what we learned from the iOS testers so far...

[Dirk Hohndel]

> On Mar 25, 2016, at 9:49 AM, Robert Helling <helling at atdotde.de> wrote:
> 
> Hi,
> 
>> On 25.03.2016, at 17:36, Dirk Hohndel <dirk at hohndel.org <mailto:dirk at hohndel.org>> wrote:
>> 
>> So you are experiencing the same things many people (Miika most vocally) have complained about. libgit2 does NOT appear to be as smart and fast when it comes to low-bandwidth situation as the git client. One of the key reasons why I picked git as the storage protocol was that the protocol itself tries to be super clever. This needs analysis by an expert. Thankfully I am taking an expert to a place that promises mediocre internet speed (Cozumel) next week, so there is hope that we can get some insights here...
> 
> 
> Too bad I cannot run a packet sniffer on the DSL router here. So far, it seems, it is not slow but never finishes. To the degree that it could be completely stuck as well.
> 
> My repository on my desktop has 90MB, almost all in .git/objects/packed. My guess is that this contained images at some point. Is there an easy way to get rid of those? As long as I keep the current state, I am happy to give up the history. Git experts to the rescue!

The always brilliant Miika posted this a couple of times already:

yep, the the photos are still stored on the git repository, even
though you cannot see them directly in Subsurface. And therefor the
size of your cloud data is massive when comparing to what it could be.
I resolved my issue by forcing new (empty) git history into the cloud
with following steps:

Using star in the commands assumes that there is only one directory
under cloudstorage, which should be the normal case.

$ cd ~/.subsurface/cloudstorage/*
$ git checkout --orphan force
$ git add -A
$ git commit
$ git branch  -D miika.turkia at gmail.com <mailto:miika.turkia at gmail.com>
$ git branch -m miika.turkia at gmail.com <mailto:miika.turkia at gmail.com>
$ git push -u --force origin miika.turkia at gmail.com <mailto:miika.turkia at gmail.com>
$ cd .. && rm -rf *

* And finally open up the cloud storage on Subsurface desktop application.


The above assumes you are on Linux machine, but the git steps should
be the same on other OSes. You basically create an orphaned branch
with only one git commit that contains your current log book. Then you
replace the master branch with this clean branch, and force push it
all to the cloud storage overwriting the old data. This cleaned a few
hundred megs from my repository allowing fast syncs after these steps.
Obviously you will need to tune some info on the commands and
hopefully have some understanding of what you are doing, if you want
to clean the cloud storage with this method.

I think we should have this as an FAQ :-)

Or create a little script that uses the verbose output of Subsurface to find the right directory and takes the account name as an argument or something... that would be awesome.

/D
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20160325/634ab2c0/attachment.html>