Updated to use suggested export _H1 env var to supply Authorization Basic credentials. This undocumented support for Basic Authorization, ContentType, etc. needs to be documented in DNSAPI Dev Guide. Removed two stray debugging lines.
Created _get_root() that tests the requested host is a subdomain to the domains hosted on MailinaBox (MIAB) DNS Server. Created common _miab_rest() used with dns_miab_add(), dns_miab_rm() and _get_root(). Also created barbaric _is_json() to test the response given by the MIAB Custom DNS API at least looks like a JSON file. We should add a hint to use _normalizeJson with JSON responses so _startswith, _endswith won't perplexingly fail.
Ok, should have noticed earlier that the calls to the private function _miab_post() never used the _needbase64_ or the __postContentType parameters. Parameters and code to handle them has been factored out.
Know I'm new to contorting to this project. I i've broke conventions please let me know what I've screwed up and I'll set it right as quickly as possible.
Propose this as a new DNS-01 validation script to dynamically add challenge DNS records to MailinaBox (MIAB) DNS. MIAB uses a custom DNS API to manage external DNS records.
The script was originally written by Darven Dissek and can be found in his repository: https://framagit.org/DarvenDissek/acme.sh-MIAB-DNS-API/). This has been forked and some slight cleanup applied and change shebang to UNIx shell. The forked repository can be found here: https://github.com/billgertz/MIAB_dns_api.
Wrote to Darven but received no reply. Support for this script has been submitted to the OPNsense project via this pull request: https://github.com/opnsense/plugins/pull/1531
* first version dns_rcode0.sh
* fixed URLs for ACME calls
* fixed challenge remove
* read & write Token/URL at rm too
* make info messages debug
* typos fixed
* update rrset only if existing challenge is found
* polish error messages and make "detect root zone" scaleable
* fixed formating issues
* code cleanup, remove some unneeded functions
* removed empty lines
* save rcode0 url only if not default
_egrep_o() function accepts extended regex and on systems that do not have egrep uses sed to emulate egrep.
This is failing on the specific regex I was using before my last commit... ae66c6f0b4
The problem is that I fixed it by passing in non-extended regex which then fails on systems that do have egrep. So I am no longer using _egrep_o.
We have a few domains that ends the same. For example :
iperfony.com
perfony.com
The problem was in the _get_root functions, when getting the domain_id :
only the first result "iperfony.com" was returned, because "perfony.com"
is contained in the "iperfony.com" string.
The correction consist of being strict in the regex, adding a slash (/)
so that it will only match on ".*/(perfony.com).*" and not
".*(perfony.com).*".
David Robles