Re: [PATCH 4/4] bisect--helper: `bisect_reset` shell function in C

On Wed, Jun 8, 2016 at 9:59 AM, Eric Sunshine [off-list ref] wrote:

On Tue, Jun 7, 2016 at 4:54 PM, Pranit Bauva [off-list ref] wrote:

quoted

Reimplement `bisect_reset` shell function in C and add a `--bisect-reset`
subcommand to `git bisect--helper` to call it from git-bisect.sh .

Using `bisect_reset` subcommand is a temporary measure to port shell
functions to C so as to use the existing test suite. As more functions
are ported, this subcommand would be retired and will be called by some
other method.

Note: --bisect-clean-state subcommand has not been retired as there are
still a function namely `bisect_start()` which still uses this
subcommand.

Signed-off-by: Pranit Bauva <redacted>
---

diff --git a/builtin/bisect--helper.c b/builtin/bisect--helper.c

@@ -118,12 +122,51 @@ int bisect_clean_state(void)
+int bisect_reset(const char *commit)

s/^/static/

quoted

+{
+       struct strbuf branch = STRBUF_INIT;
+       int status = 0;
+
+       if (file_size(git_path_bisect_start()) < 1) {

This doesn't even care about the size of the file, only if it
encountered an error while stat()'ing it. Why not just use
file_exists() instead (which you already use elsewhere in this
function)? Alternately, if you're trying to be faithful to the shell
code, then you *do* need to check that the file has non-zero size
before issuing the "not bisecting" diagnostic, so:

    if ()
        printf("... not bisecting ...");

As file_size() returns an integer, there is no difference between
"file_size(git_path_bisect_start()) <= 0" and
"file_size(git_path_bisect_start()) < 1".
Or am I missing something?