g.extension: show fatal message if installed Add-On doesn't exists in the repository

[tmszi]

Describe the bug
If you try install install non exists Add-On, g.extension module fail.

To Reproduce
Steps to reproduce the behavior:

1. Try install some non exists Add-On g.extension non_exists_addon
2. See error

GRASS world_latlong_wgs84/PERMANENT:~ > g.extension non_exists_addon
Fetching <non_exists_addon> from <https://github.com/OSGeo/grass-addons>
(be patient)...
Traceback (most recent call last):
  File "/usr/lib64/grass83/scripts/g.extension", line 2867, in <module>
    sys.exit(main())
  File "/usr/lib64/grass83/scripts/g.extension", line 2847, in main
    install_extension(source=source, url=url, xmlurl=xmlurl, branch=branch)
  File "/usr/lib64/grass83/scripts/g.extension", line 1230, in install_extension
    ) = install_extension_std_platforms(
  File "/usr/lib64/grass83/scripts/g.extension", line 1991, in install_extension_std_platforms
    srcdir = download_source_code(
  File "/usr/lib64/grass83/scripts/g.extension", line 1903, in download_source_code
    directory = download_source_code_official_github(
  File "/usr/lib64/grass83/scripts/g.extension", line 1779, in download_source_code_official_github
    ga.fetch_addons([name])
  File "/usr/lib64/grass83/scripts/g.extension", line 411, in fetch_addons
    *[self.addons[addon] for addon in addon_list],
  File "/usr/lib64/grass83/scripts/g.extension", line 411, in <listcomp>
    *[self.addons[addon] for addon in addon_list],
KeyError: 'non_exists_addon'

Expected behavior
Fatal message should be printed.

GRASS world_latlong_wgs84/PERMANENT:~ > g.extension non_exists_addon
Fetching <non_exists_addon> from <https://github.com/OSGeo/grass-addons>
(be patient)...
ERROR: Add-On name <non_exists_addon> doesn't exists in the repository. Try
       it again with correct Add-On name, please.

System description (please complete the following information):

• Operating System: all
• GRASS GIS version: 8.3.dev

[ninsbl]

Good catch. ANd thanks for the quick fix, @tmszi !
The code looks good to me. One could discuss if this should be an error or a warning message (when the attempts to download and compile are skipped). But for now, fatal seems right.

The suggested changes are just that, suggestions. So feel free to ignore.

[ninsbl]

BTW, you may add a test case for this issue to the download test!

[tmszi]

BTW, you may add a test case for this issue to the download test!

Yes, you are right, I added it.

[nilason]

To be backported with #2895