On Fri, Jul 04, 2014 at 10:09:32PM +0200, Kurt Roeckx via RT wrote: > On Fri, Jul 04, 2014 at 08:33:52PM +0200, Dan Fulger via RT wrote: > > The attached patch and notes apply to 1.0.1h. > > > > OS/400 fixes since George's port: unit tests now work better, X509 strings > > are now correctly encoded in generated > > certificates when not using prompt mode, a dependency to the OS/400 > > secure random numbers library > > was added, UTF8 strings in certificates no longer print garbage (but > > other Unicode encodings are not fixed). > > > > Still to be fixed: > > - tsa test shows "bad time value" when printing but otherwise works > > - cms verification does not work > > Such a big patches with all kind of changes is really hard to > review. It's a 4000 line patch. Could you split it in smaller > patches? > > So doing a quick look at the types of things I think most of it is > related to EBCDIC. I see things like: > - Open binary files in binary mode, so that the OS doesn't do a > charset conversion. > - Use numbers instead of characters where we are expecting ASCII > input and are checking the characters to be valid. > - In tests we do things like encoding or hashing of a string and > expect the output to be based on the ASCII version of the > string, but it's actually compiled as an EBCDIC string and so > fails.
So someone was already working on the first binary/text mode thing since this actually affects multiple OSs. I don't like the approach taking both in the current code and the patch for the other issues. But I don't have a better idea currently. We will try and look at this again at a later time. Kurt ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [email protected] Automated List Manager [email protected]
