Here are some route splits/truncations, at national park boundaries, I plan to make to the in-dev route files. These are based on legislative route definitions in the state
Streets and Highways Code; paper 2002 Caltrans route logs;
online 2015 Caltrans bridge logs (with much of the information included in the older paper logs); and GMSV imagery.
In Yosemite National Park:CA 120: Route split, with the Manteca segment ending at the west park boundary, and the Lee Vining segment starting at the east park boundary. The statutory route definition is unusually explicit that CA 120 does not enter or cross the national park. Caltrans's route and bridge logs follow the statutory definition. There is some non-Caltrans signage at a key junction within the park, pointing motorists to CA 120 on either side of the park, but that's not enough to override the state's definition of its own route. The Yosemite NP website notes that a key road connecting the two CA 120 segments was never state-maintained -- it started off as a private mine road, which then was donated to the National Park Service.
CA 41: Route truncated to end at the southern park boundary. Like with CA 120, there is non-Caltrans signage pointing motorists toward CA 41. However, Caltrans' logs cut off the route at the park boundary, and there is an End CA 41 sign to underscore that.
CA 140: Route truncated to end at the southwestern park boundary. Similar to CA 41, except I didn't see any End signs.
In Lassen Volcanic National Park:CA 89: Route split, with the Lake Tahoe segment ending at the southwestern park entrance, and the Mt. Shasta segment starting at the junction with CA 44 at the northwestern park boundary, The statutory route definition is unclear about whether there is a break in the route (it does say that one route segment proceeds north from the CA 44 junction, but is vague on where in the national park the segment to the south ends). But Caltrans' logs clearly show CA 89 does not exist within the park. No End signs I could find, and GMSV coverage within the park is from 2007 and incomplete (my guess is the camera car went there in late autumn just before the park shut down for the winter, but after a snow closure kept it from driving all the way through the park).
In Pinnacles National Park (formerly Pinnacles National Monument):
CA 146: Both route segments truncated to end at the west and east park boundaries, respectively, removing a few miles from each segment (there is no drive-able road from one side of the park to the other, so we already had split route files). This is confirmed by Caltrans' logs, and End CA 146 signs at both park entrances.
In Kings Canyon National Park:CA 180: I already truncated the route at the eastern Cedar Grove entrance to the park, which had been previously mapped to end a few miles inside that part of the park. That much is clear, from the statutory route definition, Caltrans' logs, and the End 180 sign at the Cedar Grove park boundary.
Before getting to Cedar Grove, the highway crosses the separate Grant Grove section of the park for about four miles, before exiting the park then ending about 22 miles later at Cedar Grove. I would
not split the route at Grant Grove. The legal and Caltrans route definitions suggest such a route split, though unclearly. But not only are there no End signs at the Grant Grove park boundaries, there is a Caltrans-spec CA 180 route marker at a major junction well within the Grant Grove section of the park. (But Caltrans postmile markers appear to be only outside the park.) Also, chopping up the route at Grant Grove would leave a significant section of CA 180 between Grant Grove and Cedar Grove with no state route connection to the rest of the state highway system (indeed, no road connection at all, except through Grant Grove). This situation is unclear enough for me to leave this alone.
In Death Valley National Park:CA 178 and CA 190: No changes. The park was established (in 1994) after the state highways were already in place, and neither the state legislature nor Caltrans have adjusted those route definitions to take them out of the new park. Also, I've seen Caltrans-spec route markers for both routes (including an End 178 marker) well within park boundaries.
****
Now all this removes a lot of mileage, including through routes for Yosemite and Lassen Volcanic parks, from the California State Highways route set. But not to fear! Much of that mileage, and maybe some other major park roads, would be good candidates for addition to the in-dev U.S. National Park Highways (usanp) route set, which was designed in part to fill in gaps created by national parks in other route systems (such as the gaping hole created by the official non-existence of US highways within Yellowstone NP).