Question

I need to determine state of last build (success/failure) and I do it like this:

report_url = 'http://.../ViewLatestBuildReport.aspx' success_marker = '<td class='header-title' colspan='2'>BUILD SUCCESSFUL</td>' page = urllib.urlopen(report_url) if all(success_marker not in line for line in page):   # build is not good, do something   ... 

But this is wasteful (loads entire HTML page), error-prone (I already ran into a bytes/unicode bug) and fragile.

Answer 1

OK, I sort of figured it out. CCTray tracks build status by polling XmlServerReport.aspx, which is (surprise!) XML.

So my current solution looks like this:

import sys, urllib, xml.sax, xml.sax.handler  report_url = 'http://.../CCNET/XmlServerReport.aspx'  class FoundBuildStatus(Exception):   def __init__(self, status):     self.build_status = status  class Handler(xml.sax.handler.ContentHandler):   def startElement(self, name, attrs):     if name == 'Project' and attrs.get('name') == '...':       status = attrs.get('lastBuildStatus')       if status:         raise FoundBuildStatus(status)  page = urllib.urlopen(report_url) try:   xml.sax.parse(page, Handler()) except FoundBuildStatus, ex:   if ex.build_status == 'Failure':     # build is not good, do something     ... 

In my environment it is about 8 times faster than initial HTML-based solution.