Hi,
addr2line.c:457 has: char *name = NULL; if (sscanf (string, "(%m[^)])%" PRIiMAX "%n", &name, &addr, &i) == 2 && string[i] == '\0') parsed = adjust_to_section (name, &addr, dwfl); switch (sscanf (string, "%m[^-+]%n%" PRIiMAX "%n", &name, &i, &addr, &j))
And in the portability branch that got changed to: char *name = NULL; if (sscanf (string, "(%a[^)])%" PRIiMAX "%n", &name, &addr, &i) == 2 && string[i] == '\0') parsed = adjust_to_section (name, &addr, dwfl); switch (sscanf (string, "%a[^-+]%n%" PRIiMAX "%n", &name, &i, &addr, &j))
But I think that that doesn't make sense. Isn't %m the gnu extention, and that %a now follows the C99 standard and expects a float *?
Kurt
The older GNU extension is %as or %a[ and %ms/%m[ was added after C99 usurped %a. Newer GNU implementations support both in that %a is a flag if followed by s or [ and a specifier otherwise. Hence the code is correct for portability to both old and new GNU C libraries.
On Fri, Feb 24, 2012 at 01:41:09PM -0800, Roland McGrath wrote:
The older GNU extension is %as or %a[ and %ms/%m[ was added after C99 usurped %a. Newer GNU implementations support both in that %a is a flag if followed by s or [ and a specifier otherwise. Hence the code is correct for portability to both old and new GNU C libraries.
Then I assume it's a bug in gcc to give a warning about this?
Kurt
Then I assume it's a bug in gcc to give a warning about this?
It's a debatable question. I suppose gcc's format grokking could be made to vary based on -std= or something. I suspect they just won't want to change it, since it's only for compatibility with obsolete versions of things that anybody's source would use %as nowadays.
elfutils-devel@lists.fedorahosted.org