srand函数是否跨多个翻译单元生效？主函数调用是否影响其他单元函数结果？

Great question—this gets to how C handles global state for standard library functions, which is a common point of confusion. Let's break it down clearly:

Does srand() work across multiple translation units? #

Absolutely. The rand() function relies on a single, process-wide global state to generate its sequence of pseudo-random numbers. When you call srand(), you're modifying this global state directly—regardless of which translation unit the srand() call lives in. All code in your program, across every linked translation unit, will use this updated state for subsequent rand() calls.

Will calling srand() in main() affect functions in other translation units? #

Yes, 100%—as long as those functions use rand().

For example, say you have:

• A main.c where you call srand(time(NULL)) at the start of main()
• A utils.c translation unit with a function generate_random_id() that calls rand() to create an ID

When your program runs, the srand() in main() seeds the global random state. Any call to generate_random_id() (or any other function in any translation unit that uses rand()) will generate numbers based on that seeded state, rather than the default seed (which is typically 1).

A quick note on why this works #

Translation units are just a compile-time concept—once your code is linked into a single executable, all parts of the program share the same process memory space. The random number state used by rand() is stored in a global static variable (or similar process-wide storage) managed by the standard library, so changes to it from any part of the program are visible everywhere.

内容的提问来源于stack exchange，提问作者Serket