Re: tlsdate port to NetBSD?

Thomas Klausner:
> On Thu, Apr 11, 2013 at 08:09:56AM +0000, Jacob Appelbaum wrote:
>> I've recently been working on porting my tlsdate (
>> https://github.com/ioerror/tlsdate ) program to various operating
>> systems. NetBSD is one such system that I'd like to ensure is working
>> before I tag my next release (0.0.7). The port to FreeBSD 10-CURRENT
>> (and OS X 10.8.x) seems to work for a few users and thus I think most of
>> the BSD heavy lifting should be handled.
>>
>> I suspect the specific build conditions that check for FreeBSD could
>> easily modified to also check for NetBSD. With such a change, I guess it
>> could work out of the box. Ideally, I'd like to explicitly support
>> NetBSD and not just accidentally support it as a side effect of
>> supporting FreeBSD.
> 
> Thank you :)

Sure! I love NetBSD, who doesn't?

> 
>> I wonder if the best bet is to pick NetBSD 6.0.1-RC2 and then work on a
>> pkgsrc targeting the current git tip of tlsdate master?
> 
> Sounds good.

Ok.

> 
>> If anyone has already managed to port tlsdate to NetBSD, I'd love to
>> know before I set out to port it. :)
> 
> I've just tried the git version on NetBSD-current with some packages
> installed.
> 
> git clone https://github.com/ioerror/tlsdate
> autoreconf -fiv
> ./configure
> (output attached)
> make
> 

What packages were needed? Did you find that ./autogen.sh didn't work
for you or did you run a`utoreconf -fiv` for another reason?

> and wondered that the make output is so small:
> make  all-am
>   CCLD     src/compat/libtlsdate_compat.la
> 
> Then I tried 'make check':
> make  check-am
> make         src/test/proxy-override src/test/return-argc  src/test/rotate src/test/sleep-wrap
>   CC       src/test/proxy-override.o
>   CCLD     src/test/proxy-override
>   CC       src/test/return-argc.o
>   CCLD     src/test/return-argc
>   CC       src/test/rotate.o
>   CCLD     src/test/rotate
>   CC       src/test/sleep-wrap.o
>   CCLD     src/test/sleep-wrap
> make  check-TESTS
>   CCLD     src/conf_unittest
> gcc: no input files
> *** Error code 1
> 
> 'make check V=1' then gives:
> make  check-TESTS
> /bin/bash ./libtool  --tag=CC    --mode=link gcc "-D__STDC_LIMIT_MACROS" -g -O1 -Wall -fno-strict-aliasing -D_FORTIFY_SOURCE=2 -fstack-protector-all -fwrapv -fPIE -Wstack-protector --param=ssp-buffer-size=1   -L/lib -o src/conf_unittest   
> libtool: link: gcc -D__STDC_LIMIT_MACROS -g -O1 -Wall -fno-strict-aliasing -D_FORTIFY_SOURCE=2 -fstack-protector-all -fwrapv -fPIE -Wstack-protector --param=ssp-buffer-size=1 -o src/conf_unittest  -L/lib
> gcc: no input files
> *** Error code 1


That makes sense - it likely needs to have TARGET_NETBSD explicitly or
one could define TARGET_FREEBSD to test it.

> 
> Why do you have "if TARGET_FREEBSD" etc. instead of using feature
> test( result)s?
> 

I'm open to patches that make this process less painful. Fabian from the
FreeBSD project helped with the port - so I followed his lead in how it
was implemented. For the Mac OS X port, I just decided to make some of
it work first rather than blocking on everything working.

> So at least some more work is definitely needed :)
> Let me know if I can help.

Do you want to lend me a shell? If so, I'll gladly send you a public key
and get hacking on the port.

All the best,
Jacob


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter