Attached is a patch which fixes the ifdefs. I'll spend a bit of time reworking the OpenSSL patch and resubmit it in a new thread.
On Thu, Oct 28, 2021 at 6:33 PM Eric Molitor <emoli...@molitor.org> wrote: > It's very contrarian but I like ifdefs indented like this. But this isn't > my code base so let me rebase the OpenSSL changes and fix the indenting to > match the project style. > > LibTLS is basically a very thin wrapper which constrains how you hold an > SSL library, essentially making it very hard to hold it in the wrong way. > As a nice benefit it's become a common API for abstracting away an SSL > implementation (LibreSSL, OpenSSL, BearSSL, etc) > > The OpenSSL implementation here is using all of the defaults, in other > words you are assuming OpenSSL is using sane defaults which has, > historically, not been the case. But for more recent versions of OpenSSL > (1.1.1b and higher) is reasonably safe. This is also using the BIO > abstraction which is the simplest "modern" way to hold OpenSSL correctly > that I know of. > > - Eric > > On Thu, Oct 28, 2021 at 6:12 PM Rob Landley <r...@landley.net> wrote: > >> On 10/28/21 9:57 AM, Eric Molitor wrote: >> > Ok this annoyed me so I just added direct OpenSSL support parallel to >> the LibTLS >> > support. Elliot, the OpenSSL version should work with the latest >> versions of >> > BoringSSL that Android is using. >> >> I applied the previous one (the v2 with one library), and I know this is >> pending, but... you have indented #ifdefs? >> >> What's the difference between using the openssl api and using the libtls >> api? >> >> Rob >> >
0001-wget-Make-ifdefs-sane-for-other-people.patch
Description: Binary data
_______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net
