Re: [PATCH] commit-graph: fix sparse 'NULL pointer' warning
- From
Ramsay Jones <ramsay@ramsayjones.plus.com>
- Date
- Mar 25, 2019, 17:03 UTC
- Message-ID
- <755f0706-9738-fc10-23a3-e1797c6925dd@ramsayjones.plus.com>
- In-Reply-To
- <871s2vcgb1.fsf@evledraar.gmail.com>
On 25/03/2019 12:02, Ævar Arnfjörð Bjarmason wrote:
Show 15 quoted lines
>
> On Sat, Mar 23 2019, Ramsay Jones wrote:
>
>> Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
>> ---
>>
>> Hi Ævar,
>>
>> If you need to re-roll your 'ab/commit-graph-fixes' branch, could you
>> please squash this into the relevant patch (commit aeb244adbc
>> ("commit-graph: don't early exit(1) on e.g. \"git status\"", 2019-02-21).
>
> Thanks. Will squash & re-submit. It's still just in "pu". Is there a
> compiler that warns about it? Didn't on clang/gcc, but then again it's
> just-as-valid-C & just a style issue, so compilers don't care...This is a 'sparse' warning (see subject line). :-D
I run the 'sparse' Makefile target on each branch and save to a file, sp-out, nsp-out and psp-out, then diff from branch to branch. So, currently, the diff between the 'next' and 'pu' branches look like:
$ diff nsp-out psp-out 25a26 > commit-graph.c:103:24: warning: Using plain integer as NULL pointer 123a125 > SP promisor-remote.c 188a191 > upload-pack.c:1167:56: warning: Using plain integer as NULL pointer 289d291 < SP builtin/rebase--interactive.c $
Alternatively, I can simply run 'sparse' over a single file:
$ make commit-graph.sp
SP commit-graph.c
commit-graph.c:103:24: warning: Using plain integer as NULL pointer
$ Of course, you need to have 'sparse' installed ... ;-)
Show 12 quoted lines
>
>> This same commit (aeb244adbc) also removes the last call, outside of the
>> commit-graph.c file, to the function load_commit_graph_one(). So, this
>> function is now a file-local symbol and could be marked 'static'.
>>
>> Also, the function verify_commit_graph_lite(), introduced in commit
>> d8acf37ff7 ("commit-graph: fix segfault on e.g. \"git status\"",
>> 2019-02-21), currently has no external callers. This function is also
>> a file-local symbol and could be marked 'static', unless you have plans
>> for future external calls?
>
> Fixing these too. Just missed them. Thanks.And these resulted from running my static-check.pl script.
ATB, Ramsay Jones