<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW --- - Misuse of setjmp in DCTStream" href="https://bugs.freedesktop.org/show_bug.cgi?id=63067#c4">Comment # 4</a> on <a class="bz_bug_link bz_status_NEW " title="NEW --- - Misuse of setjmp in DCTStream" href="https://bugs.freedesktop.org/show_bug.cgi?id=63067">bug 63067</a> from <a class="email" href="mailto:Thomas.Freitag@alfa.de" title="Thomas Freitag <Thomas.Freitag@alfa.de>"> Thomas Freitag</a> <pre>(In reply to <a href="show_bug.cgi?id=63067#c0">comment #0</a>) > According to POSIX and the C spec, `setjmp` should only be used as the > "entire controlling expression" of a selection of iteration statement, > possibly with `!` or `== 0` or similar. > > But in DCTStream.cc is the line: > > if (!setjmp(err.setjmp_buffer) && jpeg_read_header(&cinfo, TRUE) != > JPEG_SUSPENDED) This is NOT a iteration statement, iteration statements are "for", "while" and similar! > > If I'm not mistaken, that is undefined behavior. Not that it broke anything > yet... > > Instead it should be: > > if (!setjmp(err.setjmp_buffer)) > { > if (jpeg_read_header(&cinfo, TRUE) != JPEG_SUSPENDED) Okay, okay, the order of execution of an if condition is undefined in C/C++ (other than if Java, where You can count on that it will be done from left to right). BUT I never see an c/c++ compiler which does it in another way. But if we are puristic: Yes, it should be in that way!</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the assignee for the bug.</li> </ul> </body> </html>