I'm still focusing on Mac porting work. Without being familiar with Unix-like OS, my work efficiency is far lower than on Windows. What's more, another potential issue came out. I've got some operating system level development experience, which improved my understanding to computer system. Along with the benefits from such background, I am trapped by it too.
After reading the implementation of some basic OS APIs, and even implementing my own version of those, I start to lose the trust to all the existing OS APIs. I met a tricky issue on fclose() on Mac that it seemed not to release the handle resource at all. This mistake cost me nearly two hours and deprived chances for me to dig other spots. After testing, fclose() has no problem.-- It sounds ridiculous to test Mac internal. And I admit it.
At least I got one rule for debug : believe your working OS, or you'll be like in a black hole.
After reading the implementation of some basic OS APIs, and even implementing my own version of those, I start to lose the trust to all the existing OS APIs. I met a tricky issue on fclose() on Mac that it seemed not to release the handle resource at all. This mistake cost me nearly two hours and deprived chances for me to dig other spots. After testing, fclose() has no problem.-- It sounds ridiculous to test Mac internal. And I admit it.
At least I got one rule for debug : believe your working OS, or you'll be like in a black hole.

作者在进行Mac移植工作中遭遇效率低下问题,并反思了在非Windows环境下工作带来的挑战。由于具备操作系统级开发经验,作者对现有OS API产生了怀疑,在调试过程中耗费大量时间于基本API的验证上。
1338

被折叠的 条评论
为什么被折叠?



