[PATCH 0/1] MR52: configure: Allow linking against static SPIRV-Tools.

[Zebediah Figura (＠zfigura)]

Shared SPIRV-Tools is not shipped by Debian Bullseye, and is broken if built manually.

[PATCH 1/1] configure: Allow linking against static SPIRV-Tools.

[Zebediah Figura]

From: Zebediah Figura

Shared SPIRV-Tools is not shipped by Debian Bullseye, and is broken if built manually.
---
 configure.ac | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/configure.ac b/configure.ac
index 577b3983..3a491a49 100644
--- a/configure.ac
+++ b/configure.ac
@@ -122,11 +122,19 @@ AS_IF([test "x$with_ncurses" != "xno"],
       [PKG_CHECK_MODULES([NCURSES], [ncurses],
                          [AC_DEFINE([HAVE_NCURSES], [1], [Define to 1 if you have ncurses.]) with_ncurses=yes],
                          [with_ncurses=no])])
+
+HAVE_SPIRV_TOOLS=no
 AS_IF([test "x$with_spirv_tools" = "xyes"],
       [PKG_CHECK_MODULES([SPIRV_TOOLS], [SPIRV-Tools-shared],
-      [AC_DEFINE([HAVE_SPIRV_TOOLS], [1], [Define to 1 if you have SPIRV-Tools.])])],
+                         [HAVE_SPIRV_TOOLS=yes],
+                         [PKG_CHECK_MODULES([SPIRV_TOOLS], [SPIRV-Tools],
+                                            [HAVE_SPIRV_TOOLS=yes
+                                             AC_SUBST([SPIRV_TOOLS_LIBS],["$SPIRV_TOOLS_LIBS -lstdc++"])])])],
       [with_spirv_tools=no])

+AS_IF([test "$HAVE_SPIRV_TOOLS" = "yes"],
+      [AC_DEFINE([HAVE_SPIRV_TOOLS], [1], [Define to 1 if you have SPIRV-Tools.])])
+
 HAVE_XCB=no
 AS_IF([test "x$with_xcb" != "xno"],
       [PKG_CHECK_MODULES([XCB], [xcb xcb-event xcb-icccm xcb-keysyms],

[PATCH 1/1] configure: Allow linking against static SPIRV-Tools.

[Marvin]

Hi,

It looks like your patch introduced the new failures shown below.
Please investigate and fix them before resubmitting your patch.
If they are not new, fixing them anyway would help a lot. Otherwise
please ask for the known failures list to be updated.

The full results can be found at:
https://testbot.winehq.org/JobDetails.pl?Key=126483

Your paranoid android.

=== debian11 (build log) ===

error: patch failed: configure.ac:122
Task: Patch failed to apply

=== debian11b (build log) ===

error: patch failed: configure.ac:122
Task: Patch failed to apply

[Henri Verbeet (＠hverbeet)]

> Shared SPIRV-Tools is not shipped by Debian Bullseye, and is broken if built manually.

Is that a regression? I'm using a build of shared SPIRV-Tools I made myself, and it seems to work fine, although it's an older (v2019.5) version.

We may also want to consider simply getting rid of the SPIRV-Tools dependency. Parsing SPIR-V isn't that complicated, the SPIRV-Tools dependency has always been a bit of a pain, and it would simplify things a bit for Wine's usage of vkd3d.

[Giovanni Mascellani (＠giomasce)]

On Tue Nov 22 12:25:47 2022 +0000, Henri Verbeet wrote:
> > Shared SPIRV-Tools is not shipped by Debian Bullseye, and is broken if
>
> built manually.
> Is that a regression? I'm using a build of shared SPIRV-Tools I made
> myself, and it seems to work fine, although it's an older (v2019.5) version.
> We may also want to consider simply getting rid of the SPIRV-Tools
> dependency. Parsing SPIR-V isn't that complicated, the SPIRV-Tools
> dependency has always been a bit of a pain, and it would simplify things
> a bit for Wine's usage of vkd3d.

Last time I tried I could build vkd3d with SPIRV-Tools by just editing the configuration line in the Debian package and rebuilding it, and then rebuilding vkd3d. I saw no breakages.

Though since the shared library is not available on Debian (and, I suppose, on many derivatives) by default, it makes sense to me to use it statically.

It would be even better if SPIRV-Tools was available as a shared library, but that's not for us to decide.

[Zebediah Figura (＠zfigura)]

> > Shared SPIRV-Tools is not shipped by Debian Bullseye, and is broken if built manually.
>
> Is that a regression? I'm using a build of shared SPIRV-Tools I made myself, and it seems to work fine, although it's an older (v2019.5) version.

I'll have to check, maybe it did break at some point.
> We may also want to consider simply getting rid of the SPIRV-Tools dependency. Parsing SPIR-V isn't that complicated, the SPIRV-Tools dependency has always been a bit of a pain, and it would simplify things a bit for Wine's usage of vkd3d.

I guess. I'd rather submit this patch than spend time writing a SPIR-V parser, though. We'd also lose validation that way.

Unless we were to import into the source; that seems decidedly dispreferable, but it's not the first time I've tried to express that preference.

[Henri Verbeet (＠hverbeet)]

> > We may also want to consider simply getting rid of the SPIRV-Tools dependency. Parsing SPIR-V isn't that complicated, the SPIRV-Tools dependency has always been a bit of a pain, and it would simplify things a bit for Wine's usage of vkd3d.
>
> I guess. I'd rather submit this patch than spend time writing a SPIR-V parser, though. We'd also lose validation that way.

True. The point is mostly that SPIRV-Tools never was a particularly great fit for vkd3d-shader; we use it because it's what's available. I'm not aware of better alternatives, but we'd likely switch if we found one. Similarly, if someone were to contribute a SPIR-V parser to vkd3d-shader, there's a good chance it would make it upstream.
> Unless we were to import into the source; that seems decidedly dispreferable, but it's not the first time I've tried to express that preference.

I'd rather not do that either, no.