From 1fb1bcc9714e86ef1357fbea72243af3b5634bbf Mon Sep 17 00:00:00 2001 From: Rongsong Shen Date: Tue, 29 Sep 2026 08:11:52 +0000 Subject: [PATCH] docs(scheduler): parjob bucket-reset comment describes the job-level .host want after bug-1883 Co-Authored-By: Claude Opus 5.5 (1M context) --- src/scheduler/src/plugins/parjob_allocator.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/scheduler/src/plugins/parjob_allocator.c b/src/scheduler/src/plugins/parjob_allocator.c index c880a4ff..491c38c7 100644 --- a/src/scheduler/src/plugins/parjob_allocator.c +++ b/src/scheduler/src/plugins/parjob_allocator.c @@ -460,7 +460,8 @@ static int parjob_try_plan_(xce_sched_core_t* core, xce_sched_ctx_t* ctx, * see host_iter.c's own file header). Left un-reset, it would * still carry a DIFFERENT bucket's want (or, for the very first * bucket ever tried, the job-level ctx->worker==NULL prepare - * pass's own dead-write into .host) -- harmless for seqjob + * pass's whole-primary slots:maxTasks want, which that pass + * rebuilds per primary since bug-1883) -- harmless for seqjob * (nslots is always 1, so every bucket's want is identical by * construction) but actively wrong for parjob, where bucket size * genuinely varies call to call: a stale want here can wrongly