ERT conventions?

ERT conventions?

[Marcin Borkowski]

Hi all,

I'm learning the ERT package.  I'd like to know whether there are any
conventions/good practices regarding writing tests.  Eg., should they be
written in a file separate from the package?  (I guess yes, for obvious
reasons.)  Are there any standards for the name of that file?  In
general: could anyone using ERT (or any other testing facility, for that
matter) share some experiences worth knowing?

TIA,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University

Re: ERT conventions?

[Alexis]

Marcin Borkowski <mbork@mbork.pl> writes:

> Hi all,
>
> I'm learning the ERT package.  I'd like to know whether there 
> are any conventions/good practices regarding writing tests. 
> Eg., should they be written in a file separate from the package? 
> (I guess yes, for obvious reasons.)  Are there any standards for 
> the name of that file?  In general: could anyone using ERT (or 
> any other testing facility, for that matter) share some 
> experiences worth knowing?

i've found ERT pretty easy to use. For `org-vcard`, i've put the 
test suite in a single file within a distinct 'tests' folder, 
which also contains a 'data' folder containing various data 
sources for the tests to run on:

https://github.com/flexibeast/org-vcard/tree/master/tests


Alexis.

Re: ERT conventions?

[Marcin Borkowski]

On 2015-09-06, at 07:27, Alexis <flexibeast@gmail.com> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>> Hi all,
>>
>> I'm learning the ERT package.  I'd like to know whether there 
>> are any conventions/good practices regarding writing tests. 
>> Eg., should they be written in a file separate from the package? 
>> (I guess yes, for obvious reasons.)  Are there any standards for 
>> the name of that file?  In general: could anyone using ERT (or 
>> any other testing facility, for that matter) share some 
>> experiences worth knowing?
>
> i've found ERT pretty easy to use. For `org-vcard`, i've put the 
> test suite in a single file within a distinct 'tests' folder, 
> which also contains a 'data' folder containing various data 
> sources for the tests to run on:

Thanks!

And it's not that ERT is difficult to use, no.  What I'm asking is (for
instance): is it better to say

(ert-deftest my-function-test ()
  "Tests the `my-function' function."
  (with-temp-buffer
    (insert "input")
    (goto-char (point-min))
    (should (this-be-satisfied)
    (should (that-be-satisfied)))))

or maybe

(ert-deftest my-function-test ()
  "Tests the `my-function' function."
  (should (with-temp-buffer
	    (insert "input")
	    (goto-char (point-min))
	    (this-be-satisfied)))
  (should (with-temp-buffer
	    (insert "input")
	    (goto-char (point-min))
	    (that-be-satisfied))))

?

The former avoids code repetition and probably is (a bit) faster; the
latter /might/ result in better reporting.

Etc.

> Alexis.

Best,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University

Re: ERT conventions?

[Alexis]

Marcin Borkowski <mbork@mbork.pl> writes:

> The former avoids code repetition and probably is (a bit) 
> faster; the latter /might/ result in better reporting.

Well, based on my experiences, it seems to me that neither would 
produce better reporting than the other. i feel the `org-vcard` 
tests aren't granular enough: multiple `should`s in the same 
function can make (and has made) it more difficult to determine 
exactly which `should` assertion is failing. So at some point, 
resources permitting, i'd like to refactor the tests to 
one-assertion-per-`ert-deftest`.

(Volunteers welcome! :-) )


Alexis.