New issue 132: Middleware composition regression causes contant 406 Not Acceptable with push_ssl on https://bitbucket.org/conservancy/kallithea/issue/132/middleware-composition-regression-causes
saaj: [This commit](https://bitbucket.org/conservancy/kallithea/commits/6a0964373a3089e5a813389921c3d3e31685f14e) by @kiilerix breaks the composition. I know nothing about Pylons' middleware but through wasting half a day in a WTF mode, I know that this is broken (`HttpsFixup` is never called on VCS opertaions): ``` #!python # Enable https redirects based on HTTP_X_URL_SCHEME set by proxy if any(asbool(config.get(x)) for x in ['https_fixup', 'force_https', 'use_htsts']): app = HttpsFixup(app, config) # we want our low level middleware to get to the request ASAP. We don't # need any pylons stack middleware in them - especially no StatusCodeRedirect buffering app = SimpleHg(app, config) app = SimpleGit(app, config) ``` And this does work (the order it was in RhodeCode): ``` #!python # we want our low level middleware to get to the request ASAP. We don't # need any pylons stack middleware in them - especially no StatusCodeRedirect buffering app = SimpleHg(app, config) app = SimpleGit(app, config) # Enable https redirects based on HTTP_X_URL_SCHEME set by proxy if any(asbool(config.get(x)) for x in ['https_fixup', 'force_https', 'use_htsts']): app = HttpsFixup(app, config) ``` What's really alerting about it is that this regression important in functionality has infiltrated though the test suite. Either tests are not good enough or the commiters don't examine them. _______________________________________________ kallithea-general mailing list [email protected] http://lists.sfconservancy.org/mailman/listinfo/kallithea-general
