Terrence J. Doyle
2014-12-23 00:32:13 UTC
OK, I've managed to find fixes for some of the problems I reported
in my last message,
http://lists.research.att.com/pipermail/ast-users/2014q4/004686.html.
Along the way I discovered several more bugs, and I have fixes for some
of those, too. I'll put each into a separate message.
The first fix is for proto and its failure to build for
ast-open-2014-09-29. There are two competing redefinitions of error() in
proto.c, pperror() for standalone proto and _ast_error() to avoid
namespace collisions in normal proto. This is unique to proto which is
why the problem doesn't show up in other programs that use error().
Without additional cpp logic the standalone mechanism wins.
This problem is not unique to my system's architecture. It will
occur whenever error() is redefined as _ast_error().
Attached is proto.c.diff to fix the problem.
Terrence Doyle
in my last message,
http://lists.research.att.com/pipermail/ast-users/2014q4/004686.html.
Along the way I discovered several more bugs, and I have fixes for some
of those, too. I'll put each into a separate message.
The first fix is for proto and its failure to build for
ast-open-2014-09-29. There are two competing redefinitions of error() in
proto.c, pperror() for standalone proto and _ast_error() to avoid
namespace collisions in normal proto. This is unique to proto which is
why the problem doesn't show up in other programs that use error().
Without additional cpp logic the standalone mechanism wins.
This problem is not unique to my system's architecture. It will
occur whenever error() is redefined as _ast_error().
Attached is proto.c.diff to fix the problem.
Terrence Doyle