On Sun, 2025-09-28 at 16:27 +0200, sj wrote:
> Hello,
>
> I'm new here, but it seems like there's a bug in the configure file
> (I
> was just looking for the same thing) that means that libatril is only
> used with GTK+3 if --enable-evince is given but missing from the
> build
> environment (there's also a typo in the PKG_CHECK_MODULES macro for
> evince2_32). This patch fixes what I could find, typesetting is still
> broken for some reason, but at least it compiles. Unfortunately, I
> couldn't find every single place where atril checks are dependent on
> the
> useevince variable, so configuring with --enable-evince, then remove
> the
> dev package is still necessary.
>
> --- a/configure.ac 2025-09-28 13:05:57535618217 +0200
> +++ b/configure.ac 2025-09-28 13:10:51364641998 +0200
> @@ -275,7 +275,7 @@
> fi
>
> if test "x$have_evince2_32" = "xno"; then
> - PKG_CHECK_MODULES(ATRIL, atril-view-1.5.0 >= 1.5
> have_atril=yes,
> have_atril=no)
> + PKG_CHECK_MODULES(ATRIL, atril-view-1.5.0 >= 1.5,
> have_atril=yes,
> have_atril=no)
> CFLAGS="$CFLAGS $ATRIL_CFLAGS"
> LIBS="$LIBS $ATRIL_LIBS"
> fi
> @@ -303,13 +303,12 @@
> fi
> fi
>
> -if test "x$useevince" = "xyes"; then
> +if test "x$have_atril" = "xyes"; then
> + CFLAGS="$CFLAGS -DUSE_ATRIL"
> + LIBS="$LIBS $ATRIL_LIBS"
> +elif test "x$have_evince2" = "xyes" -o "x$have_evince2_3" = "xyes" -
> o
> "x$have_evince2_32" = "xyes" -o "x$have_evince_3_0" = "xyes"; then
> CFLAGS="$CFLAGS -DUSE_EVINCE"
> LIBS="$LIBS -DUSE_EVINCE"
> - if test "x$have_atril" = "xyes"; then
> - CFLAGS="$CFLAGS -DUSE_ATRIL"
> - LIBS="$LIBS $ATRIL_LIBS"
> - fi
> fi
>
> AC_ARG_ENABLE(
> --- a/src/printview/svgview.c 2025-09-07 14:13:44.000000000 +0200
> +++ b/src/printview/svgview.c 2025-09-28 15:54:12.532692174 +0200
> @@ -35,7 +35,5 @@
>
>
> #ifndef USE_EVINCE
> -void set_continuous_typesetting (gboolean setting){return FALSE;}
> -gboolean continuous_typesetting () {return FALSE;}
> #endif
>
> --- a/src/scripting/scheme-callbacks.c 2025-09-07 14:13:44.000000000
> +0200
> +++ b/src/scripting/scheme-callbacks.c 2025-09-28 16:01:53.998036258
> +0200
> @@ -38,5 +38,6 @@
> #include "printview/printview.h"
> #include "printview/markupview.h"
> #include "display/calculatepositions.h"
> +#include "display/displayanimation.h"
> #include "source/source.h"
> #include "source/sourceaudio.h"
>
> Not that sure about the third hunk, it would probably be better if it
> was dependent on both evince and atril but I don't know what that
> code's
> doing (maybe fallback SVG rendering?), and this at least didn't
> murder
> ld because of a duplicate definition.
As far as I can see that re-definition of set_continuous_typesetting*
is entirely spurious, so your hunk 3 is good.
> If someone knows a magic flag to
> get GTK+3 Denemo into a working condition,
What is not working? No Print View?
Richard
> please tell me, so far I
> could only break it more. In the meanwhile, I'll be looking into it.
>
> sj
>
> On 9/28/25 16:14, Andreas Schneider wrote:
> > Thanks Jeremiah.
> >
> > When compiling with --disable-evince, I get the error
> >
> > pageturner.c:28:10: fatal error: evince-view.h: Datei oder
> > Verzeichnis
> > nicht gefunden
> > 28 | #include <evince-view.h>
> > | ^~~~~~~~~~~~~~~
> >
> > The reason seems to be that the constant USE_ATRIL is not defined.
> > Yet I
> > do not see which configure option defines it.
> >
> > I can compile with evince after applying the patch mentioned in my
> > last
> > email.
> >
> > Andreas
> >
>
>
