[Qa-jenkins-scm] [jenkins.debian.net] 03/03: reproducible: scheduler: add a comment to explain how LIMITS work

[Holger Levsen]

This is an automated email from the git hooks/post-receive script.

holger pushed a commit to branch master
in repository jenkins.debian.net.

commit a6eada0f94cb015c4c57cd3dd117656fdb5d5793
Author: Mattia Rizzolo <mattia at mapreri.org>
Date:   Sat Sep 12 11:31:09 2015 +0000

    reproducible: scheduler: add a comment to explain how LIMITS work
---
 bin/reproducible_scheduler.py | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/bin/reproducible_scheduler.py b/bin/reproducible_scheduler.py
index 2d0e670..26f6783 100755
--- a/bin/reproducible_scheduler.py
+++ b/bin/reproducible_scheduler.py
@@ -26,7 +26,22 @@ from reproducible_html_live_status import generate_schedule
 from reproducible_html_packages import gen_packages_html
 from reproducible_html_packages import purge_old_pages
 
-# FIXME: these limits were understandable when the code was merged but now need some explaination
+"""
+How the scheduler chose which limit to apply:
+everything depends on how many packages are already scheduled, in a 3 steps
+process.  Let's go by an example:
+    'unstable': {1: (250, 40), 2: (350, 20), '*': 5},
+if total < 250:
+    40
+elif total < 350:
+    20
+else:
+    5
+ * 1st step, if there are less than 250 packages in the queue, schedule 40
+ * 2nd step, if there are less than 350 packages in the queue, schedule 20
+ * 3rd step, schedule 5
+So, the 3rd step happens only when there are more than 350 packages queued up.
+"""
 LIMITS = {
     'untested': {
         'amd64': {

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/qa/jenkins.debian.net.git