mirror of
https://github.com/ceph/ceph-csi.git
synced 2024-11-18 04:10:22 +00:00
32839948ef
The internal/ directory in Go has a special meaning, and indicates that those packages are not meant for external consumption. Ceph-CSI does provide public APIs for other projects to consume. There is no plan to keep the API of the internally used packages stable. Closes: #903 Signed-off-by: Niels de Vos <ndevos@redhat.com>
2.3 KiB
2.3 KiB
Coding Conventions
Please follow coding conventions and guidelines described in the following documents:
- Go proverbs - highly recommended read
- CodeReviewComments
- Effective Go
- How to Write a Git Commit Message
Here's a list of some more specific conventions that are often followed in the code and will be pointed out in the review process:
General
- Keep variable names short for variables that are local to the function.
- Do not export a function or variable name outside the package until you have an external consumer for it.
Imports
We use the following convention for specifying imports:
<import standard library packages>
<import third-party packages>
<import ceph-csi packages>
Example:
import (
"os"
"path"
"strings"
"time"
"github.com/pborman/uuid"
"github.com/pkg/errors"
"github.com/ceph/ceph-csi/internal/util"
)
Error Handling
- Use variable name
err
to denote error variable during a function call. - Reuse the previously declared
err
variable as long as it is in scope. For example, do not useerrWrite
orerrRead
. - Do not panic() for errors that can be bubbled up back to user. Use panic() only for fatal errors which shouldn't occur.
- Do not ignore errors using
_
variable unless you know what you're doing. - Error strings should not start with a capital letter.
- If error requires passing of extra information, you can define a new type
- Error types should end with
Error
.
Logging
- The inner-most utility functions should never log. Logging must almost always
be done by the caller on receiving an
error
. - Always use log level
DEBUG
to provide useful diagnostic information to developers or sysadmins. - Use log level
INFO
to provide information to users or sysadmins. This is the kind of information you'd like to log in an out-of-the-box configuration in happy scenario. - Use log level
WARN
when something fails but there's a workaround or fallback or retry for it and/or is fully recoverable. - Use log level
ERROR
when something occurs which is fatal to the operation, but not to the service or application.