HP 3312A Erratum?

[WideBandwidth]

I've been looking at the HP 3312A function gen schematics, and came across an unusual op amp configuration.
For reference, the service manual can be found from BAMA: https://bama.edebris.com/download/hp/3312a/3312A_op_ser.pdf.
The schematic in question is located on page 60 of the PDF.
I've also attached the specific schematic page in question.

On the A2 board (the triangle wave integrator board), a controllable current source and a controllable current sink are used to charge and discharge the timing capacitor. The current source and sink are set to be identical currents when a symmetric triangle wave is desired, and dissimilar currents otherwise.
Op amps U102 and U103 take a frequency control signal and generate the required current control signals for the source and sink. However, the schematic seems to indicate that U102 and U103 are both configured with positive feedback, which should rail the op amps. That is, the non-inverting terminal is used in the feedback path without any mechanism to invert the feedback gain to get negative feedback.

If anything, I would expect U102 and U103 to instead be connected in a constant current source configuration (negative feedback from the Q103/Q104's emitter resistor, setting the transistor current). That constant current would then be converted to a voltage via the collector resistors of Q103/Q104. However in their present state, it sure seems like the circuit would fail to do anything but rail.
Any thoughts?