aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorFrancois Kritzinger <francois@codesynthesis.com>2024-12-19 16:15:26 +0200
committerFrancois Kritzinger <francois@codesynthesis.com>2024-12-19 16:15:26 +0200
commitcfdf2b202d36d2332a37af51c770054bfb3d49c8 (patch)
tree6d60f710541703098af7fd5f96538486e3d0d4b0
parent62ec3a690f59814d28ff5c4c1e7dd9fd1fab7e7b (diff)
ci-github: Add questions/TODOs
-rw-r--r--INSTALL-GITHUB-DEV4
-rw-r--r--mod/mod-ci-github.cxx10
2 files changed, 14 insertions, 0 deletions
diff --git a/INSTALL-GITHUB-DEV b/INSTALL-GITHUB-DEV
index 5634072..8614979 100644
--- a/INSTALL-GITHUB-DEV
+++ b/INSTALL-GITHUB-DEV
@@ -156,6 +156,10 @@ aspects but can't redeliver webhooks.
- Re-requested check suite.
- Re-requested check run.
- Head shared with BP.
+
+ @@ TMP Does this mean the pull_request is received after the
+ check_suite?
+
- Not meargeable.
- Head behind base.
- Head commit has changed.
diff --git a/mod/mod-ci-github.cxx b/mod/mod-ci-github.cxx
index 773ca5f..9514fb8 100644
--- a/mod/mod-ci-github.cxx
+++ b/mod/mod-ci-github.cxx
@@ -960,6 +960,11 @@ namespace brep
{
if (d->archived) // Tenant is archived
{
+ // @@ TODO Why aren't we updating instead of creating? The node ids
+ // are in the service data (which is still available after
+ // archiving, which I didn't know). And both CRs should already
+ // be built.
+
// Fail (re-create) the check runs.
//
optional<gh_installation_access_token> iat (get_iat ());
@@ -1054,6 +1059,11 @@ namespace brep
// Fail if it's the conclusion check run that is being re-requested.
//
+ // @@ TMP When user selects re-run all failed checks we receive multiple
+ // check_runs, one of which is for the CCR. We then update it with the
+ // error message, triggering another check_suite(completed) right after
+ // all of the check_runs(rerequested).
+ //
if (cr.check_run.name == conclusion_check_run_name)
{
l3 ([&]{trace << "re-requested conclusion check_run";});