Automatic gene sequencing prevents research from being completed

[daniel-sch]

When automatic gene sequencing is enabled and a tech with knowledge costs barely smaller than the knowledge cap is added to the research queue, the research will never be completed because the knowledge will be spent on gene sequencing.

My hypothesis is that the game loop checks the condition for automatic gene sequencing first (which is satisfied because the knowledge cap is reached) and later checks the research queue (which will never be satisfied because the knowledge is already spent on gene sequencing).

[yarukishi]

The priority inversion issue is that the knowledge buffer (underneath cap) is different for gene sequencing (10,000 knowledge) and construction (1 day's production). Gene sequencing wins when production is under 10,000 knowledge per day.

PR already available: #1023