GitLab

  Changes:
    - Updated Raleway to Fira Sans and Inconsolata to Fira Mono fonts resp.
    - Corrected usage of custom.css in conf.py

Edited conf.py
  - html_css_files used instead of html_context (For more details refer issue and solution)

[#5] Fix rare bug where CMake attempts to build _ydbposix.so in UTF-8 mode before the M mode build has finished

This commit fixes a rare bug where CMake attempts to build both versions of _ydbposix.so at the same time without waiting for the other one to finish. We discovered this bug in an internal build where the YDBPosix plugin did not work at all. The fix is to make the M mode version of _ydbposix.so a dependency for the UTF-8 version which ensures that CMake cannot start building the UTF-8 version of _ydbposix.so until after the M mode version has already been built.

Changes made:
  - Edited conf.py to fix error (#6) and remove version value (r3)
  - Edited ydbposix.rst to comply with style guidelines
  - Edited Makefile to add clean target and htmlpatch.csh command
  - Edited htmlpatch.csh

This commit modifies CMakeLists to detect if YottaDB is installed with support for UTF-8. If UTF-8 mode support is installed, then the plugin is also built with UTF-8 support without having to run the installer twice. This is done by adding a file called ydbposixso.cmake which now handles the building of _ydbposix.so files. This file is based on sr_unix/mumps.cmake from the YDB repo and is used in a similar manner by CMakeLists.txt to ensure that the shared library is built for both M and UTF-8 modes.

  Changes made:

    - Added SPHINXOPTS to Makefile so that Sphinx reports warnings as errors
    - Updated the Copyright year
    - Changed parsed-literal to code-block to make it consistent with other documentation

See YDB!869 for what's going on here.

- Add `needs_copyright.sh`
- Change `pre-commit` to use `needs_copyright.sh`
- Change `.gitlab-ci.yml` to use the new name `verify_commit.sh` instead of `verify_gpg_commit.sh`
- Only run the `commit_verify` pipeline for MRs, not for individual
commits

  If this is not done, the environment variables for `TARGET_BRANCH`
  will not be set and there will be nothing to compare the HEAD commit
  to: the script will not know how much of the history is pre-existing and
  how much was added for the current MR.

This should not be merged until after YDB!863 lands, or it will not be
able to find the new copyright script.

[YottaDB/DB/YDBDoc#212] Use the `tcsh` in the user environment instead of hard-coding /usr/local/bin

In gitlab pipelines, tcsh is installed to /usr/bin/tcsh.

`commit_gpgkey_verify.sh` invocation is moved to separate job `gpgkey-verify`

Upstream URL is added as argument to commit_gpgkey_verify script invocation

before_script is used in gpgkey-verify job so that it is overridden at job level

[YottaDB/DBMS/YDBOcto#567] Rename MUMPS_UTF8_MODE to M_UTF8_MODE now that language name of MUMPS is renamed to M

This contains the changes renaming MUMPS to M.

This contains similar changes to YottaDB/DBMS/YDBOcto!731 and YottaDB/DBMS/YDBOcto!691.

This makes it possible to update downstream repositories using this code
with the command

```
git pull git@gitlab.com:YottaDB/Tools/YDBcmake.git
```

This makes it easy for other repos to use this one as a subtree without
having too many nested directories. Otherwise, they end up with
directories named `ydbcmake/ydbcmake/` which looks very odd.

This changes build output of the form
```
[  0%] Building MUMPS object src/CMakeFiles/_ydbocto.dir/aux/_ydboctoAdmin.m.o
```
to
```
[  0%] Building M object src/CMakeFiles/_ydbocto.dir/aux/_ydboctoAdmin.m.o
```

The `v60001/gtm_posix` subtest in the YDBTest repo failed after commit 9007dec8 with the following diff.

> FAIL regmatch^%ydbposix 4
> FAIL REGMATCH^%ydbposix 4

This is because that commit erroneously assumed that `matchresults(i,"start")` might not be set in the
for loop in case `$&ydbposix.regofft2offsets` returned a non-zero value. And because of this, it added
a `$data(matchresults(i,"start"))` check instead of the pre-existing `'matchresults(i,"start"))` check.

Turns out `$&ydbposix.regofft2offsets` always returns a 0. Therefore, the `$data` check is not
necessary and in fact incorrect since we need to actually check for 0 value of `matchresults(i,"start")`
to terminate the for loop. This is now done. And comments added in the code to help a future reader.

* The primary issue was a `if $increment(i,-1)` that was unconditionally done before the `quit:$quit`
  in `$$regmatch^%ydbposix`. I suspect the author of 3b8abd6d intended to do this in case
  the `$&ydbposix.regofft2offsets()` call returned a non-zero value (i.e. an error scenario)
  since that resulted in `matchresults(i,"start")` and `matchresults(i,"end")` not getting set.
  Therefore, this decrement is now moved to where it should have been by a `set i=i-1` just before
  we `quit` out of the for loop.

* A secondary issue in my opinion is that `if 'matchresults(i,"start")` done after the `do` call
  would result in an LVUNDEF error in case `matchresults(i,"start")` did not get set in the
  `do` body. This is now fixed by instead doing a `$data()` on it.

Journalctl won't work properly on systems not being initialized
using systemd init system. It can be tricked to work by starting
systemd logger service explicitly but the test does not assume that.

Containers with systemd entrypoint needs to be running in privileged
mode (as of today) which implies that most of the GitLab CI runners
won't be able to run systemd-enabled containers.

This commit is therefore properly detecting systemd rather than
just assuming that it is not available in the CI pipelines only.

Add EMBED_SOURCE and DYNAMIC_LITERALS mumps compiler options enabled
by default for the generated shared objects

MacOS uses .dylib as the shared library name instead of .so. Add this
to the list of library suffixes so that YottaDB can be found automatically.

Some YottaDB plugins may support multiple platforms and thus may need
to ignore warnings generated by the compiler to install correctly.

Below is the warning we got from clang.

ydbposix.c:337:17: warning: using the result of an assignment as a condition without parentheses [-Wparentheses]

This is now fixed by moving the assignment to a separate line.

[DEBIAN10] YottaDB is now supported on Debian 10 (buster/sid); icu-config is deprecated in Debian 10 so use pkg-config instead everywhere