Async getStatus?

[demian85]

getStatus is really slow for big repositories and freezes my app UI (node-webkit) during the execution of that method.
I think the solution is an async version to avoid freezing the main UI/DOM thread when performing computations.
Maybe I could use workers, but why is that method taking so much time? 2+ seconds...

[kevinsawicki]

Are you calling it with a path or with no arguments?

If you are calling it without a path then it is probably best to run it in a separate child process since it can take some time for large repositories.

Adding an async version would work too but that isn't supported yet.

[demian85]

I'm calling that method with no arguments. But still, I don't understand
why is it so slow.
I need to get the initial status when the app is loaded.
Executing 'git status' is way too fast compared to it!

On Thu, Jan 9, 2014 at 1:47 PM, Kevin Sawicki notifications@github.comwrote:

Are you calling it with a path or with no arguments?

If you are calling it without a path then it is probably best to run it in
a separate child process since it can take some time for large repositories.

Adding an async version would work too but that isn't supported yet.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/16#issuecomment-31951695
.

[kevinsawicki]

Hmm, yeah, not sure why it would be slower than git status at the command line, worth looking into.

Is the repo you are trying it out on public?

[demian85]

mm no sorry, it's a private intranet web app, but i can tell you it has at
least three thousand files inside... :P

On Thu, Jan 9, 2014 at 1:59 PM, Kevin Sawicki notifications@github.comwrote:

Hmm, yeah, not sure why it would be slower than git status at the command
line, worth looking into.

If the repo you are trying it out on public?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/16#issuecomment-31953039
.

[demian85]

Could take a look at this? Or should I report the issue in the libgit repo?

[kevinsawicki]

I'll take a look, but it make take me some time to find a repo that reproduces the behavior you are seeing.

[demian85]

I think you just need to be sure nothing weird is going on inside that function.
It seems that the whole directory contents are scanned recursively every time the function is called, and just to get status information. Shouldn't that be cached somehow? How does Git do it?
A temp solution for me could be to just parse git statusoutput :/

[demian85]

I've just implemented my own getStatus parsing git status --porcelain -z output and realized there is missing info in your method that could be useful: unmerged paths.

[billybonks]

@demian85 do you know if this is still an issue?