dfaure added a comment.
This commit breaks kfileplacesmodeltest.
$ bin/kfileplacesmodeltest testInitialState testInitialList
testInternalBookmarksHaveIds
- Start testing of KFilePlacesModelTest *********
Config: Using QtTest library 5.12.0, Qt 5.12.0 (x86_64-little_endian-lp64
shared (dynamic) debug build; by GCC 7.3.1 20180323 [gcc-7-branch revision
258812])
QDEBUG : KFilePlacesModelTest::initTestCase() 18:21:44.046
kfileplacesmodeltest(22234)
?[32mSolid::Backends::Fake::FakeManager::parseMachineFile?[0m void
Solid::Backends::Fake::FakeManager::parseMachineFile() Parsing fake computer
XML: "/d/kde/src/5/frameworks/kio/autotests/fakecomputer.xml"
PASS : KFilePlacesModelTest::initTestCase()
PASS : KFilePlacesModelTest::testInitialState()
PASS : KFilePlacesModelTest::testInitialList()
FAIL! : KFilePlacesModelTest::testInternalBookmarksHaveIds()
'!ids.contains(id)' returned FALSE. (Duplicated ID found!)
Loc: [/d/kde/src/5/frameworks/kio/autotests/kfileplacesmodeltest.cpp(315)]
PASS : KFilePlacesModelTest::cleanupTestCase()
Totals: 4 passed, 1 failed, 0 skipped, 0 blacklisted, 75ms
- Finished testing of KFilePlacesModelTest *********
I have debugged it a little bit but I don't really understand the overall
logic.
Many bookmarks have a UDI (like
/org/kde/solid/fakehw/fstab/thehost/solidpath) and no URL, and those are the
ones that are duplicated now.
When this code is called again, those bookmarks are now skipped in the first
iteration, and therefore added by the next foreach (the one which says "Add
bookmarks for the remaining devices, they were previous unknown").
Maybe this skipping should only be done for bookmarks with an empty UDI?
REPOSITORY
R241 KIO
REVISION DETAIL
https://phabricator.kde.org/D14072
To: jtamate, dfaure, #frameworks
Cc: kde-frameworks-devel, michaelh, ngraham, bruns
