rinseone / rinseweb Goto Github PK

To update from Erlang/OTP 24 to 25

• Add CircleCI tests against erlang 25
• Update github workflow to build a release with erlang 25

It seems the web CT suites repeat the same or similar steps after each request - they assert the response status, headers, returned response, and parse the answer. It seems these repetitive steps can be moved to the rinseweb_test or similar shared module to avoid lots of duplication.

Currently in use is erlang 25.0.4 which can be upgraded to 25.3, which is currently the latest version published by Erlang Solutions.

Currently 25.0.3 is used. The most recent version available it 25.0.4.

Current release is using Erlang/OTP 25.3. As of this writing the latest version of Erlang/OTP is 26.0.2.

The command could be

what is my ip

And the response would be the IP address from which the request was received.

The command could be

what is my user agent

and the response would display the user agent header values given in the request. Optionally, the response could also contain other interesting HTTP headers.

Currently in use

• cowboy 2.9.0
• uuid_erl 2.0.2

can be upgraded to

• cowboy 2.10.0
• uuid_erl 2.0.6

Now that #43 uses GitHub workflows to test on Ubuntu 22.04, do CircleCI tests have any value? If not remove them. Also, remove CircleCI access/setup, if any.

Currently rinseweb is on Ubuntu 20.04. Ubuntu 22.04 has been out for awhile and should be a straightforward upgrade.

To reproduce

Type in this command

convert 10GB to bytes

Expected result: "10 GB = 10000000000 bytes"

Actual result: 500 error response from server.

Details

Error log

  crasher:
    initial call: cowboy_stream_h:request_process/3
    pid: <0.382.0>
    registered_name: []
    exception error: bad argument
      in function  list_to_integer/1
         called as list_to_integer("1e+10")
         *** argument 1: not a textual representation of an integer
      in call from rinseweb_util:string_to_number/1 (src/rinseweb_util.erl, line 45)
      in call from rinseweb_wiz_convert:answer_using_canonical_units/3 (src/rinseweb_wiz_convert.erl, line 38)
      in call from rinseweb_wiz:handler_answer/3 (src/rinseweb_wiz.erl, line 117)
      in call from rinseweb_wiz:answer/1 (src/rinseweb_wiz.erl, line 44)
      in call from rinseweb_h_answer:answer_to_json/2 (src/rinseweb_h_answer.erl, line 61)
      in call from cowboy_rest:call/3 (_build/default/lib/cowboy/src/cowboy_rest.erl, line 1583)
      in call from cowboy_rest:set_resp_body/2 (_build/default/lib/cowboy/src/cowboy_rest.erl, line 1472)

Why

Per units man page

The default format for 'units' is '%.8g'; for greater precision, you could specify '-o %.15g'. The 'g' and 'G' format types use exponential format whenever the exponent would be less than -4, so the value 0.000013 displays as '1.3e-005'. These types also use exponential notation when the exponent is greater than or equal to the precision, so with the default format, the value 5e7 displays as '50000000' and the value 5e8 displays as '5e+008'. If you prefer fixed-point display, you might specify '-o %.8f'; however, small numbers will display very few significant digits, and values less than 0.5e-8 will show nothing but zeros.

So the units command output in the above case is 1e+10 which erlang doesn't recognize as valid. Incidentally, if it was represented as 1.0e+10, then it would have been recognized as numeric. This can also be reproduced using the units command directly

> units --terse '10GB' 'bytes'
1e+10