RFR: 8214278: Cleanup process_completed_threshold and related state
stefan.johansson at oracle.com
Tue Nov 27 09:32:14 UTC 2018
On 2018-11-26 22:17, Kim Barrett wrote:
>>> - not sure if static consts like
>>> _process_completed_buffers_threshold_never should be named like a
>>> member. I think at some point we agreed on (but did not write down) to
>>> use CamelCase for those.
> This came up in Thomas's review of another change of mine recently. I
> don't have a strong preference. I'm merely attempting to follow the
> Style Guide:
> Top Ten List:
> 9. #Names Instance variable names start with underscore ...
> - Local names (fields, variables) ... are lower-case.
> - Constant names in upper-case or mixed-case are *tolerated* [emphasis mine], according to historical necessity.
> - Constant names ... must have a distinct appearance from other names in related APIs.
> The first and second suggest the convention I used, e.g. lower case
> with leading underscore. The third argues against the proposed use of
> mixed-case (with or without leading underscore?). The last, having
> distinct appearance, seems inconsistent with the others. (Although
> who really knows, since the terminology in these rules is not standard
> C++ terminology, so it can be hard to be certain what is being
> referred to.)
I totally agree that the style guide doesn't give any real guidance
here. And as you say, we use every possible interpretation in the code
base already except maybe Camel_Snake. To me the most important is the
last bullet, that it should be easy to understand that this is a
constant. I also think the first point only talks about *instance*
variables, so not statics.
I don't see this as a blocker for this change, we have different
preferences and I'm ok with this patch as is.
> And of course, examples of all of these mentioned styles (and probably
> more) can be found in our code base.
> Anyone feel like proposing some modifications to the Style Guide?
More information about the hotspot-gc-dev