Epic: Performance Optimizations

[brentleyjones]

This issue will track know areas that we can improve the performance: Bazel memory usage, analysis phase wall-time, and generator wall-time.

Bazel memory

• [x] #2045
• [x] #2049
• [x] #2052
• [x] #2053
• [x] #2056
• [x] #2059
• [x] #2067
• [x] #2073
• [x] #2074
• [x] #2076
• [x] #2079
• [x] #2081
• [x] #2086
• [x] #2087
• [x] #2090
• [x] #2094
• [x] #2095
• [x] #2096
• [x] #2098
• [x] #2099
• [x] #2101
• [x] #2102
• [x] #2104
• [x] #2105
• [x] #2107
• [x] #2110
• [ ] #2111
• [ ] #2212
• [ ] Use a hash for TargetID instead of the raw string
• [ ] Reduce size of xcode_target

Analysis phase time

• [x] #2063
• [x] #2082
• [x] #2083
• [x] #2089
• [x] #2112
• [x] #2114
• [x] #2115
• [x] #2116
• [x] #2119
• [x] #2129
• [x] #2130
• [x] #2134
• [ ] #2212
• [ ] Normalizing file paths could happen as part of args()? Maybe move spec creation into its own action (similar to swift_debug_settings.py.

generator time

• [x] #2043
• [x] #2042
• [ ] XcodeProj.write() is largely serial

[jpsim]

I'm taking a look at parallelizing XcodeProj.write().