Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.os.linux.misc    |    Linux-specific topics not covered by oth    |    135,536 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 134,277 of 135,536    |
|    rbowman to Peter Flass    |
|    Re: naughty Python    |
|    03 Jan 26 00:52:51    |
      XPost: alt.folklore.computers       From: bowman@montana.com              On Fri, 2 Jan 2026 15:00:07 -0700, Peter Flass wrote:              > On 1/2/26 14:33, Lawrence D’Oliveiro wrote:       >> On Fri, 2 Jan 2026 08:49:25 -0800, John Ames wrote:       >>       >>> ... it's not guaranteed that the compiler won't take liberties in       >>> arranging members of a struct for optimization purposes ...       >>       >> The C23 spec (section 6.2.5, “Types”) does say the member objects of a       >> struct type need to be “sequentially allocated”. The only freedom the       >> compiler has (section 6.2.6) is to add “padding bytes”.       >       > It defeats the purpose of a structure if the compiler is free to       > rearrange it. Local variables (PL/I AUTOMATIC) can, in most languages,       > be stored however the compiler wants.              That can lead to interesting bugs. The root cause is overflowing a local       variable, say writing 6 characters to a char[4]. Which adjacent local       variable gets whacked depends on the compiler's ordering. Whether it       manifests as a bug depends on how the corrupt variable is used in the       function and where it is initialized.              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca