should use a standard python-feedparser or embed the patched one

[Vincent Ladeuil]

I can't run 'python build.py' locally, it fails with:

fetching data from http://feeds.launchpad.net/bazaar/announcements.atom
Traceback (most recent call last):
  File "build.py", line 121, in <module>
    main()
  File "build.py", line 107, in main
    build_feed("Community News", NEWS_RSS, 4),
  File "build.py", line 63, in build_feed
    d = feedparser.parse(rss_url, extra_headers={'Cache-control': 'max-age=60'})
TypeError: parse() got an unexpected keyword argument 'extra_headers'

The backtrace points to build_feed and there we have:

    # we use our own copy of feedparser with a patch to let
    # <http://code.google.com/p/feedparser/issues/detail?id=224> extra_headers
    # be passed
    d = feedparser.parse(rss_url, extra_headers={'Cache-control': 'max-age=60'})

Nice !

But... can I have this feedparser too ?

I'm on maverick (python-feedparser 4.1-14) and I wonder what is used in production...

[Vincent Ladeuil (vila)]

I worked around the problem with:

=== modified file 'build.py'
--- build.py 2011-02-08 12:20:55 +0000
+++ build.py 2011-02-10 14:23:45 +0000
@@ -60,7 +60,7 @@
     # we use our own copy of feedparser with a patch to let
     # <http://code.google.com/p/feedparser/issues/detail?id=224> extra_headers
     # be passed
- d = feedparser.parse(rss_url, extra_headers={'Cache-control': 'max-age=60'})
+ d = feedparser.parse(rss_url)
     entries = []
     for entry in d.entries[0:count]:
         entries.append(dict(

[Martin Pool (mbp)]

As the google bug says, the code is in <https://code.launchpad.net/~mbp/feedparser/224-headers/>

The easiest thing for you is probably to run from a checkout of that branch, and/or to pass the header value conditionally on checking the feedparser version.

I think including a snapshot of feedparser will just make things worse.

[Martin Pool (mbp)]

I don't think we should actually change the bzr-website scripts for this, other than perhaps to mention where the branch is.

[Роман Донченко (dpb-r)]

The recent versions of feedparser include this functionality out of the box, so the custom branch is no longer needed. The interface is slightly different:

=== modified file 'build.py'
--- build.py 2012-03-21 10:29:44 +0000
+++ build.py 2012-05-13 23:53:34 +0000
@@ -57,10 +57,7 @@

 def build_feed(title, rss_url, count):
     print "fetching data from %s" % (rss_url,)
- # we use our own copy of feedparser with a patch to let
- # <http://code.google.com/p/feedparser/issues/detail?id=224> extra_headers
- # be passed
- d = feedparser.parse(rss_url, extra_headers={'Cache-control': 'max-age=60'})
+ d = feedparser.parse(rss_url, request_headers={'Cache-control': 'max-age=60'})
     entries = []
     for entry in d.entries[0:count]:
         entries.append(dict(

[Richard Wilbur (richard-wilbur)]

Ran into this while working on lp:963742.

Anybody know what triggers a rebuild of the Bazaar home page (index.html)?