On Wed, 19 Oct 2016, Anton Khirnov wrote:

It didn't take into account the new pattern used for bitstream filters
and protocols.
---
Makefile | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/Makefile b/Makefile
index 34f0c9c..d76173c 100644
--- a/Makefile
+++ b/Makefile
@@ -121,8 +121,14 @@ $(TOOLS): %$(EXESUF): %.o $(EXEOBJS)

tools/cws2fws$(EXESUF): ELIBS = $(ZLIB)

+CONFIGURABLE_COMPONENTS =                                           \
+    $(wildcard $(FFLIBS:%=$(SRC_PATH)/lib%/all*.c))                 \
+    $(SRC_PATH)/libavformat/protocols.c                             \
+    $(SRC_PATH)/libavcodec/bitstream_filters.c                      \
+
+
config.h: .config
-.config: $(wildcard $(FFLIBS:%=$(SRC_PATH)/lib%/all*.c))
+.config: $(CONFIGURABLE_COMPONENTS)
        @-tput bold 2>/dev/null
        @-printf '\nWARNING: $(?F) newer than config.h, rerun configure\n\n'
        @-tput sgr0 2>/dev/null
--
2.0.0

Thanks, this is a step in the right direction. I sometimes ignore these warnings though, e.g. if checking out versions back and forth (jumping across a version where e.g. a new decoder was added), I just configure with the latest version and build despite the warnings about potentially requiring reconfiguration. If config.h/config.mak are too old, I'd get a build error in e.g. allcodecs.c about CONFIG_XYZ_DECODER being undefined.

I thought a little about a way to extend e.g. bitstream_filters.c and bsf_list.c to get this effect back, but didn't come up with something sensible that doesn't require duplicating the list somehow. (E.g. bitstream_filters.c would have some sort of reference to all config items, that bsf_list.c and/or config.h would supply, and cause a failure if missing, but this would at least require adding one more line per item in bitstream_filters.c.)

So ideally I'd want to get that effect back, but until then, I'll just try to reconfigure more often if warned about it.

// Martin
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to