exclude termkey

[dundargoc]

No description provided.

[justinmk]

fwiw, I did pass along this to Leonerd (seems like it could be legitimate?). That's for coverity, not clang scan, but if it's not noisy it might be worth keeping (especially if we are maintaining a fork). If it is noisy, then sure, let's ignore it.

*** CID 469488:  API usage errors  (SWAPPED_ARGUMENTS)
/src/termkey/termkey.c: 1531 in termkey_strpkey()
1525         key->code.mouse[0] = code;
1526
1527         unsigned int line = 0, col = 0;
1528         if((format & TERMKEY_FORMAT_MOUSE_POS) && sscanf(str, " @
(%u,%u)%zn", &col, &line, &snbytes) == 2) {
1529           str += snbytes;
1530         }
>>>     CID 469488:  API usage errors  (SWAPPED_ARGUMENTS)
>>>     The positions of arguments in the call to "termkey_key_set_linecol" do not match the ordering of the parameters:
* "col" is passed to "line".
* "line" is passed to "col".
1531         termkey_key_set_linecol(key, col, line);
1532       }
1533       // Unicode must be last
1534       else if(parse_utf8((unsigned const char *)str, strlen(str),
&key->code.codepoint, &nbytes) == TERMKEY_RES_KEY) {
1535         key->type = TERMKEY_TYPE_UNICODE;
1536         fill_utf8(key);

[dundargoc]

I think we should be consistent with keeping the the scan-build and clang-analyzer CI check in sync (since the goal is to eventually remove scan-build completely). Since the clang-analyzer only checks under src/nvim I thought this was more appropriate. I don't think it matters much tbh either way, shall I revert this?

[justinmk]

Don't need to revert, can revisit later if libtermkey continues to be something we are maintaining.