seriyps
commented
3 years ago Is anyone available to review this? @arjan @mworrell ? prop_mimemail suite mainly just generates arbitrary mimemail:mimetuple() structures (plain as well as multipart) and
1) Just tries to encode them with mimemail:encode
2) Tries to encode and then decode them with mimemail and checks that the result of encode+decode sequence is more or less the same as the original mimetuple()
We can later introduce additional checks. Say, validate that for any mimetuple(), when encoded, none of the lines in resulting mime text is longer than 80 chars. Or that resulting mime text only contains 7-bit ASCII etc.
> proper_gen:pick(prop_mimemail:gen_plaintext_mail()).
{ok,{<<"text">>,<<"plain">>,
[{<<"6tjAzEe">>,<<"Y<ul\\H">>},
{<<"7ZV">>,
<<240,144,128,133,240,144,128,129,225,172,153,88,224,189,
134,107>>},
{<<"B-">>,<<207,190,2,90>>},
{<<"From">>,<<"test@example.com">>},
{<<"GXz5l1KW">>,<<218,138,111,47,54>>},
{<<"V6">>,
<<199,131,208,142,240,144,128,132,235,146,171,240,144,
128,129,117,...>>},
{<<"aC_Iw">>,<<"Bi%\\P2\n">>},
{<<"to-t1M9R0">>,<<>>},
{<<"tq_x4Kn">>,<<"W<h8 fJ|">>}],
#{content_type_params => [{<<"charset">>,<<"utf-8">>}],
disposition => <<"attachment">>,
transfer_encoding => <<"quoted-printable">>},
<<"pi@'">>}}
4> proper_gen:pick(prop_mimemail:gen_plaintext_mail()).
{ok,{<<"text">>,<<"plain">>,
[{<<"1N">>,
<<219,165,46,240,144,128,129,240,144,128,136,216,130,240,
144,129,128,42>>},
{<<"From">>,<<"test@example.com">>},
{<<"PKbaID">>,<<"j%)\"I">>},
{<<"To">>,<<"to@example.com">>},
{<<"c">>,<<"M'VW : O)">>},
{<<"fsHoM7-">>,<<>>},
{<<"mh9dpzII">>,<<"{">>}],
#{disposition => <<"attachment">>,
transfer_encoding => <<"base64">>},
<<227,150,145,103,240,144,128,142,214,163,206,156,239,
162,171,231,153,171,238,150,149,16,...>>}}
5> proper_gen:pick(prop_mimemail:gen_multipart_mail()).
{ok,{<<"multipart">>,<<"alternative">>,
[{<<"Cc">>,<<"to@example.com">>},
{<<"From">>,<<"test@example.com">>},
{<<"gG">>,<<100,240,144,128,139,45,240,144,128,130>>},
{<<"oG">>,<<"vf">>}],
#{},
[{<<"text">>,<<"html">>,
[{<<"03">>,
<<214,156,219,168,240,144,128,133,239,150,159,240,144,
128,135,...>>},
{<<"0XF">>,<<>>},
{<<"DoTxLp4v62">>,<<"Zi|2">>},
{<<"LM1V0-zO">>,<<>>},
{<<"Reply-To">>,<<"to@example.com">>},
{<<"nII4hN">>,<<"}">>}],
#{content_type_params => [{<<"charset">>,<<"utf-8">>}],
disposition => <<"inline">>},
<<"<!doctype html><html><body><p>GX Kp},pU+</p></body></html>">>},
{<<"text">>,<<"plain">>,
[{<<"0LXHtM2et">>,<<">m[hLT">>},
{<<"84g1xX-qR">>,<<>>},
{<<"Fx">>,<<"ED\t2:FNX">>},
{<<"MY0YFh_">>,
<<240,144,128,137,201,151,240,144,128,151,74>>},
{<<"ymWky">>,<<70,53,240,144,128,130,240,144,128,139,...>>}],
#{content_type_params => [{<<"charset">>,<<"utf-8">>}],
disposition => <<"attachment">>,
transfer_encoding => <<"base64">>},
<<"b}OTtcK05">>},
{<<"text">>,<<"plain">>,[],
#{content_type_params => [{<<"charset">>,<<"utf-8">>}],
disposition => <<"inline">>,
transfer_encoding => <<"base64">>},
<<240,144,128,132,204,139,3,240,144,128,128,239,130,176>>},
{<<"text">>,<<"html">>,[],
#{content_type_params => [{<<"charset">>,<<"utf-8">>}],
disposition => <<"inline">>},
<<"<!doctype html><html><body><p></p></body></html>">>}]}}
There are 2 commits: 1st one just fixes #246 (I have choosen, unlike Python, to do QP-encoding of
\rand\n, because otherwisemimemailis not able to decode the result).2nd commit is some "proper" property-based test suite for
mimemail.Fix for #246 and test suite were created together, but I decided t split to 2 commits, because they are technically unrelated.
It might be interesting to think about having some "proper" test suite where we compare encoding/decoding of
mimemailwith some other mime codec (eg, some PHP or Python one) to see if they can inter-communicate (mail encoded by Erlang can be read by Python and vice versa), but next time.