Improve CSV parsing to allow quoted and escaped fields

[EricTendian]

The RadioReference talkgroup CSV output looks like the following:

Decimal,Hex,Alpha Tag,Mode,Description,Tag,Category
1,001,",T,"Installation / Techs / Evaluation",Business,"Motorola Installation / Testing"
2,002,"Motorola Tech 2",D,"Installation / Techs / Evaluation",Business,"Motorola Installation / Testing"
3,003,"Motorola Tech 3",DE,"Installation / Techs / Evaluation",Business,"Motorola Installation / Testing"

However, when trunk-recorder parses the CSV, it does not strip out the quotes, which leads to some malformed entry errors, as well as the call JSON to have syntax errors like the following:

{
    "talkgroup_tag": ""DC Tac 2"",
    "talkgroup_description": ""Tac 2"",
    "talkgroup_group_tag": "Law Dispatch",
    "talkgroup_group": ""ISP District Chicago""
}

This PR fixes that error by switching to the Boost escaped_list_separator class, which is able to handle text being quoted and escaped. The previous class char_separator did not handle this.

Both comma-separated values and tab-separated values are still supported by this parser. I've done some testing to make sure everything still works.

Besides this, I made a tiny change to the Dockerfile to make compiling trunk-recorder use all available CPU cores for faster compilation.

[robotastic]

These are great changes - thanks @EricTendian