commit | cfa7934d4e579d4e7e589042c3af1044e63d45a5 | [log] [tgz] |
---|---|---|
author | Pavel Maltsev <pavelm@google.com> | Mon Nov 21 22:50:18 2016 -0800 |
committer | Martijn Coenen <maco@google.com> | Tue Nov 22 08:53:51 2016 +0100 |
tree | 0cc1abd4d9663bce484612c8ebcf72d808b55590 | |
parent | 23a0a3a055aac5d3a20ae44a8b4d918844fe45ea [diff] |
Fix java-client generation of HIDL code hidl_vec<T> and hidl_string was recently changed to use uint32_t for mSize field instead of size_t. See ag/1640383 As a result Java clients started to crash because the size of hidl_vec and hidl_string was reduced to 16 bytes and Java gen code was not adjusted. Test: our java client is no longer crashing Change-Id: I506ecd61277e3a4df96a10180c46ea924a42235d Fix: b/33038659
croot make hidl-gen
hidl-gen -o output-path -L language (-r interface-root) fqname output-path: directory to store the output files. language: output file for given language. e.g.c++, vts.. fqname: fully qualified name of the input files. For singe file input, follow the format: package@version::fileName For directory input, follow the format: package@version interface-root(optional): prefix and root path for fqname. If not set, use the default prefix: android.hardware and default root path defined in $TOP. examples: hidl-gen -o output -L c++ android.hardware.nfc@1.0::INfc.hal hidl-gen -o output -L vts android.hardware.nfc@1.0 hidl-gen -o test -L c++ -r android.hardware:/home/android/master/hardware/interfaces android.hardware.nfc@1.0