You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I would like to know more about the process of our bower registry.
I used an Ubuntu server, install Node/NPM, Git SVN and download your NPM. Then, I set the config JSON file similary to your example, with Git cache enabled.
I run the command, all seems fine.
I run an "bower install" on a remote desktop, the depencies are cloned and put into the repository JSON file. And I retrieve on my remote desktop the expected dependecies
BUT when I remove on my remote desktop the bower components and make an "obwer cache clean", and then a "bower install", I've got the following error on my server:
private-bower error fatal: destination path '.' already exists and is not an empty directory.
This error seems to occur when "private-bower" try to clone again the repository. The dependecies are still downloaded on the remote desktop.
So my question are:
This is a normal case ?
What is happening on the server ? It will try to update the cloned git repository ? Or this is a snapshot only for the first time ?
Can we force to refresh the git cache (by fetching, remove and clone again ) ?
Many thanks for your feedbacks
Cheers
Julien Roche
The text was updated successfully, but these errors were encountered:
I ran into this as well.
I threw in some debugging, and it looks like bower is hitting the /packages/:name route twice.
I was thinking that this case should be guarded against and if a second call comes in while we are in the process of fetching the repo, it should have to wait until the first request is complete. Pretty much through a mutex up around it.
I'll play with it this weekend and throw up a pull
I managed to get this working on a basic level by putting a mutex up around the /packages/:name route to ensure that it is only trying to clone the repo once. I'm almost thinking of abandoning the git cache in my case as I am hitting #54 on my laptop. The original calls from bower are also timing out as the clone of some of these repos can take a while.
I'll try a bit more to see what I can get working.
Hi,
I would like to know more about the process of our bower registry.
I used an Ubuntu server, install Node/NPM, Git SVN and download your NPM. Then, I set the config JSON file similary to your example, with Git cache enabled.
I run the command, all seems fine.
I run an "bower install" on a remote desktop, the depencies are cloned and put into the repository JSON file. And I retrieve on my remote desktop the expected dependecies
BUT when I remove on my remote desktop the bower components and make an "obwer cache clean", and then a "bower install", I've got the following error on my server:
private-bower error fatal: destination path '.' already exists and is not an empty directory.
This error seems to occur when "private-bower" try to clone again the repository. The dependecies are still downloaded on the remote desktop.
So my question are:
Many thanks for your feedbacks
Cheers
Julien Roche
The text was updated successfully, but these errors were encountered: