* meson.build dependencies PIC, seticor lib and dep definitions

[radonnachie]

This allows seticore to easily be included as a meson subproject.

[radonnachie]

With these changes I can submodule and compile seticore into BLADE.

[lacker]

Cool. Can you just verify that the unit tests run after this change? Run ‘unit_tests.sh’

[eli-schwartz]

For fmt, at least, I would actually suggest:

• use dependency('fmt'), as many people might already have it installed and can then use the one they already have
• delete the current subproject, and instead use meson wrap install fmt, which installs a pure-meson version of fmt from https://wrapdb.mesonbuild.com/ that handles this correctly in one line instead of four. :)

[radonnachie]

radonn@cosmic-gpu-0:/home/cosmic/dev/blade/subprojects/seticore$ ./unit_tests.sh 
unit testing...
===============================================================================
All tests passed (22 assertions in 6 test cases)

downloading h5 data for regression testing...
--2022-10-13 18:28:52--  https://bldata.berkeley.edu/pipeline/AGBT21B_999_31/blc17_blp17/blc17_guppi_59544_62191_HIP99317_0059.rawspec.0000.h5
Resolving bldata.berkeley.edu (bldata.berkeley.edu)... 208.68.240.101
Connecting to bldata.berkeley.edu (bldata.berkeley.edu)|208.68.240.101|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 3230416399 (3.0G) [application/octet-stream]
Saving to: ‘data/blc17_guppi_59544_62191_HIP99317_0059.rawspec.0000.h5’

blc17_guppi_59544_62191_HIP99317_0059.rawspec.0000.h5           100%[======================================================================================================================================================>]   3.01G  4.02MB/s    in 18m 26s 

2022-10-13 18:47:19 (2.78 MB/s) - ‘data/blc17_guppi_59544_62191_HIP99317_0059.rawspec.0000.h5’ saved [3230416399/3230416399]

regression testing...
welcome to seticore, version 0.2.10
running in dedoppler mode.
loading input from data/blc17_guppi_59544_62191_HIP99317_0059.rawspec.0000.h5
dedoppler parameters: max_drift=0.40 min_drift=0.0000 snr=10.00
writing output to data/testout.hits
hit: coarse channel = 23, index = 566231, snr = 10.06962, drift rate = -0.00000 (0 bins)
hit: coarse channel = 30, index = 384478, snr = 215.74893, drift rate = -0.00000 (0 bins)
hit: coarse channel = 30, index = 388533, snr = 10.81678, drift rate = -0.10204 (10 bins)
hit: coarse channel = 60, index = 418095, snr = 37.85838, drift rate = 0.08163 (-8 bins)
hit: coarse channel = 60, index = 436060, snr = 16.08596, drift rate = 0.02041 (-2 bins)
hit: coarse channel = 60, index = 438649, snr = 10.71509, drift rate = -0.05102 (5 bins)
hit: coarse channel = 60, index = 440394, snr = 23.78380, drift rate = -0.21429 (21 bins)
dedoppler elapsed time: 9s
diffing against expected output.
output matches. regression test looks good
radonn@cosmic-gpu-0:/home/cosmic/dev/blade/subprojects/seticore$ git show
commit c978e515f10b874eec9d3342676039da9a71442b (HEAD, mydonsol/as_subproject, as_subproject)
Author: RocketRoss <code@radonn.co.za>
.
.
.

[lacker]

Looks good to me, thanks!

[lacker]

I had to revert this change, more or less, in https://github.com/lacker/seticore/commit/829f32cbf0c2159e1bdfc6c746eeaa2a3e41174c - it's breaking meerkat production because the built binary, when copied to a different location during the deployment process, is now failing with

error while loading shared libraries: libfmt.so: cannot open shared object file: No such file or directory

To double-check you can just ensure the built binary continues to work when copied to a different location. I think this is probably related to the wrap stuff that Eli suggested rather than the library stuff, but it wasn't clear to me how to have one without the other.

[eli-schwartz]

CMake is defaulting to build a static fmt, probably. Because BUILD_SHARED_LIBS (global setting) defaults to off.

You can trivially do that with Meson too, simply configure Meson with -Dfmt:default_library=static, or more likely (so that you don't have to manually pass arguments to the build), pass default_options: ['default_library=static'] as a kwarg to the fmt dependency subproject lookup.

Either approach causes the fmt subproject to build its libraries as static archives instead of a DSO -- hence, fmt becomes included in the binary rather than requiring to load some *.so file.

(It is still worth using dependency('fmt') instead of subproject('fmt').get_variable('fmt_dep'), because you can avoid building fmt if you already have one.)

[lacker]

I took another look at using this meson wrap but with default_options: ['default_library=static'] I get the same error. I tried both on the subproject lookup and on the overall project default_options.