aboutsummaryrefslogtreecommitdiff
path: root/mod/mod-ci-github-service-data.hxx
diff options
context:
space:
mode:
Diffstat (limited to 'mod/mod-ci-github-service-data.hxx')
-rw-r--r--mod/mod-ci-github-service-data.hxx12
1 files changed, 10 insertions, 2 deletions
diff --git a/mod/mod-ci-github-service-data.hxx b/mod/mod-ci-github-service-data.hxx
index c9218f3..a6cb742 100644
--- a/mod/mod-ci-github-service-data.hxx
+++ b/mod/mod-ci-github-service-data.hxx
@@ -11,6 +11,8 @@
namespace brep
{
+ // @@@ Check is any data members are unused.
+
// Service data associated with the tenant (corresponds to GH check suite).
//
// It is always a top-level JSON object and the first member is always the
@@ -44,8 +46,7 @@ namespace brep
};
// We have two kinds of service data that correspond to the following two
- // scenarios (those are the only possible ones, until/unless we add support
- // for merge queues):
+ // typical scenarios (until/unless we add support for merge queues):
//
// 1. Branch push (via check_suite) plus zero or more local PRs (via
// pull_request) that share the same head commit id.
@@ -58,6 +59,13 @@ namespace brep
// can be created and is not behind base. We do all this before we actually
// create the CI tenant.
//
+ // Note that the above two cases are typical but not the only possible
+ // scenarios. Specifically, it is possible to have a mixture of all three
+ // kinds (branch push, local PR, and remote PR) since the same head commit
+ // id can be present in both local and remote branches. There is no way to
+ // handle this case perfectly and we do the best we can (see
+ // build_unloaded_pre_check() for details).
+ //
struct service_data
{
// The data schema version. Note: must be first member in the object.