Re: [PATCH 3/3] clean: improve performance when removing lots of directories

On Tue, Apr 7, 2015 at 12:10 AM, Eric Sunshine [off-list ref] wrote:

On Mon, Apr 6, 2015 at 7:48 AM, Erik Elfström [off-list ref] wrote:

quoted

Before this change, clean used resolve_gitlink_ref to check for the
presence of nested git repositories. This had the drawback of creating
a ref_cache entry for every directory that should potentially be
cleaned. The linear search through the ref_cache list caused a massive
performance hit for large number of directories.

Teach clean.c:remove_dirs to use setup.c:is_git_directory
instead. is_git_directory will actually open HEAD and parse the HEAD
ref but this implies a nested git repository and should be rare when
cleaning.

Using is_git_directory should give a more standardized check for what
is and what isn't a git repository but also gives a slight behavioral
change. We will now detect and respect bare and empty nested git
repositories (only init run). Update t7300 to reflect this.

The time to clean an untracked directory containing 100000 sub
directories went from 61s to 1.7s after this change.

Impressive.

quoted

Signed-off-by: Erik Elfström <redacted>
Helped-by: Jeff King [off-list ref]

It is customary for your sign-off to be last.

More below...

quoted

---

diff --git a/builtin/clean.c b/builtin/clean.c
index 98c103f..e951bd9 100644
--- a/builtin/clean.c
+++ b/builtin/clean.c

@@ -148,6 +147,24 @@ static int exclude_cb(const struct option *opt, const char *arg, int unset)
        return 0;
 }

+static int is_git_repository(struct strbuf *path)
+{
+       int ret = 0;
+       if (is_git_directory(path->buf))
+               ret = 1;
+       else {
+               int orig_path_len = path->len;
+               if (path->buf[orig_path_len - 1] != '/')

Minor: I don't know how others feel about it, but I always find it a
bit disturbing to see a potential negative array access without a
safety check that orig_path_len is not 0, either directly in the
conditional or as a documenting assert().

I think I would prefer to accept empty input and return false rather
than assert. What to you think about:

static int is_git_repository(struct strbuf *path)
{
    int ret = 0;
    size_t orig_path_len = path->len;
    if (orig_path_len == 0)
        ret = 0;
    else if (is_git_directory(path->buf))
        ret = 1;
    else {
        if (path->buf[orig_path_len - 1] != '/')
            strbuf_addch(path, '/');
        strbuf_addstr(path, ".git");
        if (is_git_directory(path->buf))
            ret = 1;
        strbuf_setlen(path, orig_path_len);
    }

    return ret;
}


Also I borrowed this pattern from remove_dirs and it has the same
problem. Should I add something like this as a separate commit?

diff --git a/builtin/clean.c b/builtin/clean.c
index ccffd8a..88850e3 100644
--- a/builtin/clean.c
+++ b/builtin/clean.c

@@ -173,7 +173,8 @@ static int remove_dirs(struct strbuf *path, const

char *prefix, int force_flag,
        DIR *dir;
        struct strbuf quoted = STRBUF_INIT;
        struct dirent *e;
-       int res = 0, ret = 0, gone = 1, original_len = path->len, len;
+       int res = 0, ret = 0, gone = 1;
+       size_t original_len = path->len, len;
        struct string_list dels = STRING_LIST_INIT_DUP;

        *dir_gone = 1;

@@ -201,6 +202,7 @@ static int remove_dirs(struct strbuf *path, const

char *prefix, int force_flag,
                return res;
        }

+       assert(original_len > 0 && "expects non-empty path");
        if (path->buf[original_len - 1] != '/')
                strbuf_addch(path, '/');

quoted

+                       strbuf_addch(path, '/');
+               strbuf_addstr(path, ".git");
+               if (is_git_directory(path->buf))
+                       ret = 1;
+               strbuf_setlen(path, orig_path_len);
+       }
+
+       return ret;
+}
+
 static int remove_dirs(struct strbuf *path, const char *prefix, int force_flag,
                int dry_run, int quiet, int *dir_gone)
 {