Frequently Asked Questions (original) (raw)

1. Why can't I randomly access fields within a message?

One of the design principles for SBE is to stream through messages in a ascending sequential access pattern for performance reasons.

2. What are the targets for C++?

There are a number of ant targets for C++ that use cpptask for compilation. The ant cpp:test target compiles and run tests on Linux (Ubuntu 16.04 LTS) fine. There is some compiler warns on the generated code about signed/unsigned compares. Those are OK for now and will be cleaned up. Here are the various targets related to cpp.

• cpp: compile OTF files, tests, run tests, generate Doxygen doc, and package doc into zip
• cpp:test: compile cpp OTF files, compile Google Test files, run unit tests for OTF decoder
• cpp:doc: generate Doxygen doc for OTF decoder
• cpp:dist: generate Doxygen doc for OTF decoder and package into zip
• examples:cpp: compiles and runs generated and OTF decoder examples
• cpp:examples:gen: compile and run cpp example
• cpp:examples:otf: compile and run cpp OTF decoder example

NOTE: Doxygen is required. If you don't have it, the cpp target will not succeed. But cpp:test should still work.

3. What if a 3rd party schema contains names that are language keywords?

Some 3rd parties have not considered that the names they use could cause a clash with language keywords such as Java final. Generated property names that clash with keywords can have a string appended, such as the _ character, by setting the sbe.keyword.append.token system property when running the SbeTool. The templates_FixBinary.xml from the CME is such an example.