Re: Windows SSPI/Schannel winbuild makefiles
Date: Thu, 14 Jun 2012 17:59:36 +0200
2012/6/14 Yang Tse <yangsita_at_gmail.com>:
> Although I would like to comment just a couple of things relative to
> the description of the patch
> 1) Even being true that a patch fixes some one's incomplete or
> problematic code, as a general rule we don't mention whose code or
> whatever the patch fixes. We try to be as aseptic as possible and only
> mention what it fixes and if we deem necessary we mention the hash of
> the problematic commit.
> Can you imagine what would happen if I blamed in each fix or
> adjustment I make the name of the person whose code is being fixed? I
> don't even want to imagine it.
Understood, even though I didn't want to "blame" you for anything.
>> The patch does also include some comments on the general define naming
>> strategy I would like to recommend for the Windows makefiles.
> 2) If you feel so, you could bring that on the list and comment with
> others. I'm not saying I'm against what you propose there for the
> 'winbuild' system. I'm only saying that the comment will most likely
> be missed in the future.
Understood, so here is the original recommendation I made in the
Since the makefiles are currently switching between ENABLE_, WITH_
and USE_ prefix I would like to propose the following strategy for
- WITH_ prefix: External define with custom path or mode support
- ENABLE_ prefix: External define with yes or no values
- USE_ prefix: Internal define with true or false value
USE_ should also probably be the only define passed around from
winbuild/Makefile.vc to winbuild/MakefileBuild.vc
> Once I've said those things. I ask you to tell me if you want this
> patch pushed as-is or if you want to provide an updated one. Either
> way I have no problem.
Sure, attached is a patch that does not include your name, but only
the commit hash and also does not include the recommendation above.
- application/octet-stream attachment: 0001-winbuild-Allow-SSPI-build-with-or-without-Schannel.patch