Still useful to use init-repository because of all the other logic, such
as repository urls. More general than --no-webkit, although it is still
a valid convenience as the most common case.
Also makes qt.pro check for existence of modules, since it is no longer
certain that you'll have everything but webkit.
Change-Id: Iaaaa30e24e0a793218e2e307bdde56067760c513
Reviewed-on: http://codereview.qt.nokia.com/1258
Reviewed-by: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Some of the depends were referring to qmake targets and some of them
were referring to make targets. That's unnecessarily confusing. Make
them consistent.
For the sake of a uniform test procedure, it is cleaner to have
`make module-qtactiveqt' always working, but simply not building
anything on platforms where ActiveQt is not supported.
Without this, any test procedure for building Qt5 needs to be able
to know or determine that ActiveQt should only be built on Windows.
Reviewed-by: Jason McDonald
Change-Id: I3869bbacdc6598a764561a4b9f6042c49e6ffae8
The previous method of adding qtlocation did not allow
`make module-qtlocation' to work.
Reviewed-by: Alex
Change-Id: I75caab75ddf8ed5800c75933097df297b0021c6c
Force all top-level makefile rules to be named `module-$GITSUBMODULE',
including the addition of the missing qlalr and qtqa targets.
This naming convention must be upheld, otherwise there is no simple
way to programmatically build a single git submodule and all of its
dependencies.
Reviewed-by: Liang Qi
