On Thu, Jun 18, 2020 at 11:39 PM David Gow davidgow@google.com wrote: [...]
On Fri, Jun 19, 2020 at 4:28 AM Brendan Higgins brendanhiggins@google.com wrote:
On Tue, Jun 16, 2020 at 9:21 PM David Gow davidgow@google.com wrote:
On Tue, Jun 16, 2020 at 5:40 PM Alan Maguire alan.maguire@oracle.com wrote:
On Tue, 16 Jun 2020, David Gow wrote:
[...]
keescook@chromium.org wrote:
On Sat, Jun 13, 2020 at 02:51:17PM +0800, David Gow wrote:
[...]
- Test names: Personally, I'd kind-of like to not prefix these at all,
as they're already part of the suite. If we do want to, though, prefix them with <subsystem> and <suite>.
Eh, I did that to remain consistent with the kernel naming conventions, but I think those have diverged too. If maintainers are cool with it, I agree that the prefixes are redundant on tests and generally way too long.
Do you have a link to the conventions you're talking about?
A link no. This is only of those undocumented rules that most people follow.
The rule is something like this:
Global identifiers should be named: <subsystem_name_n>_<subsystem_name_n-1>_..._<subsystem_name_1>_<subsystem_name_0>_foo.
For example, let's say I am working on Synopsis' DesignWare I2C master driver. The outermost namespace is i2c, and because DesignWare is long, we might prefix each function with i2c_dw_*.
It is a practice that is not universally maintained around the kernel, but it seems to be the most common method of namespacing aside from just randomly throwing characters together in a prefix that hasn't been used before.
Anyway, standardized or not, that is the convention I was trying to follow.