pan_auto on termux (arm) gdb debug
fun fact - but directly pasting this log {just 29 kb} to gmail's mobile web mail interface resulted in page reload... so i copypasted it via pre-installed wps office. most suspecting imo audio_output_ptr = {0xf66050a0, 0xf66050d0, 0x0 <repeats 30 times> but may be those just pointers on one for each channel, so 0x0 is normal for unused channels
On Tue, 2021-05-18 at 17:59 +0300, Andrew Randrianasulu via Cin wrote:
fun fact - but directly pasting this log {just 29 kb} to gmail's mobile web mail interface resulted in page reload...
so i copypasted it via pre-installed wps office.
most suspecting imo
audio_output_ptr = {0xf66050a0, 0xf66050d0, 0x0 <repeats 30 times>
but may be those just pointers on one for each channel, so 0x0 is normal for unused channels
Andrew, if I understood correctly, you are trying to get CinGG to work on a specific arm64 device? Would such an CinGG also work on say qemu with fedora34 arm64? Or is is very specific? MatN
On Tuesday, May 18, 2021, mat via Cin <[email protected]> wrote:
On Tue, 2021-05-18 at 17:59 +0300, Andrew Randrianasulu via Cin wrote:
fun fact - but directly pasting this log {just 29 kb} to gmail's mobile web mail interface resulted in page reload...
so i copypasted it via pre-installed wps office.
most suspecting imo
audio_output_ptr = {0xf66050a0, 0xf66050d0, 0x0 <repeats 30 times>
but may be those just pointers on one for each channel, so 0x0 is normal for unused channels
Andrew, if I understood correctly, you are trying to get CinGG to work on a specific arm64 device? Would such an CinGG also work on say qemu with fedora34 arm64? Or is is very specific?
well, more like arm8 / (cortex-a53) (all userspace here is 32-bit, not sure if cpu can do 64 bit? ) i think usual glibc-based arm(64) distro should be even easier as compilation environment.. just add your case into guicast/Makefile: diff --git a/cinelerra-5.1/guicast/Makefile b/cinelerra-5.1/guicast/Mak efile index cf189dea..fc086f92 100644 --- a/cinelerra-5.1/guicast/Makefile +++ b/cinelerra-5.1/guicast/Makefile @@ -12,6 +12,11 @@ ifeq ($(OBJDIR), amd64) BOOTSTRAPFLAGS := -DBOOTSTRAP="\"objcopy -B i386 -I binary -O elf64-x8 6-64\"" endif +ifeq ($(OBJDIR), armv8l) +BOOTSTRAPFLAGS := -DBOOTSTRAP="\"objcopy -B arm -I binary -O elf32-littlearm\"" +endif + + for arm64 -O probably will be elf64-littlearm.. and put
CFLAGS += -DNO_BTRACE CFLAGS += -DNO_CTX CFLAGS += -DNO_GDB { mostly for bcsignals.C } in global_config after configure but before make (i still not figured out how to put this in configure.ac depending on arch.. there already exist x86(-64) test for adding or not adding nasm, so i probably can just invert this logic? ) some libs in thirdparty may need their config.guess updated before they recognize relatively new architecture.. i just copied system's (automake?) config.guess into each dir manually when i was building for e2k... but most likely thirdparty/Makefile can be used for automating this.. +libogg.cfg_vars?= autoreconf -ifv -I m4 && automake -caf; libogg.cfg_params?= --enable-shared=no i'll attach my e2k patch, only as reference... (some __i386__ ifdefs are correct, but some missing relative to my current set of termux/arm patches....) good luck! (-dev libs + build itself may take nearly 4gb, at least my termux install weights some more than 4gb already.. but i have some extras like gtk3 installed) MatN
-- Cin mailing list [email protected] https://lists.cinelerra-gg.org/mailman/listinfo/cin
participants (2)
-
Andrew Randrianasulu -
mat