Cisco WS-C2960X-24PS RSTP

Yes. RSTP will block both providing the cabling is cory.

If  there is more than one VLAN, or you are using some other VLAN than VLAN1 is hard to predict behavior. CIsco's RSTP implementation (Rapid PVST+) is not fully compatible with IEEE 802.1w standard.

Rapid is exactly as per the standard.

Not really. Beginning from STP cost. :)
Typically there can be found tests of STP interoperability between Cisco's STP versions with other vendors in mixed environment. If I remember correctly, there is also a different way to calculate STP cost of STP ports etc...

Arista - Spanning Tree Protocol Interoperability With Cisco PVST+/PVRST+/MSTP
HP - ProCurve & Cisco Spanning Tree Interoperability  

Generally, MSTP is recommended, but again as you can see in Arista testing it lead to unpredictable result. Interoperability with Juniper is also problematic.

Actually the Arista article explains interoperability with PVST+, not RPVST.

This will clarify...

https://supportforums.cisco.com/discussion/11442331/stp-8021d-vs-pvst-and-rstp-8021w-vs-rapid-pvst

I am not talking about interoperability between PVST+ or RPVST or MTSP...
But generally, interoperability between Cisco's implementation and other vendors. There are problems. MSTP suppose to be most standardized and there are still errors even with that one. It was tested and sometimes gives unpredictable results, that's my point.

But, enough about it... I agree that we disagree.
:)

Cisco's implementation has "enhancements" on the native VLAN but all other VLANs run the IEEE versions of each STP protocol.

If you know how to config trunks it's all good. There's no actual interoperability issues, just bad config.

It is not just about trunks.

IEEE is using destination multicast MAC address 0180.C200.0000, in Cisco's implementation only VLAN1 runs both multicast BPDU destination for compatibility (0180.C200.0000 & 0100.0CCC.CCCD). All other VLANs do not run IEEE compatible version since all VLANs use multicast MAC address 0100.0CCC.CCCD.

In that case for all other VLANs except VLAN1 Cisco will for sure choose itself as Root Bridge (RB) of spanning tree, only for VLAN1 Cisco is doing RB election with other switches. That would mean that only in the case that Cisco switch is configured as RB in topology have chance to work properly. In all other cases that some other switch is configured with best priority and become RB - all VLANs except VLAN1 will have 2 RBs. As a consequence - on Cisco switch all other VLANs will never block any non VLAN1 port.
Some vendors have implementation that are also per VLAN and listening for BPDUs with 0100.0CCC.CCCD multicast address to be compatible with Cisco's implementation (it typically include additional configuration steps on non Cisco switches), not sure about Comnet switches.

My understanding is the opposite. The native VLAN uses RapidPVST but each tagged VLAN runs IEEE 802.1w RSTP. Therefore if your trunks aren't configured properly you'll get issues, otherwise compatibility is guaranteed.

There is a nice DELL article - Dell - Cisco Spanning Tree Interoperability Testing and Recommendations

From part: What exactly happened and why was reconvergence so slow?

For all 802.1Q tagged VLANs (more specifically, all VLANs besides VLAN 1), Cisco switches send their BPDUs only to the reserved Cisco multicast address of 0100.0CCC.CCCD. Therefore, unless the third party switch is also listening to this multicast address, its STP process will only have visibility to, and an understanding of, the logical topology of the CST. That means the rapid failover and convergence offered by RSTP will only be evident on VLAN 1. Regardless of the existence of other VLANs, from an STP perspective, the standards - compliant bridge “sees” only VLAN 1, and therefore only one logical STP topology exists.

For all VLANs, except VLAN 1, the Dell 6224 switch simply received the multicast packets destined for the reserved Cisco multicast address and flooded them out all corresponding VLAN ports, much the same way broadcasts are handled. In other words, the BPDUs for the tagged 802.1Q VLANs were simply tunneled through the non-Cisco region.

There are non-Cisco switches that do listen to the Cisco multicast address of 0100.0cCC.CCCD for the sake of interoperability, such as Extreme Networks, Force 10, some Foundry (now Brocade) switches and Juniper's VSTP.

OVERCOMING THE LIMITATIONS OF RSTP/rapid-PVST+ INTEROPERABILITY - explains why MSTP should be used instead of RSTP and all recommendations for STP interoperability that I read were pointing into same direction.

I would like to say a big thank you to everyone, on site on Wednesday, a lot of Cisco knowledge refreshing this weekend.