commit | 2802c1d7c286788437a7c904f7b5a2514f4400df | [log] [tgz] |
---|---|---|
author | davg <davg@google.com> | Fri Aug 31 15:34:20 2018 -0400 |
committer | kelvinchan-google <kelvinchan@google.com> | Wed Sep 05 18:48:05 2018 -0400 |
tree | 6570f663ece23725d5f742f8e2f6eb10eee285ce | |
parent | b7c9c714e4e19cd90cf87ac280294d9fc623d271 [diff] |
Fix Xcode 10 on-device testing framework injection Xcode 10 injects a dylib into test host targets which Tulsi needs to resign. Tulsi previously only resigned .framework bundles. PiperOrigin-RevId: 211126953
Open src/Tulsi.xcodeproj, and within Xcode, build the TulsiApp.
Run the TulsiApp.
Tulsi-generated Xcode projects use Bazel to build, not Xcode via xcbuild. This means that many common components of an Xcode project are handled differently than you may be used to. Notable differences:
bazel
invocations, some of which may affect Bazel caching. In order to maximize cache re-use when building from the command line, try using the user_build.py
script which is located in the generated xcodeproj at <xcodeproj>/.tulsi/Scripts/user_build.py
.Tulsi projects contain a few settings which control various behaviors during project generation and builds.
build
flags, customizable per compilation mode (dbg
and opt
)build
startup flags, also customizable per compilation modedbg
or opt
, no fastbuild
) used during project generation.dbg
, swap to opt
if you normally build Release builds in Xcode (i.e. profiling your app). Setting this improperly shouldn't break your project although it may potentially worsen generation and build performance.No
, swap to Yes
if your project contains Swift (even in its dependencies). Setting this improperly shouldn't break your project although it may potentially worsen generation and build performance.