Scrapydd

- agent daemon mode, agent cmdline help
- server daemon stdout/stderr redirect to /dev/null
- server: project upload non-blocking run.
- server and agent: run project in isolated virtualenv
- fix: item file is not written completely on job complete

- agent logging file
- agent remove local temp files when job complete
- agent local files in workspace folder
- server check node id and job status when committing job complete request

- job history limitation
- webhook use isolated file against the job's original items file, the file will also be cleared when webhook job complete.

- Webhook

- stream upload items file
- async upload items with file stream when task complete
- agent execute process stdout to fd
- agent use a slot container to control concurrency, instead of the queue
- fix: agent remove tmp egg file after requirements test
- heartbeat post running jobs

- items file upload
- log and item file error handling
- fix: spider page display all jobs history
- scrapydd cmdline -v/--version to check version
- double check task_queue get, and also check database consistence
- heartbeat response contains whether there is new task on server info, client do not need pool GETTASK any more
- fix: get_task might cause task status in queue inconsistency with database