Feature Request: accept long non-splittable texts with MAX_LINE_LENGTH about protolint HOT 6 OPEN

@bochengyang We already have a disable comment feature. So you can append // protolint:disable:this MAX_LINE_LENGTH at the end of the target line to skip the check.

from protolint.

@sulume Thank you for your feature request!
It makes sense a lot overall. I understand the need to accept some exceptions.

Do you have any good ideas or references to a specific implementation for that?
I think protolint should conservatively handle the exceptions because it already allows end-users to disable the rule by a line.

At first glance, at least an exception for import paths would be right without any nasty surprises.

[memo] Here is a current implementation.

• protolint/internal/addon/rules/maxLineLengthRule.go

  Lines 94 to 110 in c99b4b9

  	func(index int, line string) {
  	line = strings.Replace(line, "\t", strings.Repeat(" ", r.tabChars), -1)
  	lineCount := utf8.RuneCountInString(line)
  	if r.maxChars < lineCount {
  	failures = append(failures, report.Failuref(
  	meta.Position{
  	Filename: fileName,
  	Line: index + 1,
  	Column: 1,
  	},
  	r.ID(),
  	"The line length is %d, but it must be shorter than %d",
  	lineCount,
  	r.maxChars,
  	))
  	}
  	},

from protolint.

Thank you for your quick response.

I think protolint should conservatively handle the exceptions because it already allows end-users to disable the rule by a line.

I'm not really sure how the consistency is maintained, but I think one way is to add another rule option alongside max_chars and tab_chars so that users have options on its behavior and existing users would see it working the same by default.

For example,

• a boolean rule option to ignore import lines, and another for URLs
  • Separate options may have different default values.
  • There would be no single general criteria to decide if a line of URL is not splittable, so it might need a heuristic rule, e.g. a comment line with only a single URL.
• a string rule option to specify a regexp pattern to match lines to ignore their lengths
  • This way leaves users some extensibility for arbitrary long ID strings.

from protolint.

@sulume

so that users have options on its behavior and existing users would see it working the same by default.

That sounds great. Thank you for providing nice examples.
Thanks to those, it's time to implement each exception incrementally.

[memo] I'm thinking of working on them when I finish existing priority issues.
If you need them (asap) before I will assign myself, PRs are always welcome.

from protolint.

@yoheimuta

Since we can extract each line individually, can we have a regex for extracting annotation, e.g., #protolint:ignore: MAX_LINE_LENGTH at the end of the original protobuf definition? Once a rule is specified to be ignored, the line length is not a problem.

from protolint.

Great! It saves my day :)

Thanks

from protolint.