devblog
This commit is contained in:
parent
7234b1f9a7
commit
b3c2ae2fc7
1 changed files with 17 additions and 0 deletions
17
doc/devblog/day_596__back_from_summer_vacation.mdwn
Normal file
17
doc/devblog/day_596__back_from_summer_vacation.mdwn
Normal file
|
@ -0,0 +1,17 @@
|
||||||
|
I've been back from summer vacation for a couple of days.
|
||||||
|
My contract to work on git-annex has expired, at least for now,
|
||||||
|
but I have a lot of Patreon rewards to catch up on anyway. I've been
|
||||||
|
pushing hard for months on that contract and made a lot of progress on
|
||||||
|
long-term goals. Plan for the next little while is to cut back a little
|
||||||
|
bit, and work on easier stuff.
|
||||||
|
|
||||||
|
Today I improved how git-annex uses Copy-On-Write when copying between
|
||||||
|
two repositories on the same drive. It had relied on matching up device
|
||||||
|
numbers, but it turns out that with eg BTRFS subvolumes, CoW is supported
|
||||||
|
even when the device numbers don't match. Also, it was using cp even on
|
||||||
|
filesystems that don't support CoW, which prevented resuming after an
|
||||||
|
interruption. The new approach is to try to make a CoW copy once per remote,
|
||||||
|
and if it fails, fall back to rsync.
|
||||||
|
|
||||||
|
Today's work was sponsored by Trenton Cronholm
|
||||||
|
<a href="https://patreon.com/joeyh">on Patreon</a>.
|
Loading…
Add table
Add a link
Reference in a new issue