Renovate scrapy upstart job a bit

[CameronNemo]

No need to redirect stdout/err, upstart will catch that. Use chdir stanza to keep up to date with init script, do not source empty/useless defaults file.

[pablohoffman]

Does upstart provide an alternative to the defaults file?. It is common to pre-populate some environment vars there.

[CameronNemo]

You can use override files for that. All you need to do is make a file scrapyd.override in the same directory (/etc/init) and add lines like env VAR1=val1. You can also put stuff like limit nofile 1024 4096 where 1024 is the soft resource limit for the number of file descripters, and 4096 is the hard limit. In the latest release of Upstart, you can actually add cgroup constraints too.

By environment vars, I take it scrapy reads them itself, right? A default file is usually used to configure how the init script operates, not how the daemon operates. So you would have vars like DAEMON_ARGS that the init script would append to the daemon's arguments, but that variable would not be exported to the daemon. That is the usual use case.

[dangra]

You can use override files for that. All you need to do is make a file scrapyd.override in the same directory (/etc/init)

interesting tip! is this possible in the version of upstart shipped with ubuntu lucid?

[CameronNemo]

No, it was added in Upstart 0.9 (0.6.5 is in lucid). Sorry, I did not think about older releases. I have added back the default file sourcing.

[jayzeng]

LGTM, @pablohoffman good to merge?

[dangra]

LGTM.