summaryrefslogtreecommitdiff
path: root/TODO.md
diff options
context:
space:
mode:
Diffstat (limited to 'TODO.md')
-rw-r--r--TODO.md52
1 files changed, 29 insertions, 23 deletions
diff --git a/TODO.md b/TODO.md
index 337639e..d1af9d0 100644
--- a/TODO.md
+++ b/TODO.md
@@ -1,23 +1,29 @@
-# refactor core
- * move state from memory to disk, sqlite probably
- * allows us to retain stats etc over restart
- * simplifies code
- * keep only running nodes in memory
- * negligible I/O cost, as majority is I/O wait getting config
-
-# separate login to owon package
- * oxidized-script is not only use-case
- * it would be good to have minimal package used to login to routers
- * oxidized just one consumer of that functionality
- * what to do with models, we need model to know how to login. Should models be separated to another package? oxidized-core, oxidized-models and oxidized-login?
- * how can we allow interactive login in oxidized-login? With functional VTY etc? REPL loop in input/ssh and input/telnet?
-
-# thread number
- * think about algo
- * if job ended later than now-iteration have rand(node.size) == 0 to add thread
- * if now is less than job_ended+iteration same chance to remove thread?
- * should we try to avoid max threads from being hit? (like maybe non-success thread is pulling average?)
-
-# docs, testing
- * yard docs
- * minitest tests
+# To Do
+
+## refactor core
+
+* move state from memory to disk, sqlite probably
+* allows us to retain stats etc over restart
+* simplifies code
+* keep only running nodes in memory
+* negligible I/O cost, as majority is I/O wait getting config
+
+## separate login to own package
+
+* oxidized-script is not only use-case
+* it would be good to have minimal package used to login to routers
+* oxidized just one consumer of that functionality
+* what to do with models, we need model to know how to login. Should models be separated to another package? oxidized-core, oxidized-models and oxidized-login?
+* how can we allow interactive login in oxidized-login? With functional VTY etc? REPL loop in input/ssh and input/telnet?
+
+## thread number
+
+* think about algo
+* if job ended later than now-iteration have rand(node.size) == 0 to add thread
+* if now is less than job_ended+iteration same chance to remove thread?
+* should we try to avoid max threads from being hit? (like maybe non-success thread is pulling average?)
+
+## docs, testing
+
+* yard docs
+* minitest tests