[Slackbuilds-users] Updates - 20170923.1

[Duncan Roe]

Hi Willy,

I just noticed you're the maintainer ;)

On Sun, Sep 24, 2017 at 07:29:01PM +1000, Duncan Roe wrote:
> Hi Dave,
>
> On Sat, Sep 23, 2017 at 05:06:17PM +0100, David Spencer wrote:
> > Hello everybody,
> >
> > This week's public update is a bit later than usual, because of our
> > personal schedules :)
> >
> ...
> > libraries/webkit2gtk: Fix build issue on single job.
> ...
>
> It seems you need this:
>
> diff --git a/libraries/webkit2gtk/webkit2gtk.SlackBuild b/libraries/webkit2gtk/webkit2gtk.SlackBuild
> index 57e4f2e..2f7c42e 100644
> --- a/libraries/webkit2gtk/webkit2gtk.SlackBuild
> +++ b/libraries/webkit2gtk/webkit2gtk.SlackBuild
> @@ -57,7 +57,7 @@ fi
>
>  # This is needed to fix build issue on single make job
>  # Thanks to ponce for investigating this issue.
> -MAKEFLAGS=${MAKEFLAGS:--j2}
> +export MAKEFLAGS=${MAKEFLAGS:--j2}
>
>  set -eu
>
>
> ------------------
>
> My build just failed after 18 hrs. It only ever ran 1 cc1plus.
>
> With the patch, it's running 2 of them,
>
> Cheers ... Duncan.

My re-build ran to completion with the above patch.

However, I do wonder whether "make -j2" is simply masking a bug in the make
process.

Perhaps you might like to report the failure to upstream. Here's what I found:

As Chris Willing noted in
https://lists.slackbuilds.org/pipermail/slackbuilds-users/2017-September/019991.html,
sources from the tarball #include WebCore/BackingStoreBackendCairoImpl.h whereas
the file exists at
webkitgtk-2.18.0/Source/WebCore/platform/HysteresisActivity.h.

The successful build generates
webkitgtk-2.18.0/build/DerivedSources/ForwardingHeaders/WebCore/BackingStoreBackendCairoImpl.h
which has the line

#include "WebCore/platform/graphics/cairo/BackingStoreBackendCairoImpl.h"

(i.e. where the file really is). The single-job make doesn't even get to create
the directory webkitgtk-2.18.0/build/DerivedSources/ForwardingHeaders/WebCore/
before failing. Could this indicate an error in the build sequence?

Cheers ... Duncan.