build failure with glibc-2.30

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

build failure with glibc-2.30

Eray Aslan-2
glibc-2.30 removed RES_INSECURE1, RES_INSECURE2 and RES_USE_INET6
symbols[1] resulting in:

dns_str_resflags.c:55:13: warning: RES_AAONLY is deprecated
   55 |     "RES_AAONLY", RES_AAONLY,
      |             ^~~~~~~~~~~~~~~~~        
dns_str_resflags.c:57:13: warning: RES_PRIMARY is deprecated
   57 |     "RES_PRIMARY", RES_PRIMARY,
      |             ^~~~~~~~~~~~~~~~~~~        
dns_str_resflags.c:63:22: error: ‘RES_INSECURE1’ undeclared here (not in a function); did you mean ‘RES_RECURSE’?
   63 |     "RES_INSECURE1", RES_INSECURE1,
      |                      ^~~~~~~~~~~~~
      |                      RES_RECURSE
dns_str_resflags.c:64:22: error: ‘RES_INSECURE2’ undeclared here (not in a function); did you mean ‘RES_RECURSE’?
   64 |     "RES_INSECURE2", RES_INSECURE2,
      |                      ^~~~~~~~~~~~~
      |                      RES_RECURSE
dns_str_resflags.c:66:22: error: ‘RES_USE_INET6’ undeclared here (not in a function)
   66 |     "RES_USE_INET6", RES_USE_INET6,
      |                      ^~~~~~~~~~~~~
dns_str_resflags.c:71:13: warning: RES_NOCHECKNAME is deprecated
   71 |     "RES_NOCHECKNAME", RES_NOCHECKNAME,
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~    
dns_str_resflags.c:76:13: warning: RES_KEEPTSIG is deprecated
   76 |     "RES_KEEPTSIG", RES_KEEPTSIG,
      |             ^~~~~~~~~~~~~~~~~~~~~      
dns_str_resflags.c:79:13: warning: RES_BLAST is deprecated
   79 |     "RES_BLAST", RES_BLAST,
      |             ^~~~~~~~~~~~~~~          

FYI
--
Eray

[1] https://sourceware.org/ml/libc-alpha/2019-08/msg00029.html

Reply | Threaded
Open this post in threaded view
|

Re: build failure with glibc-2.30

Wietse Venema
Eray Aslan:
> dns_str_resflags.c:63:22: error: ?RES_INSECURE1? undeclared here (not in a function); did you mean ?RES_RECURSE??
>    63 |     "RES_INSECURE1", RES_INSECURE1,

etc. Fixed in Postfix 3.5. You cn delete the offending lines.
Which distro are you building on?

        Wietse
Reply | Threaded
Open this post in threaded view
|

Re: build failure with glibc-2.30

Eray Aslan-2
On Mon, Aug 19, 2019 at 08:51:47AM -0400, Wietse Venema wrote:
> Eray Aslan:
> > dns_str_resflags.c:63:22: error: ?RES_INSECURE1? undeclared here (not in a function); did you mean ?RES_RECURSE??
> >    63 |     "RES_INSECURE1", RES_INSECURE1,
>
> etc. Fixed in Postfix 3.5. You cn delete the offending lines.
> Which distro are you building on?

That was on gentoo.  Sorry didnt think of checking on postfix-3.5.

--
Eray
Reply | Threaded
Open this post in threaded view
|

Re: build failure with glibc-2.30

Eliza
HI

on 2019/8/20 14:23, Eray Aslan wrote:
>> etc. Fixed in Postfix 3.5. You cn delete the offending lines.
>> Which distro are you building on?
> That was on gentoo.  Sorry didnt think of checking on postfix-3.5.

Just be curious for what reasons we choose to build a software from
sources? today it's very convenient to use package management tools to
install linux software.

regards.
Reply | Threaded
Open this post in threaded view
|

Re: build failure with glibc-2.30

Bill Cole-3
On 20 Aug 2019, at 2:29, Eliza wrote:

> HI
>
> on 2019/8/20 14:23, Eray Aslan wrote:
>>> etc. Fixed in Postfix 3.5. You cn delete the offending lines.
>>> Which distro are you building on?
>> That was on gentoo.  Sorry didnt think of checking on postfix-3.5.
>
> Just be curious for what reasons we choose to build a software from
> sources? today it's very convenient to use package management tools to
> install linux software.

1. Postfix isn't just for Linux.

2. Even for Linux, packaging is variable. e.g. Gentoo's Portage package
is mostly used in a build from source mode.

3. Many people want a stable mature base platform (e.g EL7 or even EL6 )
but want to be more current with critical functionality, like the MTA on
a production mail server.

4. Some people prioritize the security advantages of bespoke minimized
systems with known build conditions over those of a system with builds
designed for the broadest possible audience and maintained by an expert
team. This is more commonly seen in the BSD world, but it is not
infeasible with Linux.



--
Bill Cole
[hidden email] or [hidden email]
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)