Capistrano composer step failure

That end of tunnel light is getting a little closer. I’m using Bitbucket for my repo and working through an example.ca scenario I have my local dev working properly to stand up WordPress. The capistrano piece is works if I comment out the composer line from gemfile and capfile, but that obviously doesn’t make for a successful full deploy. I’m left with a page loading that indicates /releases/20140110141008/wp/wp-blog-header.php) [function.require]: failed to open stream that is understandable b/c it hasn’t attempted to push the dependancies.

I’ve put the cap log from the last step which fails at https://gist.github.com/Idealien/eec9e85c0fead2a45156

I’m running this on MT:GS for hosting and had to do a small tweak to enable composer. But I have confirmed that when I manually login with same account via ssh that composer command responds appropriately.

Let me know if any other file samples would help - The capistrano related file modifications are pretty much stock from the bedrock github save for my credential details.

Thanks,
J

Looks like your Composer problem is the same as this: Deploying WordPress with Capistrano screencast

You’ll need something like:

SSHKit.config.command_map[:composer] = "php -d memory_limit=512M -d allow_url_fopen=1 -d suhosin.executor.include.whitelist=phar composer.phar"

Beautiful! In my case (or anyone else using MT:GS following that same snippet) is:

SSHKit.config.command_map[:composer] = "php -d memory_limit=512M -d allow_url_fopen=1 -d suhosin.executor.include.whitelist=phar ~/composer.phar"

Thanks for this! Works for me…almost. (I’m on GS, as well.)

After the composer command runs, I get an error:

cap aborted! No such file or directory - whoami

However, the current symlink is on the server and the releases directory has the most current release (though the symlink doesn’t appear to be working).

Is this just because I’m running Windows 7, or am I missing something else?

EDIT:
This is a problem with sshkit. I managed to find the fix in this commit. Not sure why 3.1.0 doesn’t have this fix…

However, I’m still getting some issues: my deploy seems to work fine, but I get a 403 error when I visit my URL. The server returns a 500 error when I visit myurl.com/current. Have I mucked up my config somewhere?

Actually, now that I’m looking at it, no symlinks are being followed, both via HTTP and FTP. I can cd through symlinks in SSH, but trying to access any of the /shared or /current symlinks returns an error both via FTP and HTTP.

My .htaccess has a line that says Options +FollowSymLinks…not sure what I’m missing here.