This is the mail archive of the newlib@sourceware.org mailing list for the newlib project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
On Aug 27 14:31, Jeff Johnston wrote:Well sure, if one wants to be practical :) Patch checked in.
Corinna Vinschen wrote:
I tracked this down to a problem in the function _wcsrtombs_r.Thanks for looking at this Corinna.
The condition to write a converted character into the destination string and to continue the loop is (wcsrtombs.c, line 48):
if (n <= len - bytes && bytes < len)
The second half of this condition is wrong, AFAICS. First of all, it should be bytes <= len instead of bytes < len. Second, this part of the condition is already catched by the first part of the condition:
n <= len - bytes <==> n + bytes <= len
with n always >= 0, bytes is always <= len if n + bytes is <= len. So the condition is redundant.
Actually, the second part is needed because the first equation is unsigned. Let's say len is 1 but bytes is 2. len - bytes is (unsigned)-1. We don't want to write at this point. Thus, we make the additional check that bytes is actually <= len. I just modified your change and checked it in.
Thank you. But... wouldn't
if (n + bytes <= len)
have the same result?
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |