ae19ac0739
[required variant with jstestshell changes](https://spruce.mongodb.com/version/6643e8bf8571a30007c6562d/tasks?sorts=STATUS%3AASC%3BBASE_STATUS%3ADESC) <- 242 hours [required variant without jstestshell changes](https://spruce.mongodb.com/version/6643e9dd8571a30007c6576b/tasks?sorts=STATUS%3AASC%3BBASE_STATUS%3ADESC) <- 301 hours 20% reduction in compute. [AUBSAN After](https://spruce.mongodb.com/version/6644ecfaf87baf0007e4b124/tasks): 412 hours [AUBSAN Before](https://spruce.mongodb.com/version/6644e4b131a6190007b0526c/tasks) 539 hours 24% reduction in compute [TSAN After](https://spruce.mongodb.com/version/6644ecabe2618d0007a54406/tasks) 356 hours [TSAN Before](https://spruce.mongodb.com/version/6644e47f9b110e0007e9820f/tasks) 580 hours 39% reduction in compute According to project outliers this should reduce our spend on mongodb-mongo-master by (20%*20.32%) = 4.06% (39%*14.82%) = 5.78% (24%*8.73%) = 2.10% **Total: 11.94%** In all these patches we use a jstest shell compiled without debug symbols and statically linked. This increased the startup speed from 1.2 seconds to .02 seconds. Since each javascript test is run with a separate invocation of the shell this speeds up ever javascript test by about 1.2 seconds. The tradeoff is that we are not going to catch tsan or aubsan bugs in the jstestshell and it makes our test running process just that much more complicated. Not included here is another optimization to run batches of javascript tests together to avoid having to re-handshake with the database. [Final PB](https://spruce.mongodb.com/version/664b780cceb3230007a77382/tasks?sorts=STATUS%3AASC%3BBASE_STATUS%3ADESC) showing mostly green. GitOrigin-RevId: db2f54f832512676f6f8159e1267d5ae69aa6b3d