Recent progress in large language models (LLMs) has led to impressive code generation capabilities. However, existing evaluations of LLMs primarily focus on generating isolated, small-scale code units (e.g., single functions or statements) under default or unspecified software environments. As a result, it remains unclear whether LLMs can reliably generate executable code tailored to specific user environments.. To fill this knowledge gap, we make the first systematic study of Environment-Aware Code Generation (EACG), which requires generating code that is both functionally correct and directly executable under arbitrary software configurations. We introduce VersiBCB, a large-scale benchmark constructed from real-world Python projects, featuring diverse environment specifications and realistic, multi-package scenarios. Building on this benchmark, we develop three representative adaptation strategies for LLMs: retrieval-augmented generation (data-based), mixture-of-experts (parameter-based), and memory-augmented generation (cache-based). We empirically evaluate these methods across tasks including code completion, function repair, and API migration. Our results reveal that existing LLMs struggle with environment-specific code generation, but our adaptation strategies yield improvements in environment compatibility and executability. These findings highlight critical challenges and opportunities for deploying LLMs in practical, heterogeneous software engineering workflows, and underscore the need for further research into environment-aware AI-assisted programming.