Rule: defer-assignment
#1215
Rule: defer-assignment
#1215
Conversation
Finally a new rule in the `performance` category! And this is quite a fun one. I was curious to see how many violations this would find in the Regal codebase, if any. It did find a few, and good ones to fix! In order to keep it simple, this first implementation is perhaps a bit too cautious about what it considers deferrable. Check the implementation to see what exactly, but pretty much only assignments where the next line is a not an assignment, a potential loop, or something else that *might* make deferring the assignment undesirable. Fixes #1147 Signed-off-by: Anders Eknert <anders@styra.com>
anderseknert
force-pushed
the
defer-assignment
branch
from
1c32a5a to
1867055
Compare
|
|
||
| not _ref_with_vars(rhs) | ||
|
|
||
| # for now, only simple var assignment counts.. later we can |
There was a problem hiding this comment.
| # for now, only simple var assignment counts.. later we can | |
| # for now, only simple var assignment counts, later we can |
|
|
||
| Assignments are normally cheap, but certainly not always. If the right-hand side of an assignment is expensive, | ||
| deferring the assignment to where it's needed can save a considerable amount of time. Even for less expensive | ||
| assignments, code tends to be more readable when assignments are placed close to where they're used. |
There was a problem hiding this comment.
code tends to be more readable when assignments are placed close to where they're used.
I think this is the number one reason, cheap/expensive is the secondary reason for this rule since it does not always apply / is more complicated.
Something like this might work:
Placing assignments close to where they’re used improves code readability. Following this rule is also particularly important when the right-hand side of an assignment is expensive, such as making an HTTP call, as deferring the assignment can also save significant time.
There was a problem hiding this comment.
This rule is in the performance category, which is why it's emphasized here. I think there are good examples of both really, and there are of course examples of both where location doesn't matter much.
Finally a new rule in the `performance` category! And this is quite a fun one. I was curious to see how many violations this would find in the Regal codebase, if any. It did find a few, and good ones to fix! In order to keep it simple, this first implementation is perhaps a bit too cautious about what it considers deferrable. Check the implementation to see what exactly, but pretty much only assignments where the next line is a not an assignment, a potential loop, or something else that *might* make deferring the assignment undesirable. Fixes StyraInc#1147 Signed-off-by: Anders Eknert <anders@styra.com>
Finally a new rule in the
performancecategory! And this is quite a fun one. I was curious to see how many violations this would find in the Regal codebase, if any. It did find a few, and good ones to fix!In order to keep it simple, this first implementation is perhaps a bit too cautious about what it considers deferrable. Check the implementation to see what exactly, but pretty much only assignments where the next line is a not an assignment, a potential loop, or something else that might make deferring the assignment undesirable.
Fixes #1147