]> git.ipfire.org Git - thirdparty/git.git/commitdiff
trace2: avoid "futile conditional"
authorJohannes Schindelin <johannes.schindelin@gmx.de>
Thu, 15 May 2025 13:11:46 +0000 (13:11 +0000)
committerJunio C Hamano <gitster@pobox.com>
Thu, 15 May 2025 20:46:47 +0000 (13:46 -0700)
CodeQL reports empty `if` blocks that only contain a comment as "futile
conditional". The comment talks about potential plans to turn this into
a warning, but that seems not to have been necessary. Replace the entire
construct with a concise comment.

Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
trace2/tr2_tmr.c

index 51f564b07a4091a24db71b3a756b4730173cae98..038181ad9be05b4aa7e2cf4930643b699617b15c 100644 (file)
@@ -102,25 +102,11 @@ void tr2_update_final_timers(void)
                struct tr2_timer *t_final = &final_timer_block.timer[tid];
                struct tr2_timer *t = &ctx->timer_block.timer[tid];
 
-               if (t->recursion_count) {
-                       /*
-                        * The current thread is exiting with
-                        * timer[tid] still running.
-                        *
-                        * Technically, this is a bug, but I'm going
-                        * to ignore it.
-                        *
-                        * I don't think it is worth calling die()
-                        * for.  I don't think it is worth killing the
-                        * process for this bookkeeping error.  We
-                        * might want to call warning(), but I'm going
-                        * to wait on that.
-                        *
-                        * The downside here is that total_ns won't
-                        * include the current open interval (now -
-                        * start_ns).  I can live with that.
-                        */
-               }
+               /*
+                * `t->recursion_count` could technically be non-zero, which
+                * would constitute a bug. Reporting the bug would potentially
+                * cause an infinite recursion, though, so let's ignore it.
+                */
 
                if (!t->interval_count)
                        continue; /* this timer was not used by this thread */