On Sat, Jul 5, 2014 at 4:03 PM, Steven M. Schweda <s...@antinode.info> wrote:
> From: "Salz, Rich" <rs...@akamai.com> > > > There's a bunch of hacks in apps/openssl.c to work around some old VMS > rele= > > ases; the coment is dated 2011-03-22. > > I am going to delete it. > > Thanks for the notice. > > Why? Is your primary goal to remove all support for VMS, or only to > cause defective behavior on most/all Alpha systems when 64-bit pointers > are used? Does this code cause any performance or other portability > problems? Is the comment too confusing? Does this (rather small, I > thought) "bunch of hacks" fluff up the source code excessively? If > _you_ don't need it, then it's not worth keeping? > > > Speak up now if you can justify keeping it. > > I know nothing, but I'd guess that few VMS users follow this forum, > so you may be unlikely to get a response here from anyone (else). The > problem being worked around is not confined to "old VMS releases", it's > in a particular compiler (or set of compilers), which. I believe, may > still be widely used (for appropriate values of "widely"). Yes, HP > should provide a compiler without this bug, and all users should obtain > and use it. This, I fear, is a low-probability outcome. > > If the Voices (or even wise judgement) tell you that this ugly blot > on the OpenSSL code must be purged, then you might consider replacing it > with a comment which explains the original problem, the work-around, and > the last release which still contained it, so that any poor slob who is > affected by it in the future will be able to find easily the (hard-won) > solution after you discard it. > > ------------------------------------------------------------------------ > > Steven M. Schweda sms@antinode-info > 382 South Warwick Street (+1) 651-699-9818 > Saint Paul MN 55105-2547 > ______________________________________________________________________ > OpenSSL Project http://www.openssl.org > Development Mailing List openssl-dev@openssl.org > Automated List Manager majord...@openssl.org > What about replacing it all with something like #if defined(OPENSSL_SYS_VMS) int app_main(int Argc, char *Argv[]) /* arguments may need to be munged depending on build */ #else int main(int Argc, char *argv[]) #endif and somebody that cares and can actually build/test on affected systems needs to make it work before 1.0.3 is released. Version control has the old code. The code gets decrapified by just a little bit and VMS-savants have guidance on what they need to do.