extra javac opts should override default ones

[thomasbao12]

Description

Fixes #1550

Currently the default javac flags will be appended last, overriding the manual ones.

Motivation

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[liucijus]

@thomasbao12 could you add a test to prevent regressions?

[thomasbao12]

@liucijus I added a test and verified it fails on master #1615.

I tried to follow the pattern established in other javac flag tests.

One issue though is I assume the default java toolchain sets a -target flag.

3 potential options:

0. merge as is
1. I could rename the test to clarify that I actually am only testing that I want to be able to override the java target version. This is actually the behavior I want to fix.
2. I could try a more elaborate unit-test that verifies that two javac flags exist, one from the default toolchain and one from javacopts and that the javacopt one is ordered after the default toolchain.

I'm okay with 0 & 1. I worry 2 is not worth doing and overly specific.

[liucijus]

3 potential options:

0. merge as is

1. I could rename the test to clarify that I actually am only testing that I want to be able to override the java target version.  This is actually the behavior I want to fix.

2. I could try a more elaborate unit-test that verifies that two javac flags exist, one from the default toolchain and one from javacopts and that the javacopt one is ordered after the default toolchain.

@thomasbao12 if I haven't missed anything, option 0 seems the best to me - it communicates your need plus the name implies generic expectation

[liucijus]

Thanks, @thomasbao12!

[thomasbao12]

@liucijus Can you merge?