno));
+    return NULL;
+  }
+
+  if (fread(&changes, sizeof(uint8_t), 1, f) != 1) {
+    L("fread: %s", strerror(errno));
+    return NULL;
+  }
+  fclose(f);
+
+  hash->hash = data;
+  hash->changes = changes;
+  return hash;
+}
+
+int
+write_hash(git_repository *repo, repohash hash)
+{
+  FILE *f;
+  char path[PATH_MAX] = { 0 };
+
+  const char *gitpath = git_repository_path(repo);
+  if (strlen(gitpath) + strlen(XD_HASH_PATH) > PATH_MAX - 1) {
+    L("strlen: %s", strerror(errno));
+    return 1;
+  }
+  strcat(path, gitpath);
+  strcat(path, XD_HASH_PATH);
+
+  f = fopen(path, "wb");
+  if (!f) {
+    L("fopen: %s", strerror(errno));
+    return 1;
+  }
+
+  if (fwrite(&hash.hash, sizeof(uint64_t), 1, f) != 1) {
+    L("fwrite: %s", strerror(errno));
+    return 1;
+  }
+
+  if (fseek(f, sizeof(uint64_t), SEEK_SET) < 0) {
+    L("fseek: %s", strerror(errno));
+    return 1;
+  }
+
+  if (fwrite(&hash.changes, sizeof(uint8_t), 1, f) != 1) {
+    L("fwrite: %s", strerror(errno));
+    return 1;
+  }
+
+  fclose(f);
+  return 0;
+}
+#endif

hash.h

@@ -0,0 +1,37 @@
+#pragma once
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <git2/repository.h>
+
+#ifdef GIT
+typedef struct {
+  uint64_t hash;
+  bool changes;
+} repohash;
+
+#define XD_HASH_PATH "/XDhash"
+/**
+ * @brief generate a hash from the repository state
+ *
+ * @param repo the git repository
+ * @return the hash
+ */
+uint64_t generate_hash(git_repository *repo);
+
+/**
+ * @brief read the hash from the git repo
+ *
+ * @param repo the git repository
+ * @return the hash
+ */
+repohash *read_hash(git_repository *repo);
+
+/**
+ * @brief write a new hash to the repository
+ *
+ * @param repo the repository
+ * @param hash the hash to write
+ */
+int write_hash(git_repository *repo, repohash hash);
+#endif

helpers.c

@@ -0,0 +1,28 @@
+#include <stdarg.h>
+#include <stdio.h>
+#include <string.h>
+
+#include "helpers.h"
+
+#ifdef EXPLAIN
+int explain = 0;
+#endif
+
+#if defined(ERR) || defined(EXPLAIN)
+void
+l(const char *fmt, ...)
+{
+  va_list ap;
+
+  va_start(ap, fmt);
+  vfprintf(stderr, fmt, ap);
+  va_end(ap);
+
+  if (fmt[0] && fmt[strlen(fmt) - 1] == ':') {
+    fputc(' ', stderr);
+    perror(NULL);
+  } else {
+    fputc('\n', stderr);
+  }
+}
+#endif

helpers.h

@@ -0,0 +1,20 @@
+#pragma once
+
+#if defined(ERR) || defined(EXPLAIN)
+void l(const char *fmt, ...);
+#endif
+
+#ifdef ERR
+#define L(...) l(__VA_ARGS__)
+#else
+#define L(...)
+#endif
+
+#ifdef EXPLAIN
+extern int explain;
+#define E(...) if (explain) {    \
+                 l(__VA_ARGS__); \
+               } else
+#else
+#define E(...)
+#endif