Permalink
Commits on Oct 19, 2015
  1. Migration to git.merproject.org

    lbt committed Oct 19, 2015
    Signed-off-by: David Greaves <david.greaves@jolla.com>
Commits on May 7, 2015
  1. Merge pull request #58 from amtep/master

    amtep committed May 7, 2015
    Fix MTP compatibility with some android kernels
Commits on May 4, 2015
  1. Fix compatibility with some android kernels, alternate approach

    amtep committed Apr 29, 2015
    The length field was wrong in the previous version
  2. [compat] Fix compatibility with some android kernels, fixes JB#26329

    amtep committed Apr 9, 2015
    There's an ad-hoc patch for superspeed support that breaks the API,
    which is different from how it's supported upstream.
    
    We don't use superspeed either way yet so our only concern needs
    to be compatibility.
    
    Working around this so that Sailfish can support MTP on affected kernels.
Commits on Nov 6, 2014
  1. Merge pull request #57 from amtep/master

    amtep committed Nov 6, 2014
    [tests] Make autotests pass again
  2. [fsstorageplugin] pass ObjHandle by value

    amtep committed Sep 11, 2014
    Changed methods removeFromStorage and deleteItemHelper to take
    their ObjHandle by value instead of by reference, on the grounds
    that they're going to remove that handle from some maps and this
    is highly awkward if the calling function passed in a reference
    taken from those maps.
    
    Specifically, the autotests found an error where handleFSMove
    passes m_pathNamesMap[oldPath] as the argument, and the []
    operator returns a reference to the map element, which is
    then deleted in removeFromStorage.
    
    This is the kind of subtle bug that makes C++ so beloved.
  3. [fsstorageplugin] Initialize the cached inotify event

    amtep committed Sep 11, 2014
    Class member initialization was missing, found with valgrind.
    This was causing crashes in the autotests.
  4. [tests] Initialize m_opcode in mtpresponder_test

    amtep committed Sep 11, 2014
    found by valgrind
  5. [tests] don't use global QEventLoop in protocol_test

    amtep committed Sep 10, 2014
    QEventLoop complains if it is constructed before QApplication,
    and this may be involved in a hard-to-reproduce segfault in
    initTestCase. Move it into the function that uses it.
  6. [deviceinfo] initialize m_batteryLevel

    amtep committed Sep 10, 2014
    issue found by valgrind
  7. [tests] Added storagefactor-test to tests.xml

    amtep committed Sep 10, 2014
    Also removed some inaccurate comments.
    It would be nice to change the suite and set names to reflect the
    new package name, but that would mess with the continuity of
    test results for no good reason.
Commits on Oct 7, 2014
  1. Merge pull request #56 from faenil/nemomerge_fixes

    faenil committed Oct 7, 2014
    [compilation] Fix compilation on glibc 2.19/gcc 4.8
Commits on Sep 18, 2014
  1. Merge pull request #55 from monich/build-fix

    monich committed Sep 18, 2014
    Fix build break
  2. [buteo-mtp] Fix build break

    monich committed Sep 18, 2014
    Buteo SyncDBusConnection is no more
Commits on Aug 27, 2014
  1. [rpm] Also pull in libqt5sparql-tracker

    amtep committed Aug 27, 2014
    The tests fail without it, so presumably buteo-mtp relies on it.
    In the release image libqt5sparql-tracker is pulled in by other
    things so this doesn't affect the release.
Commits on Jul 28, 2014
  1. Merge pull request #54 from nemomobile/fixbuild

    chriadam committed Jul 28, 2014
    [buteo-mtp] Fix build break
  2. [buteo-mtp] Fix build break

    chriadam committed Jul 28, 2014
Commits on Jun 12, 2014
  1. Merge pull request #53 from nemomobile/fixmtpserver

    chriadam committed Jun 12, 2014
    [buteo-mtp] Fix build of mtpserver
Commits on May 19, 2014
  1. Merge PR#52 from xhaakon/master

    amtep committed May 19, 2014
  2. [tests] fix MTPResponder_test::testSetObjectReferences()

    xhaakon committed May 19, 2014
    On devices with many stored files, 1000 might be a too low object handle
    to be invalid.
  3. [tests] wait until storages are ready

    xhaakon committed May 19, 2014
    We have to wait until the storages are initialized before proceeding
    with the test case execution.
Commits on May 6, 2014
  1. Merge pull request #51 from xhaakon/master

    amtep committed May 6, 2014
    Fix PUOID storage and related stuff
    
    The puoid persistence logic was broken by the sd card support and this PR fixes it
Commits on Apr 17, 2014
  1. [tests] don't link storage-test with libblkid and libmount

    xhaakon committed Apr 17, 2014
    Avoid storage-test crash on device:
    
    [W] QCoreApplicationPrivate::QCoreApplicationPrivate:416 - WARNING: \
    QApplication was not created in the main() thread.
    Segmentation fault (core dumped)
    
    It's probably caused by TLS variable currentThreadData in Qt's
    qthread_unix.cpp not getting initialized to zero, which seems like a
    compiler or run-time linker issue. Once sorted out, this commit may be
    reverted.
Commits on Apr 16, 2014
  1. [fsstorage] remove legacy PUOID database

    xhaakon committed Apr 15, 2014
    The file used in previous versions of buteo-mtp now only wastes device
    memory.
  2. [fsstorage] use unique filenames for PUOID databases

    xhaakon committed Apr 14, 2014
    All instances of FsStoragePlugin were using file ~/.local/mtp/mtppuoids,
    overwriting each other's data. The new naming scheme is
    
     mtppuoids-{name}-{fsuuid}
    
    Where name is a storage name attribute from its configuration file in
    /etc/fsstorage.d and fsuuid is a UUID of storage's filesystem.
    
    UUID alone is not enough for unique distinction since different parts of
    a single filesystem may be exported as different MTP storages. In that
    case, user-configured storage names ensure unique database filenames.
Commits on Apr 15, 2014
  1. [fsstorage] ensure absolute FSStoragePlugin::m_storagePath

    xhaakon committed Apr 14, 2014
    User may write the storage path in /etc/fsstorage.d/*.xml with extra
    slashes or '.' and '..' elements. We want to always have the storage
    path in the absolute format to allow correct string comparison etc.
  2. [storage] move StoragePlugin::m_storagePath to FsStoragePlugin

    xhaakon committed Apr 15, 2014
    This protected class member doesn't play any role in the abstract
    interface of a storage plugin. In general, for some potential kinds of
    storage plugin, implemented in the future, the concept of "storage path"
    might not even make any sense. Consequently, let's move the member to
    the place of its use, which is FsStoragePlugin.
  3. [fsstorage] remove unused m_uniqueObjectHandle

    xhaakon committed Apr 15, 2014
    The member is assigned but never read. In unit tests, use a static
    variable to provide object handles when StorageFactory isn't available.
  4. [fsstorage] remove unused m_rootFolderPath

    xhaakon committed Apr 14, 2014
    FsStoragePlugin::m_rootFolderPath wasn't used anywhere, likely being a
    duplicate of StoragePlugin::m_storagePath.
  5. [storage] remove HalProxy

    xhaakon committed Apr 14, 2014
    HAL passed away, never quite loved. Rest in peace.
Commits on Apr 8, 2014
  1. Merge pull request #50 from amtep/jb12235

    amtep committed Apr 8, 2014
    Do the storage enumeration (file scan) in the background in order to keep OS/X from becoming impatient.