The initial code for bug #27673 accidentally nuked all buildid service
concurrency, not just identical concurrent requests. Correct this
with one-liner patch. Observing the effect in the automated testsuite
is difficult, so hand-tested against large requests and short ones,
run in an interleaved way.
Signed-off-by: Frank Ch. Eigler <fche@redhat.com>
+2021-08-19 Frank Ch. Eigler <fche@redhat.com>
+
+ PR28249
+ * debuginfod.cxx (handler_cb): Fix after_you unique_set key
+ to the entire incoming URL.
+
2021-07-26 Noah Sanci <nsanci@redhat.com>
PR27982
struct timespec tsay_start, tsay_end;
clock_gettime (CLOCK_MONOTONIC, &tsay_start);
static unique_set<string> busy_urls;
- unique_set_reserver<string> after_you(busy_urls, url1);
+ unique_set_reserver<string> after_you(busy_urls, url_copy);
clock_gettime (CLOCK_MONOTONIC, &tsay_end);
afteryou = (tsay_end.tv_sec - tsay_start.tv_sec) + (tsay_end.tv_nsec - tsay_start.tv_nsec)/1.e9;
add_metric ("thread_busy", "role", "http-buildid-after-you", -1);