[help needed][packaging] add pkgconfig

[djdisodo]

i'm not sure where should i put these codes can i get some feedback?

tested on windows(using mingw)

[tencent-adm]

All committers have signed the CLA.

[nihui]

put ncnn.pc.in and cmake rules in src directory

install to ${CMAKE_INSTALL_LIBDIR}/pkgconfig

prefix=@CMAKE_INSTALL_PREFIX@

[codecov-commenter]

Codecov Report

Merging #3984 (1cee058) into master (43912c7) will decrease coverage by 0.96%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #3984      +/-   ##
==========================================
- Coverage   93.89%   92.92%   -0.97%     
==========================================
  Files         728      675      -53     
  Lines      177006   159274   -17732     
==========================================
- Hits       166200   148007   -18193     
- Misses      10806    11267     +461     

Impacted Files	Coverage Δ
src/layer/riscv/flatten_riscv.cpp	49.55% <0.00%> (-45.25%)	⬇️
src/layer/riscv/packing_riscv.cpp	60.51% <0.00%> (-27.55%)	⬇️
src/layer/riscv/crop_riscv.cpp	78.57% <0.00%> (-18.69%)	⬇️
src/layer/riscv/convolution_riscv.cpp	81.40% <0.00%> (-17.56%)	⬇️
src/command.cpp	74.60% <0.00%> (-13.05%)	⬇️
src/layer/riscv/padding_riscv.cpp	83.09% <0.00%> (-11.37%)	⬇️
src/layer/riscv/convolutiondepthwise_riscv.cpp	88.54% <0.00%> (-8.68%)	⬇️
src/layer/riscv/innerproduct_riscv.cpp	93.38% <0.00%> (-6.23%)	⬇️
src/layer/riscv/deconvolutiondepthwise_riscv.cpp	93.15% <0.00%> (-5.14%)	⬇️
src/layer/riscv/concat_riscv.cpp	90.47% <0.00%> (-5.11%)	⬇️
... and 185 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0de869e...1cee058. Read the comment docs.

[djdisodo]

prefix=@CMAKE_INSTALL_PREFIX@

this will hardcode absolute path of install prefix when it's builded
so it won't work if blobs aren't in their original path

which is not ideal especially when we are providing prebuilt releases

[nihui]

LGTM

[nihui]

Thanks for your contribution !