使用arm-gcc3.3.1在32位pic上进行循环

Question

Im going to be downvoted to oblivion but: this question : https://stackoverflow.com/questions/43510049/c99-for-loop-looses-static-variable-values as been tagged as incomplete but its not.

User pointed out all elements of a reccurent bug under gcc-arm-toochain3.3.1, and even more reccurent on 32-bit pics (all question from this user are related to those), when working with array of structures inside a structure.

This bug is created by the pic considering the succession on pointer in the structure of structure as a single variable name. when the last bit of the index isnt 0, the buffer overflow reading values with the wrong type as multiple of max int.

Changing the code from:

structure1.structure2[i].var1 = blabla;

by

structure2_ * ptr = NULL;
ptr = &structure1.structure2[i];
ptr->var1 = blabla;

will fix it instantly (example recreated under keil3.0 & windows xp with arm-gcc-toolchain 3.3.1 and tested on pic)

Why am I posting this as a question?

1) Post is on hold thus I cant answer it.

2) I dont have enough reputation thus I can't comment on it.

3) peers whom reviewed the edit didnt considered to push the edit as an answer/comment

So my question - it should be in meta, but once again not enough reputation - is: Why does edit reviewer don't accept my edit for an on hold question if the edit is the answer to said question?

Downvote away!

edit if someone could link this thread to the other question, I'll appreciate it.

Answer 1

我不确定这是否真的是答案，但建议编辑与此帖子相关的内容，而我的问题确实解决了我的问题。