Linear optics is one of the earliest hardware candidates for quantum computation. It offers several key advantages: robustness against decoherence, compatibility with quantum communication, and room-temperature operation. However, its development has faced significant challenges. Most notably the susceptibility to photon loss and the lack of deterministic entangling gates. The emergence of quantum-emitter systems that combine a spin qubit with the ability to emit entangled photons offers a promising path forward. These systems provide new opportunities to overcome the long-standing challenges of linear optics, particularly in the context of fault-tolerant quantum computation (FTQC).In this thesis, I study how to efficiently generate entanglement between photonic qubits using linear optics. I explore new architectures for FTQC that combine linear optical tools with emitter-based entanglement generation. A central objective is to design a computation scheme that is both robust to photon loss and adapted to the physical constraints of entanglement creation in photonic platforms. I present results that clarify some of the fundamental limitations of entanglement generation in linear optics. In particular, I investigate the constraints associated with preparing two-photon entangled states, both in post-selected and heralded generation schemes. Furthermore, I introduce a new interpretation to entangling measurement interferometers, as the optical Bell state measurement, which led to the discovery of novel heralded entangling gates for photonic qubits. I also investigate a novel architecture for FTQC based on quantum emitters and analyze its fault-tolerance performance under relevant noise models. Computation occurs on spin-degrees of freedom of quantum emitters while entangling gates are performed via entangling measurements of photons emitted. This hybrid architecture aims to keep photonics assets while alleviating its two main flaws: susceptibility to photon loss and probabilistic entanglement. This architecture is resource-efficient and introduces a new set of opportunities and constraints for FTQC design. In particular, it enables the efficient implementation of FTQC circuits based on Floquet codes.