bootsnap 1.16.0 → 1.18.4

data/ext/bootsnap/bootsnap.c

@@ -18,8 +18,19 @@
 #include <sys/types.h>
 #include <errno.h>
 #include <fcntl.h>
+#include <unistd.h>
 #include <sys/stat.h>
 
+#ifdef __APPLE__
+  // The symbol is present, however not in the headers
+  // See: https://github.com/Shopify/bootsnap/issues/470
+  extern int fdatasync(int);
+#endif
+
+#ifndef O_NOATIME
+#define O_NOATIME 0
+#endif
+
 /* 1000 is an arbitrary limit; FNV64 plus some slashes brings the cap down to
  * 981 for the cache dir */
 #define MAX_CACHEPATH_SIZE 1000
@@ -30,7 +41,7 @@
 #define MAX_CREATE_TEMPFILE_ATTEMPT 3
 
 #ifndef RB_UNLIKELY
-  #define RB_UNLIKELY(x) (x)
+#define RB_UNLIKELY(x) (x)
 #endif
 
 /*
@@ -54,8 +65,10 @@ struct bs_cache_key {
   uint32_t ruby_revision;
   uint64_t size;
   uint64_t mtime;
-  uint64_t data_size; /* not used for equality */
-  uint8_t  pad[24];
+  uint64_t data_size; //
+  uint64_t digest;
+  uint8_t digest_set;
+  uint8_t pad[15];
 } __attribute__((packed));
 
 /*
@@ -69,7 +82,7 @@ struct bs_cache_key {
 STATIC_ASSERT(sizeof(struct bs_cache_key) == KEY_SIZE);
 
 /* Effectively a schema version. Bumping invalidates all previous caches */
-static const uint32_t current_version = 4;
+static const uint32_t current_version = 6;
 
 /* hash of e.g. "x86_64-darwin17", invalidating when ruby is recompiled on a
  * new OS ABI, etc. */
@@ -87,25 +100,36 @@ static VALUE rb_mBootsnap_CompileCache;
 static VALUE rb_mBootsnap_CompileCache_Native;
 static VALUE rb_cBootsnap_CompileCache_UNCOMPILABLE;
 static ID instrumentation_method;
-static VALUE sym_miss;
-static VALUE sym_stale;
+static VALUE sym_hit, sym_miss, sym_stale, sym_revalidated;
 static bool instrumentation_enabled = false;
 static bool readonly = false;
+static bool revalidation = false;
+static bool perm_issue = false;
 
 /* Functions exposed as module functions on Bootsnap::CompileCache::Native */
 static VALUE bs_instrumentation_enabled_set(VALUE self, VALUE enabled);
 static VALUE bs_readonly_set(VALUE self, VALUE enabled);
+static VALUE bs_revalidation_set(VALUE self, VALUE enabled);
 static VALUE bs_compile_option_crc32_set(VALUE self, VALUE crc32_v);
 static VALUE bs_rb_fetch(VALUE self, VALUE cachedir_v, VALUE path_v, VALUE handler, VALUE args);
 static VALUE bs_rb_precompile(VALUE self, VALUE cachedir_v, VALUE path_v, VALUE handler);
 
 /* Helpers */
+enum cache_status {
+  miss,
+  hit,
+  stale,
+};
 static void bs_cache_path(const char * cachedir, const VALUE path, char (* cache_path)[MAX_CACHEPATH_SIZE]);
 static int bs_read_key(int fd, struct bs_cache_key * key);
-static int cache_key_equal(struct bs_cache_key * k1, struct bs_cache_key * k2);
+static enum cache_status cache_key_equal_fast_path(struct bs_cache_key * k1, struct bs_cache_key * k2);
+static int cache_key_equal_slow_path(struct bs_cache_key * current_key, struct bs_cache_key * cached_key, const VALUE input_data);
+static int update_cache_key(struct bs_cache_key *current_key, struct bs_cache_key *old_key, int cache_fd, const char ** errno_provenance);
+
+static void bs_cache_key_digest(struct bs_cache_key * key, const VALUE input_data);
 static VALUE bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args);
 static VALUE bs_precompile(char * path, VALUE path_v, char * cache_path, VALUE handler);
-static int open_current_file(char * path, struct bs_cache_key * key, const char ** errno_provenance);
+static int open_current_file(const char * path, struct bs_cache_key * key, const char ** errno_provenance);
 static int fetch_cached_data(int fd, ssize_t data_size, VALUE handler, VALUE args, VALUE * output_data, int * exception_tag, const char ** errno_provenance);
 static uint32_t get_ruby_revision(void);
 static uint32_t get_ruby_platform(void);
@@ -122,15 +146,6 @@ struct s2o_data;
 struct i2o_data;
 struct i2s_data;
 
-/* https://bugs.ruby-lang.org/issues/13667 */
-extern VALUE rb_get_coverages(void);
-static VALUE
-bs_rb_coverage_running(VALUE self)
-{
-  VALUE cov = rb_get_coverages();
-  return RTEST(cov) ? Qtrue : Qfalse;
-}
-
 static VALUE
 bs_rb_get_path(VALUE self, VALUE fname)
 {
@@ -161,15 +176,14 @@ Init_bootsnap(void)
 
   instrumentation_method = rb_intern("_instrument");
 
+  sym_hit = ID2SYM(rb_intern("hit"));
   sym_miss = ID2SYM(rb_intern("miss"));
-  rb_global_variable(&sym_miss);
-
   sym_stale = ID2SYM(rb_intern("stale"));
-  rb_global_variable(&sym_stale);
+  sym_revalidated = ID2SYM(rb_intern("revalidated"));
 
   rb_define_module_function(rb_mBootsnap, "instrumentation_enabled=", bs_instrumentation_enabled_set, 1);
   rb_define_module_function(rb_mBootsnap_CompileCache_Native, "readonly=", bs_readonly_set, 1);
-  rb_define_module_function(rb_mBootsnap_CompileCache_Native, "coverage_running?", bs_rb_coverage_running, 0);
+  rb_define_module_function(rb_mBootsnap_CompileCache_Native, "revalidation=", bs_revalidation_set, 1);
   rb_define_module_function(rb_mBootsnap_CompileCache_Native, "fetch", bs_rb_fetch, 4);
   rb_define_module_function(rb_mBootsnap_CompileCache_Native, "precompile", bs_rb_precompile, 3);
   rb_define_module_function(rb_mBootsnap_CompileCache_Native, "compile_option_crc32=", bs_compile_option_crc32_set, 1);
@@ -185,6 +199,14 @@ bs_instrumentation_enabled_set(VALUE self, VALUE enabled)
   return enabled;
 }
 
+static inline void
+bs_instrumentation(VALUE event, VALUE path)
+{
+    if (RB_UNLIKELY(instrumentation_enabled)) {
+       rb_funcall(rb_mBootsnap, instrumentation_method, 2, event, path);
+    }
+}
+
 static VALUE
 bs_readonly_set(VALUE self, VALUE enabled)
 {
@@ -192,6 +214,13 @@ bs_readonly_set(VALUE self, VALUE enabled)
   return enabled;
 }
 
+static VALUE
+bs_revalidation_set(VALUE self, VALUE enabled)
+{
+  revalidation = RTEST(enabled);
+  return enabled;
+}
+
 /*
  * Bootsnap's ruby code registers a hook that notifies us via this function
  * when compile_option changes. These changes invalidate all existing caches.
@@ -290,17 +319,59 @@ bs_cache_path(const char * cachedir, const VALUE path, char (* cache_path)[MAX_C
  * The data_size member is not compared, as it serves more of a "header"
  * function.
  */
-static int
-cache_key_equal(struct bs_cache_key * k1, struct bs_cache_key * k2)
+static enum cache_status cache_key_equal_fast_path(struct bs_cache_key *k1,
+                                     struct bs_cache_key *k2) {
+  if (k1->version == k2->version &&
+          k1->ruby_platform == k2->ruby_platform &&
+          k1->compile_option == k2->compile_option &&
+          k1->ruby_revision == k2->ruby_revision && k1->size == k2->size) {
+      if (k1->mtime == k2->mtime) {
+        return hit;
+      }
+      if (revalidation) {
+        return stale;
+      }
+  }
+  return miss;
+}
+
+static int cache_key_equal_slow_path(struct bs_cache_key *current_key,
+                                     struct bs_cache_key *cached_key,
+                                     const VALUE input_data)
 {
-  return (
-    k1->version        == k2->version        &&
-    k1->ruby_platform  == k2->ruby_platform  &&
-    k1->compile_option == k2->compile_option &&
-    k1->ruby_revision  == k2->ruby_revision  &&
-    k1->size           == k2->size           &&
-    k1->mtime          == k2->mtime
-  );
+  bs_cache_key_digest(current_key, input_data);
+  return current_key->digest == cached_key->digest;
+}
+
+static int update_cache_key(struct bs_cache_key *current_key, struct bs_cache_key *old_key, int cache_fd, const char ** errno_provenance)
+{
+  old_key->mtime = current_key->mtime;
+  lseek(cache_fd, 0, SEEK_SET);
+  ssize_t nwrite = write(cache_fd, old_key, KEY_SIZE);
+  if (nwrite < 0) {
+      *errno_provenance = "update_cache_key:write";
+      return -1;
+  }
+
+#ifdef HAVE_FDATASYNC
+  if (fdatasync(cache_fd) < 0) {
+      *errno_provenance = "update_cache_key:fdatasync";
+      return -1;
+  }
+#endif
+
+  return 0;
+}
+
+/*
+ * Fills the cache key digest.
+ */
+static void bs_cache_key_digest(struct bs_cache_key *key,
+                                const VALUE input_data) {
+  if (key->digest_set)
+    return;
+  key->digest = fnv1a_64(input_data);
+  key->digest_set = 1;
 }
 
 /*
@@ -356,17 +427,34 @@ bs_rb_precompile(VALUE self, VALUE cachedir_v, VALUE path_v, VALUE handler)
 
   return bs_precompile(path, path_v, cache_path, handler);
 }
+
+static int bs_open_noatime(const char *path, int flags) {
+  int fd = 1;
+  if (!perm_issue) {
+    fd = open(path, flags | O_NOATIME);
+    if (fd < 0 && errno == EPERM) {
+      errno = 0;
+      perm_issue = true;
+    }
+  }
+
+  if (perm_issue) {
+    fd = open(path, flags);
+  }
+  return fd;
+}
+
 /*
  * Open the file we want to load/cache and generate a cache key for it if it
  * was loaded.
  */
 static int
-open_current_file(char * path, struct bs_cache_key * key, const char ** errno_provenance)
+open_current_file(const char * path, struct bs_cache_key * key, const char ** errno_provenance)
 {
   struct stat statbuf;
   int fd;
 
-  fd = open(path, O_RDONLY);
+  fd = bs_open_noatime(path, O_RDONLY);
   if (fd < 0) {
     *errno_provenance = "bs_fetch:open_current_file:open";
     return fd;
@@ -377,7 +465,9 @@ open_current_file(char * path, struct bs_cache_key * key, const char ** errno_pr
 
   if (fstat(fd, &statbuf) < 0) {
     *errno_provenance = "bs_fetch:open_current_file:fstat";
+    int previous_errno = errno;
     close(fd);
+    errno = previous_errno;
     return -1;
   }
 
@@ -387,6 +477,7 @@ open_current_file(char * path, struct bs_cache_key * key, const char ** errno_pr
   key->ruby_revision  = current_ruby_revision;
   key->size           = (uint64_t)statbuf.st_size;
   key->mtime          = (uint64_t)statbuf.st_mtime;
+  key->digest_set     = false;
 
   return fd;
 }
@@ -430,7 +521,12 @@ open_cache_file(const char * path, struct bs_cache_key * key, const char ** errn
 {
   int fd, res;
 
-  fd = open(path, O_RDONLY);
+  if (readonly || !revalidation) {
+    fd = bs_open_noatime(path, O_RDONLY);
+  } else {
+    fd = bs_open_noatime(path, O_RDWR);
+  }
+
   if (fd < 0) {
     *errno_provenance = "bs_fetch:open_cache_file:open";
     return CACHE_MISS;
@@ -467,7 +563,6 @@ open_cache_file(const char * path, struct bs_cache_key * key, const char ** errn
 static int
 fetch_cached_data(int fd, ssize_t data_size, VALUE handler, VALUE args, VALUE * output_data, int * exception_tag, const char ** errno_provenance)
 {
-  char * data = NULL;
   ssize_t nread;
   int ret;
 
@@ -479,8 +574,8 @@ fetch_cached_data(int fd, ssize_t data_size, VALUE handler, VALUE args, VALUE *
     ret = ERROR_WITH_ERRNO;
     goto done;
   }
-  data = ALLOC_N(char, data_size);
-  nread = read(fd, data, data_size);
+  storage_data = rb_str_buf_new(data_size);
+  nread = read(fd, RSTRING_PTR(storage_data), data_size);
   if (nread < 0) {
     *errno_provenance = "bs_fetch:fetch_cached_data:read";
     ret = ERROR_WITH_ERRNO;
@@ -491,7 +586,7 @@ fetch_cached_data(int fd, ssize_t data_size, VALUE handler, VALUE args, VALUE *
     goto done;
   }
 
-  storage_data = rb_str_new(data, data_size);
+  rb_str_set_len(storage_data, nread);
 
   *exception_tag = bs_storage_to_output(handler, args, storage_data, output_data);
   if (*output_data == rb_cBootsnap_CompileCache_UNCOMPILABLE) {
@@ -500,7 +595,6 @@ fetch_cached_data(int fd, ssize_t data_size, VALUE handler, VALUE args, VALUE *
   }
   ret = 0;
 done:
-  if (data != NULL) xfree(data);
   return ret;
 }
 
@@ -607,17 +701,22 @@ atomic_write_cache_file(char * path, struct bs_cache_key * key, VALUE data, cons
 
 
 /* Read contents from an fd, whose contents are asserted to be +size+ bytes
- * long, into a buffer */
-static ssize_t
-bs_read_contents(int fd, size_t size, char ** contents, const char ** errno_provenance)
+ * long, returning a Ruby string on success and Qfalse on failure */
+static VALUE
+bs_read_contents(int fd, size_t size, const char ** errno_provenance)
 {
+  VALUE contents;
   ssize_t nread;
-  *contents = ALLOC_N(char, size);
-  nread = read(fd, *contents, size);
+  contents = rb_str_buf_new(size);
+  nread = read(fd, RSTRING_PTR(contents), size);
+
   if (nread < 0) {
     *errno_provenance = "bs_fetch:bs_read_contents:read";
+    return Qfalse;
+  } else {
+    rb_str_set_len(contents, nread);
+    return contents;
   }
-  return nread;
 }
 
 /*
@@ -668,38 +767,67 @@ static VALUE
 bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args)
 {
   struct bs_cache_key cached_key, current_key;
-  char * contents = NULL;
   int cache_fd = -1, current_fd = -1;
   int res, valid_cache = 0, exception_tag = 0;
   const char * errno_provenance = NULL;
 
-  VALUE input_data;   /* data read from source file, e.g. YAML or ruby source */
+  VALUE status = Qfalse;
+  VALUE input_data = Qfalse;   /* data read from source file, e.g. YAML or ruby source */
   VALUE storage_data; /* compiled data, e.g. msgpack / binary iseq */
   VALUE output_data;  /* return data, e.g. ruby hash or loaded iseq */
 
   VALUE exception; /* ruby exception object to raise instead of returning */
+  VALUE exception_message; /* ruby exception string to use instead of errno_provenance */
 
   /* Open the source file and generate a cache key for it */
   current_fd = open_current_file(path, &current_key, &errno_provenance);
-  if (current_fd < 0) goto fail_errno;
+  if (current_fd < 0) {
+    exception_message = path_v;
+    goto fail_errno;
+  }
 
   /* Open the cache key if it exists, and read its cache key in */
   cache_fd = open_cache_file(cache_path, &cached_key, &errno_provenance);
   if (cache_fd == CACHE_MISS || cache_fd == CACHE_STALE) {
     /* This is ok: valid_cache remains false, we re-populate it. */
-    if (RB_UNLIKELY(instrumentation_enabled)) {
-      rb_funcall(rb_mBootsnap, instrumentation_method, 2, cache_fd == CACHE_MISS ? sym_miss : sym_stale, path_v);
-    }
+    bs_instrumentation(cache_fd == CACHE_MISS ? sym_miss : sym_stale, path_v);
   } else if (cache_fd < 0) {
+    exception_message = rb_str_new_cstr(cache_path);
     goto fail_errno;
   } else {
     /* True if the cache existed and no invalidating changes have occurred since
      * it was generated. */
-    valid_cache = cache_key_equal(&current_key, &cached_key);
-    if (RB_UNLIKELY(instrumentation_enabled)) {
-      if (!valid_cache) {
-        rb_funcall(rb_mBootsnap, instrumentation_method, 2, sym_stale, path_v);
+
+    switch(cache_key_equal_fast_path(&current_key, &cached_key)) {
+    case hit:
+      status = sym_hit;
+      valid_cache = true;
+      break;
+    case miss:
+      valid_cache = false;
+      break;
+    case stale:
+      valid_cache = false;
+      if ((input_data = bs_read_contents(current_fd, current_key.size,
+                                     &errno_provenance)) == Qfalse) {
+        exception_message = path_v;
+        goto fail_errno;
       }
+      valid_cache = cache_key_equal_slow_path(&current_key, &cached_key, input_data);
+      if (valid_cache) {
+        if (!readonly) {
+          if (update_cache_key(&current_key, &cached_key, cache_fd, &errno_provenance)) {
+              exception_message = path_v;
+              goto fail_errno;
+          }
+        }
+        status = sym_revalidated;
+      }
+      break;
+    };
+
+    if (!valid_cache) {
+      status = sym_stale;
     }
   }
 
@@ -713,13 +841,18 @@ bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args
     else if (res == CACHE_UNCOMPILABLE) {
       /* If fetch_cached_data returned `Uncompilable` we fallback to `input_to_output`
         This happens if we have say, an unsafe YAML cache, but try to load it in safe mode */
-      if (bs_read_contents(current_fd, current_key.size, &contents, &errno_provenance) < 0) goto fail_errno;
-      input_data = rb_str_new(contents, current_key.size);
+      if (input_data == Qfalse && (input_data = bs_read_contents(current_fd, current_key.size, &errno_provenance)) == Qfalse) {
+        exception_message = path_v;
+        goto fail_errno;
+      }
       bs_input_to_output(handler, args, input_data, &output_data, &exception_tag);
       if (exception_tag != 0) goto raise;
       goto succeed;
     } else if (res == CACHE_MISS || res == CACHE_STALE) valid_cache = 0;
-    else if (res == ERROR_WITH_ERRNO) goto fail_errno;
+    else if (res == ERROR_WITH_ERRNO){
+      exception_message = rb_str_new_cstr(cache_path);
+      goto fail_errno;
+    }
     else if (!NIL_P(output_data)) goto succeed; /* fast-path, goal */
   }
   close(cache_fd);
@@ -727,8 +860,10 @@ bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args
   /* Cache is stale, invalid, or missing. Regenerate and write it out. */
 
   /* Read the contents of the source file into a buffer */
-  if (bs_read_contents(current_fd, current_key.size, &contents, &errno_provenance) < 0) goto fail_errno;
-  input_data = rb_str_new(contents, current_key.size);
+  if (input_data == Qfalse && (input_data = bs_read_contents(current_fd, current_key.size, &errno_provenance)) == Qfalse) {
+    exception_message = path_v;
+    goto fail_errno;
+  }
 
   /* Try to compile the input_data using input_to_storage(input_data) */
   exception_tag = bs_input_to_storage(handler, args, input_data, path_v, &storage_data);
@@ -747,6 +882,7 @@ bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args
    * We do however ignore any failures to persist the cache, as it's better
    * to move along, than to interrupt the process.
    */
+  bs_cache_key_digest(&current_key, input_data);
   atomic_write_cache_file(cache_path, &current_key, storage_data, &errno_provenance);
 
   /* Having written the cache, now convert storage_data to output_data */
@@ -765,6 +901,7 @@ bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args
       * No point raising an error */
       if (errno != ENOENT) {
         errno_provenance = "bs_fetch:unlink";
+        exception_message = rb_str_new_cstr(cache_path);
         goto fail_errno;
       }
     }
@@ -775,16 +912,22 @@ bs_fetch(char * path, VALUE path_v, char * cache_path, VALUE handler, VALUE args
   goto succeed; /* output_data is now the correct return. */
 
 #define CLEANUP \
-  if (contents != NULL) xfree(contents);   \
   if (current_fd >= 0)  close(current_fd); \
-  if (cache_fd >= 0)    close(cache_fd);
+  if (cache_fd >= 0)    close(cache_fd); \
+  if (status != Qfalse) bs_instrumentation(status, path_v);
 
 succeed:
   CLEANUP;
   return output_data;
 fail_errno:
   CLEANUP;
-  exception = rb_syserr_new(errno, errno_provenance);
+  if (errno_provenance) {
+    exception_message = rb_str_concat(
+      rb_str_new_cstr(errno_provenance),
+      rb_str_concat(rb_str_new_cstr(": "), exception_message)
+    );
+  }
+  exception = rb_syserr_new_str(errno, exception_message);
   rb_exc_raise(exception);
   __builtin_unreachable();
 raise:
@@ -802,13 +945,16 @@ invalid_type_storage_data:
 static VALUE
 bs_precompile(char * path, VALUE path_v, char * cache_path, VALUE handler)
 {
+  if (readonly) {
+    return Qfalse;
+  }
+
   struct bs_cache_key cached_key, current_key;
-  char * contents = NULL;
   int cache_fd = -1, current_fd = -1;
   int res, valid_cache = 0, exception_tag = 0;
   const char * errno_provenance = NULL;
 
-  VALUE input_data;   /* data read from source file, e.g. YAML or ruby source */
+  VALUE input_data = Qfalse;   /* data read from source file, e.g. YAML or ruby source */
   VALUE storage_data; /* compiled data, e.g. msgpack / binary iseq */
 
   /* Open the source file and generate a cache key for it */
@@ -824,7 +970,26 @@ bs_precompile(char * path, VALUE path_v, char * cache_path, VALUE handler)
   } else {
     /* True if the cache existed and no invalidating changes have occurred since
      * it was generated. */
-    valid_cache = cache_key_equal(&current_key, &cached_key);
+    switch(cache_key_equal_fast_path(&current_key, &cached_key)) {
+    case hit:
+      valid_cache = true;
+      break;
+    case miss:
+      valid_cache = false;
+      break;
+    case stale:
+      valid_cache = false;
+      if ((input_data = bs_read_contents(current_fd, current_key.size, &errno_provenance)) == Qfalse) {
+        goto fail;
+      }
+      valid_cache = cache_key_equal_slow_path(&current_key, &cached_key, input_data);
+       if (valid_cache) {
+         if (update_cache_key(&current_key, &cached_key, cache_fd, &errno_provenance)) {
+             goto fail;
+         }
+      }
+      break;
+    };
   }
 
   if (valid_cache) {
@@ -836,8 +1001,7 @@ bs_precompile(char * path, VALUE path_v, char * cache_path, VALUE handler)
   /* Cache is stale, invalid, or missing. Regenerate and write it out. */
 
   /* Read the contents of the source file into a buffer */
-  if (bs_read_contents(current_fd, current_key.size, &contents, &errno_provenance) < 0) goto fail;
-  input_data = rb_str_new(contents, current_key.size);
+  if ((input_data = bs_read_contents(current_fd, current_key.size, &errno_provenance)) == Qfalse) goto fail;
 
   /* Try to compile the input_data using input_to_storage(input_data) */
   exception_tag = bs_input_to_storage(handler, Qnil, input_data, path_v, &storage_data);
@@ -852,13 +1016,13 @@ bs_precompile(char * path, VALUE path_v, char * cache_path, VALUE handler)
   if (!RB_TYPE_P(storage_data, T_STRING)) goto fail;
 
   /* Write the cache key and storage_data to the cache directory */
+  bs_cache_key_digest(&current_key, input_data);
   res = atomic_write_cache_file(cache_path, &current_key, storage_data, &errno_provenance);
   if (res < 0) goto fail;
 
   goto succeed;
 
 #define CLEANUP \
-  if (contents != NULL) xfree(contents);   \
   if (current_fd >= 0)  close(current_fd); \
   if (cache_fd >= 0)    close(cache_fd);
 

data/ext/bootsnap/extconf.rb

@@ -1,23 +1,30 @@
 # frozen_string_literal: true
 
-require("mkmf")
+require "mkmf"
 
-if RUBY_ENGINE == "ruby"
-  $CFLAGS << " -O3 "
-  $CFLAGS << " -std=c99"
+if %w[ruby truffleruby].include?(RUBY_ENGINE)
+  have_func "fdatasync", "unistd.h"
+
+  unless RUBY_PLATFORM.match?(/mswin|mingw|cygwin/)
+    append_cppflags ["-D_GNU_SOURCE"] # Needed of O_NOATIME
+  end
+
+  append_cflags ["-O3", "-std=c99"]
 
   # ruby.h has some -Wpedantic fails in some cases
   # (e.g. https://github.com/Shopify/bootsnap/issues/15)
   unless ["0", "", nil].include?(ENV["BOOTSNAP_PEDANTIC"])
-    $CFLAGS << " -Wall"
-    $CFLAGS << " -Werror"
-    $CFLAGS << " -Wextra"
-    $CFLAGS << " -Wpedantic"
+    append_cflags([
+      "-Wall",
+      "-Werror",
+      "-Wextra",
+      "-Wpedantic",
 
-    $CFLAGS << " -Wno-unused-parameter" # VALUE self has to be there but we don't care what it is.
-    $CFLAGS << " -Wno-keyword-macro" # hiding return
-    $CFLAGS << " -Wno-gcc-compat" # ruby.h 2.6.0 on macos 10.14, dunno
-    $CFLAGS << " -Wno-compound-token-split-by-macro"
+      "-Wno-unused-parameter", # VALUE self has to be there but we don't care what it is.
+      "-Wno-keyword-macro", # hiding return
+      "-Wno-gcc-compat", # ruby.h 2.6.0 on macos 10.14, dunno
+      "-Wno-compound-token-split-by-macro",
+    ])
   end
 
   create_makefile("bootsnap/bootsnap")

data/lib/bootsnap/bundler.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 module Bootsnap
-  extend(self)
+  extend self
 
   def bundler?
     return false unless defined?(::Bundler)

data/lib/bootsnap/cli.rb

@@ -36,16 +36,16 @@ module Bootsnap
     end
 
     def precompile_command(*sources)
-      require "bootsnap/compile_cache/iseq"
-      require "bootsnap/compile_cache/yaml"
-      require "bootsnap/compile_cache/json"
+      require "bootsnap/compile_cache"
 
       fix_default_encoding do
-        Bootsnap::CompileCache::ISeq.cache_dir = cache_dir
-        Bootsnap::CompileCache::YAML.init!
-        Bootsnap::CompileCache::YAML.cache_dir = cache_dir
-        Bootsnap::CompileCache::JSON.init!
-        Bootsnap::CompileCache::JSON.cache_dir = cache_dir
+        Bootsnap::CompileCache.setup(
+          cache_dir: cache_dir,
+          iseq: iseq,
+          yaml: yaml,
+          json: json,
+          revalidation: true,
+        )
 
         @work_pool = WorkerPool.create(size: jobs, jobs: {
           ruby: method(:precompile_ruby),
@@ -60,14 +60,16 @@ module Bootsnap
         precompile_json_files(main_sources)
 
         if compile_gemfile
-          # Some gems embed their tests, they're very unlikely to be loaded, so not worth precompiling.
-          gem_exclude = Regexp.union([exclude, "/spec/", "/test/"].compact)
-          precompile_ruby_files($LOAD_PATH.map { |d| File.expand_path(d) }, exclude: gem_exclude)
-
           # Gems that include JSON or YAML files usually don't put them in `lib/`.
           # So we look at the gem root.
+          # Similarly, gems that include Rails engines generally file Ruby files in `app/`.
+          # However some gems embed their tests, they're very unlikely to be loaded, so not worth precompiling.
+          gem_exclude = Regexp.union([exclude, "/spec/", "/test/", "/features/"].compact)
+
           gem_pattern = %r{^#{Regexp.escape(Bundler.bundle_path.to_s)}/?(?:bundler/)?gems/[^/]+}
-          gem_paths = $LOAD_PATH.map { |p| p[gem_pattern] }.compact.uniq
+          gem_paths = $LOAD_PATH.map { |p| p[gem_pattern] || p }.uniq
+
+          precompile_ruby_files(gem_paths, exclude: gem_exclude)
           precompile_yaml_files(gem_paths, exclude: gem_exclude)
           precompile_json_files(gem_paths, exclude: gem_exclude)
         end
@@ -220,6 +222,9 @@ module Bootsnap
 
     def parser
       @parser ||= OptionParser.new do |opts|
+        opts.version = Bootsnap::VERSION
+        opts.program_name = "bootsnap"
+
         opts.banner = "Usage: bootsnap COMMAND [ARGS]"
         opts.separator ""
         opts.separator "GLOBAL OPTIONS"