aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2018-07-31 22:46:21 -0700
committerBryan Newbold <bnewbold@robocracy.org>2018-07-31 22:46:21 -0700
commit7164f10355fcdc45d3be58bfcee2939937e01623 (patch)
tree839a26b217e8690ac2160959457b1b3b0904ae3c
parent4548577d0017611574bc6a8bd30941b74f3267ad (diff)
downloadgeniza-7164f10355fcdc45d3be58bfcee2939937e01623.zip
geniza-7164f10355fcdc45d3be58bfcee2939937e01623.tar.gz
backup more old notes
-rw-r--r--THOUGHTS6
-rw-r--r--notes/new_apis.md17
2 files changed, 23 insertions, 0 deletions
diff --git a/THOUGHTS b/THOUGHTS
new file mode 100644
index 0000000..a90a6b1
--- /dev/null
+++ b/THOUGHTS
@@ -0,0 +1,6 @@
+
+Should probably stick with simple hyperdrives only in synchronizer
+implementation, instead of biting off arbitary feeds-per-peer.
+
+Was intenting to abandon this project, but still sort of thinking of using
+crossbeam-channels or actix to hack something out.
diff --git a/notes/new_apis.md b/notes/new_apis.md
new file mode 100644
index 0000000..9a683ee
--- /dev/null
+++ b/notes/new_apis.md
@@ -0,0 +1,17 @@
+
+## HOWTO: clone drive from scratch
+
+- create empty metadata register (on disk) from key (to ensure we can?)
+- do discovery to find peers
+- from first peer, discover data key
+- create data register
+- create drive from empty registers
+- create synchronizer from all the above
+- ... off it goes
+
+## HOWTO: sync existing drive
+
+- open the drive (on disk)
+- create synchronizer from drive
+- synchronizer does discovery of peers using metadata key (based on config)
+- ... off it goes