summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAndreas Baumann <abaumann@yahoo.com>2012-04-25 21:44:49 +0200
committerAndreas Baumann <abaumann@yahoo.com>2012-04-25 21:44:49 +0200
commit244cdd5d5a9641f494d0e1edb0737e167e5568f3 (patch)
tree9182d59f5a8b317e5aa4f2b2d0c7e8f956a78ace
parentff33cf46862c7c88ae2250e787dd00f2b946c9e6 (diff)
downloadpgfuse-244cdd5d5a9641f494d0e1edb0737e167e5568f3.tar.gz
pgfuse-244cdd5d5a9641f494d0e1edb0737e167e5568f3.tar.bz2
updated some docu
-rw-r--r--DEVELOPERS7
-rw-r--r--FILELIST1
-rw-r--r--TODO4
-rw-r--r--tests/README4
4 files changed, 10 insertions, 6 deletions
diff --git a/DEVELOPERS b/DEVELOPERS
index dbee0e0..d930ec4 100644
--- a/DEVELOPERS
+++ b/DEVELOPERS
@@ -29,6 +29,9 @@ Coding guidelines
are simple enough. If possible avoid string manipulations as
for timestamps (we are on low-level OS-abstraction layer, so
'struct timespec' and epochs are fine).
+
+ Consistency of the data should be ensured in the database.
+ We don't want to implement an 'fsck' for pgfuse.
Design desicions
----------------
@@ -77,6 +80,10 @@ How to tune the block sizes? What factors influence the experiment?
At the moment we store padded blocks of fixes size (STANDARD_BLOCK_SIZE),
not really sure if that is good or bad.
+The block size should be computed (small files have only one block,
+all other have a a blocksize of the first full block). Gets us more
+independent on some configuration or options.
+
Directory tree in database
--------------------------
diff --git a/FILELIST b/FILELIST
index 1c27c9b..09381f2 100644
--- a/FILELIST
+++ b/FILELIST
@@ -1,5 +1,4 @@
schema.sql - create schema for PgFuse in PostgreSQL database
-clean.sql - remove data and schema from PostgreSQL database
config.h - global limitations of the program
pgfuse.c - main and hooks for FUSE operations
pgsql.c - implementation of PostgreSQL access functions
diff --git a/TODO b/TODO
index 153d0ec..cf544ba 100644
--- a/TODO
+++ b/TODO
@@ -1,7 +1,7 @@
TODO list (in order of priority)
---------
-- rename (after a redo of the dir table)
+- rename
- handling of most file system metadata
- ownership: how is this done depending on
per-user or root mounts? think about security
@@ -13,6 +13,8 @@ TODO list (in order of priority)
- strategy for half-blocks, help PostgreSQL optimize disk usage
of data in BYTEA columns. Try to support tails of growing files
and tiny files (without padding to the block size)
+- make block size detectable per file (there is no reason why
+ different files should not have a different block size)
- allow for size_t and off_t (check int4 which should
be int8, use be64ton and not htonl), flag for 64-bit
(large filesystem)?
diff --git a/tests/README b/tests/README
deleted file mode 100644
index a4cb163..0000000
--- a/tests/README
+++ /dev/null
@@ -1,4 +0,0 @@
-testfsync.c - tests flush, fsync, fdatasync, can't be done eaily
- with file system commands only
-testpgfsql.c - standalone tests of libpq interface (for instance
- how to handle timestamps)