linux-kernel - Re: [KTAP V2 PATCH] ktap_v2: add recognized test name line

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sun, 26 Mar 2023 21:44:30 -0500
From:   Frank Rowand <frowand.list@...il.com>
To:     Daniel Latypov <dlatypov@...gle.com>, Rae Moar <rmoar@...gle.com>
Cc:     davidgow@...gle.com, skhan@...uxfoundation.org,
        keescook@...omium.org, Tim.Bird@...y.com,
        brendanhiggins@...gle.com, corbet@....net,
        guillaume.tucker@...labora.com, kernelci@...ts.linux.dev,
        kunit-dev@...glegroups.com, linux-kselftest@...r.kernel.org,
        linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [KTAP V2 PATCH] ktap_v2: add recognized test name line

On 3/20/23 14:21, Daniel Latypov wrote:
> On Thu, Mar 16, 2023 at 3:59 PM Rae Moar <rmoar@...gle.com> wrote:
>>
>> Add recognition of the test name line ("# Subtest: <name>") to the KTAP v2
>> spec.
>>
>> The purpose of this line is to declare the name of a test before its
>> results. This functionality is especially useful when trying to parse test
>> results incrementally and when interpretting results after a crash.
>>
>> This line is already compliant with KTAP v1 as it is interpretted as a
> 
> minor nit for if there's a v2, s/interprett/interpret (here and above)
> 
> Also, I want to elaborate on the previous paragraph a bit more, in
> case the motivation isn't clear.
> The problem with TAP and KTAP as-is is that the name of a test case is
> only known *after* it completes.
> 
> So the scenario being referred to is
> 
> KTAP version 1
> 1..1
> <lots of output>
> <kernel crash, no more output>
> 
> It would be nice if parsers could report "test FOO caused a crash" as
> opposed to "the first test case caused a crash, good luck figuring out
> which one that was"
> 
> Daniel

It would be useful to take some of the motivation explanation from TAP14,
which says (everything to the end of this email):

Commented Subtests are encouraged, as they provide the following benefits:

Easier for humans to read. For example:

  TAP version 14
              1..1
              ok 1 - hmm, what level is this?
vs:

  TAP version 14
  # Subtest: level 1
      # Subtest: level 2
          # Subtest: level 3
              1..1
              ok 1 - clearly level 3

Additional strictness around matching the Test Point description to Subtest Name can catch errors and detect accidentally interleaved output.