af7e5983e5
NeonSession::NeonAuth requires an AuthListener to be set in the RequestEnvironment (RE). After a HTTP 401, the Negotiate challenge should be handled. The whole RE however is reset in HandleError, which is called after every request. Usually, this is not a problem because the RE is re-set at the beginning of almost all requests (PUT, GET, PROPFIND, ...). But when refreshing locks this is not the case, as the LockStore does not know any RequestEnvironment. NeonSession::NeonAuth is then called in a post-send hook after a HTTP 401, the AuthListener is null and in consequence Neon's ah_post_send does not return NE_RETRY. So there will be no retry with the correct authorization header. This patch removes the reset of the RequestEnvironment in HandleError, thus keeping it in the session for further use. Change-Id: Ib1a5eff08072f9ef003e760c5f40d9ad26c6cad1 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/96191 Tested-by: Jenkins Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com> |
||
|---|---|---|
| .. | ||
| qa | ||
| source | ||
| test/com/sun/star/comp/ucb | ||
| CppunitTest_ucb_webdav_local_neon.mk | ||
| CppunitTest_ucb_webdav_neon_opts.mk | ||
| CppunitTest_ucb_webdav_propfindcache.mk | ||
| CppunitTest_ucb_webdav_res_access.mk | ||
| IwyuFilter_ucb.yaml | ||
| JunitTest_ucb_complex.mk | ||
| JunitTest_ucb_unoapi.mk | ||
| Library_cached1.mk | ||
| Library_srtrs1.mk | ||
| Library_ucb1.mk | ||
| Library_ucpcmis1.mk | ||
| Library_ucpdav1.mk | ||
| Library_ucpexpand1.mk | ||
| Library_ucpext.mk | ||
| Library_ucpfile1.mk | ||
| Library_ucpftp1.mk | ||
| Library_ucpgio1.mk | ||
| Library_ucphier1.mk | ||
| Library_ucpimage.mk | ||
| Library_ucppkg1.mk | ||
| Library_ucptdoc1.mk | ||
| Makefile | ||
| Module_ucb.mk | ||
| README | ||
Universal Content Broker (has ucp) which do things like convert files to strings in content broker world. mmeeks: so - I renamed the old LGPLv3 webdav code to webdav-neon, and imported the (not built) serf webdav ucp into the old space. so that in future, we can merge changes more easily - and still choose which to use. cbosdonnat kindly volunteered to do some comparative analysis of the two codebases to decide which is best for what etc.