Re: CVS commit: pkgsrc/mk/check

[Ryo ONODERA <ryo%tetera.org@localhost>]

Hi,

From: Jonathan Perkin <jperkin%joyent.com@localhost>, Date: Tue, 3 Oct 2017 14:00:58 +0100

> * On 2017-10-03 at 13:56 BST, Ryo ONODERA wrote:
> 
>> This commit enables SSP check for PKG_DEVELOPER.
>> However my NetBSD/amd64-current does not have SSP support.
>> At least my environment has no libssp.so.
>> Maybe any older NetBSD releases do not have SSP support too.
>> 
>> So almost all packages including pkgtools/digest failed in package phase.
>> 
>> Could you take a look at this problem?
> 
> I'm a bit confused as to how this can happen, the check is for
> PKG_DEVELOPER and _PKGSRC_USE_SSP=yes, and _PKGSRC_USE_SSP is only set
> to yes if both the user requests SSP and the OS supports it.
> 
> What are all the relevant variables set to in your environment?

I have PKG_DEVELOPER=yes in my /etc/mk.conf.
no _PKGSRC_USE_SSP in my /etc/mk.conf.

$ cd /usr/pkgsrc/pkgtools/digest
$ make show-vars VARNAMES=PKGSRC_USE_SSP
yes

This 'yes' is from /usr/pkgsrc/mk/defaults/mk.conf.


$ cd /usr/pkgsrc/pkgtools/digest
$ make show-vars VARNAMES=_PKGSRC_USE_SSP
yes

This 'yes' is from /usr/pkgsrc/mk/bsd.prefs.mk.
Because PKGSRC_USE_SSP==yes and _OPSYS_SUPPORTS_SSP!=no.
_OPSYS_SUPPORTS_SSP=yes is from /usr/pkgsrc/mk/platform/NetBSD.mk.

I feel that determining _OPSYS_SUPPORTS_SSP with 
OPSYS and MACHINE_ARCH is bad idea.

Thank you.

> -- 
> Jonathan Perkin  -  Joyent, Inc.  -  www.joyent.com

--
Ryo ONODERA // ryo%tetera.org@localhost
PGP fingerprint = 82A2 DC91 76E0 A10A 8ABB  FD1B F404 27FA C7D1 15F3