From: Stefan Beller <sbeller@google.com>
Date: Thu, 13 Apr 2017 19:25:28 GMT
Subject: Re: [PATCH 3/2] ls-files: only recurse on active submodules
Message-ID: <CAGZ79kaRTrgHMnP6mA3V2rDHzf8iOidqLTpp2aXmi6x+6YeZxw@mail.gmail.com>
In-Reply-To: <CA+P7+xp=1PUsq1_or=J8ED+-1NMaF=BckC9bK4jqWbA+RFeszw@mail.gmail.com>

On Thu, Apr 13, 2017 at 12:12 PM, Jacob Keller <jacob.keller@gmail.com> wrote:
> On Thu, Apr 13, 2017 at 12:05 PM, Stefan Beller <sbeller@google.com> wrote:
>> On Thu, Apr 13, 2017 at 11:57 AM, Brandon Williams <bmwill@google.com> wrote:
>>> Add in a check to see if a submodule is active before attempting to
>>> recurse.  This prevents 'ls-files' from trying to operate on a submodule
>>> which may not exist in the working directory.
>>
>> What would currently happen on recursing into non-active submodules?
>> Can we have a test for this?
>>
>> Thanks,
>> Stefan
>
> We should be able to test for this. Is it possible that we can recurse
> into a submodule as long as we have the clone in .git/modules/<name>
> even if we don't have it checked out currently?

Conceptually that should be possible, e.g.

    git ls-files --recurse-submodules <ancient ref>

where the ancient ref contained submodules that are not present any more.
In that case we would need to do

    struct strbuf sb = STRBUF_INIT;
    struct child_process = CHILD_PROCESS_INIT;
    struct submodule *sub = submodule_from_path( \
        <path as recorded in ancient tree>, <ancient ref>)
    strbuf_git_path(&sb, "modules/%s", sub->name);

    argv_array_pushl(&cp.args, "git", "ls-files", "--recurse", ...);
    cp.dir = sb.buf;
    run_command(&cp);

Stefan