BF: limit CMD_MAX_ARG if obnoxious value is encountered.

[yarikoptic]

Apparently the issue observed with elderly python 3.4 (no longer
supported) on Debian system, is likely to do nothing with the
Python but some peculiarity of the system since now was observed on
CentOS conda env with Python 3.9.

Since our hardcoded limits are quite generous, and 3.4 is no longer
support anyways, I decided just to add "flexible" check that if
returned value is a million times larger (still have a buffer of
thousand times in this particular case) -- take hardcoded value.

Closes #5943

[codecov]

Codecov Report

Merging #5945 (1c532fb) into maint (dfdd352) will decrease coverage by 5.63%.
The diff coverage is 100.00%.

❗ Current head 1c532fb differs from pull request most recent head 8aad62e. Consider uploading reports for the commit 8aad62e to get more accurate results

@@            Coverage Diff             @@
##            maint    #5945      +/-   ##
==========================================
- Coverage   90.28%   84.65%   -5.64%     
==========================================
  Files         300      297       -3     
  Lines       42516    42480      -36     
==========================================
- Hits        38387    35961    -2426     
- Misses       4129     6519    +2390     

Impacted Files	Coverage Δ
datalad/utils.py	81.56% <100.00%> (-3.55%)	⬇️
datalad/support/tests/utils.py	0.00% <0.00%> (-100.00%)	⬇️
datalad/metadata/extractors/tests/test_audio.py	19.35% <0.00%> (-80.65%)	⬇️
datalad/metadata/extractors/xmp.py	12.96% <0.00%> (-79.63%)	⬇️
datalad/metadata/extractors/exif.py	18.75% <0.00%> (-78.13%)	⬇️
datalad/metadata/extractors/image.py	19.35% <0.00%> (-77.42%)	⬇️
datalad/metadata/extractors/audio.py	20.00% <0.00%> (-77.15%)	⬇️
datalad/metadata/extractors/tests/test_exif.py	24.00% <0.00%> (-76.00%)	⬇️
datalad/metadata/extractors/tests/test_xmp.py	26.92% <0.00%> (-73.08%)	⬇️
datalad/metadata/extractors/tests/test_image.py	26.92% <0.00%> (-73.08%)	⬇️
... and 106 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 dfdd352...8aad62e. Read the comment docs.

[yarikoptic]

appveyor -- mac not even bothering to start

travis

======================================================================
ERROR: datalad.tests.test_archives.test_decompress_file('strip',)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/tmp/dl-miniconda-e7dty8st/lib/python3.9/site-packages/nose/case.py", line 198, in runTest
    self.test(*self.arg)
  File "/tmp/dl-miniconda-e7dty8st/lib/python3.9/site-packages/datalad/tests/utils.py", line 585, in _wrap_with_tree
    return t(*(arg + (d,)), **kw)
  File "/tmp/dl-miniconda-e7dty8st/lib/python3.9/site-packages/datalad/tests/test_archives.py", line 78, in check_decompress_file
    decompress_file(op.join(path, fn_archive_obscure_ext), outdir,
  File "/tmp/dl-miniconda-e7dty8st/lib/python3.9/site-packages/datalad/support/archives.py", line 99, in decompress_file
    rmdir(widow_dir)
  File "/tmp/dl-miniconda-e7dty8st/lib/python3.9/site-packages/datalad/utils.py", line 476, in rmdir
    os.rmdir(path)
OSError: [Errno 39] Directory not empty: '/tmp/nfsmount/datalad-e4vqr_en/simple-extracted/ |;&%b5{}\'"ΔЙקم๗あ .dbtc '

which is #4496 .

[yarikoptic]

Will merge tomorrow of no objections

[mih]

Should be safe ;-)

[yarikoptic]

travis -- AFAIK known NFS failing test, appveyor -- regular OSX fiasco. Let's proceed