Re: [Bug-wget] Wget 1.17 doesn't compile on Windows (hsts.c)

[Tim Ruehsen]

Hi Dagobert,

On Tuesday 17 November 2015 12:16:15 Dagobert Michelsen wrote:
> Hi Tim,
> 
> Am 17.11.2015 um 10:36 schrieb Tim Ruehsen <address@hidden>:
> > On Monday 16 November 2015 22:34:53 Dagobert Michelsen wrote:
> >> Same here on Solaris 10 with Oracle Studio 12:
> >>> /opt/csw/bin/gcc -DHAVE_CONFIG_H
> >>> -DSYSTEM_WGETRC=\"/usr/local/etc/wgetrc\"
> >>> -DLOCALEDIR=\"/usr/local/share/locale\" -I.  -I../lib -I../lib
> >>> -D_REENTRANT  -I/opt/csw/include   -I/opt/csw/include -I/opt/csw/include
> >>> -I/opt/csw/include -I/opt/csw/include/p11-kit-1   -DHAVE_LIBGNUTLS
> >>> -I/opt/csw/include   -I/opt/csw/include   -DNDEBUG  -MT hsts.o -MD -MP
> >>> -MF .deps/hsts.Tpo -c -o hsts.o hsts.c
> >>> 
> >>> hsts.c: In function 'hsts_store_save':
> >>> hsts.c:505:22: error: 'LOCK_EX' undeclared (first use in this function)
> >>> 
> >>>           flock (fd, LOCK_EX);
> >>>           
> >>>                      ^
> >>> 
> >>> hsts.c:505:22: note: each undeclared identifier is reported only once
> >>> for
> >>> each function it appears in
> >>> 
> >>> Makefile:1573: recipe for target 'hsts.o' failed
> >> 
> >> Also the builds are failing for quite some time:
> >>  https://buildfarm.opencsw.org/buildbot/waterfall?category=wget
> >> 
> >> Unfortunately I didn’t get any feedback at all when I mentioned compiling
> >> issues: http://lists.gnu.org/archive/html/bug-wget/2015-07/msg00068.html
> > 
> > This is still open here as an issue (but with low priority).
> > We are definitely not enough developers here…
> 
> Ok.
> 
> > These failures might be simple things like a missing include directive.
> > It would would be very helpful if you could just tell us "hey, adding this
> > line ... fixes it". Thats 10 minutes for you and 10 for us, and voila.
> 
> Unfortunately it is not that easy as Solaris does not have „flock“:
>   http://www.perkin.org.uk/posts/solaris-portability-flock.html
> The requested solution would be to use the POSIX-compliant fcntl(),
> at least on Solaris.
> 
> I’d be happy to implement a fix, just let me know which version you prefer:
> (1) check for flock and fcntl, prefer flock if available
>     (least invasive to current mode)
> (2) Use #ifdef for LOCK_EX and use fcntl instead
> (3) always use fcntl
> 
> I was worried my fix wouldn’t be good enough for direct inclusion...

I guess, Darshit's patch also fixes this issue for Solaris.

I was more referring to
> issues: http://lists.gnu.org/archive/html/bug-wget/2015-07/msg00068.html

It seems that metalink.c has to include <errno.h> instead <sys/errno.h>.
Maybe could could give it try or otherwise let us know how to fix it for 
Solaris.

> > But if compilation fails within your build farm, couldn't you send a mail
> > to address@hidden ? I know you already suggested something like this…
> Sure, but I didn’t want to set this up without asking to not spam the list
> with stuff nobody wants to read :-) To minimize noise I can send email
> only when a build transitions from good to bad or vice versa.

"Only on build transitions" sounds good to me.
Please feel free to set notifications up that way.

Regards, Tim