text.Scanner is outputting unnecessary stderr

[arran4]

text.Scanner is outputting errors. Such as:

<input>:1:1: literal not terminated
<input>:1:1: invalid char literal
<input>:1:1: literal not terminated
<input>:1:1: invalid char literal

Such as when I use tokens like '':

		layoutData{
			layout: "' Mon, 2 Jan 2006 15:4:5 -0700",
			matches: []string{
				"' Нд, 22 Гру 2019 00:35:44 +0200",
			},
			unmatches: []string{},
			locales: []Locale{
				LocaleUkUA,
				LocaleRuRU,
				LocaleEnUS,
				LocaleEnUS,
			},
		},

I believe this is because text.Scanner is build to parse go tokens for a compiler.

Which technically means I should be able to break tests by passing in tokens like /* and when it /could/ be a valid format string.

Ie:

package main

import (
	"fmt"
	"github.com/goodsign/monday"
	"time"
)

func main() {
	fmt.Println(time.Parse("/* Mon, 02 Jan 2006 15:04:05 */", "/* Mon, 02 Nov 2007 17:42:36 */"))
	fmt.Println(monday.NewLocaleDetector().Parse("/* Mon, 02 Jan 2006 15:04:05 */", "/* Mon, 02 Nov 2007 17:42:36 */"))
	fmt.Println(time.Parse(".. Mon, 02 Jan 2006 15:04:05 ..", ".. Mon, 02 Nov 2007 17:42:36 .."))
	fmt.Println(monday.NewLocaleDetector().Parse(".. Mon, 02 Jan 2006 15:04:05 ..", ".. Mon, 02 Nov 2007 17:42:36 .."))
}

Outputs:

2007-11-02 17:42:36 +0000 UTC <nil>
0001-01-01 00:00:00 +0000 UTC parsing time "/* Mon, 02 Nov 2007 17:42:36 */"'/* Mon, 02 Nov 2007 17:42:36 */' not matches to '/* Mon, 02 Jan 2006 15:04:05 */' last error position = 0

2007-11-02 17:42:36 +0000 UTC <nil>
2007-11-02 17:42:36 +0000 UTC <nil>

[arran4]

bufio.Scanner is more appropriate (and what I have used in similar cases) but requires writing your own identifier for the lexer

[sw360cab]

@arran4 is it possible that I am encountering the same pseudo issue?

With this example code:

import (
	"fmt"
	"time"
	"github.com/goodsign/monday"
)


func main () {
	var (
		timeLocaleDetector  *monday.LocaleDetector = monday.NewLocaleDetector()
		parsedStartDatetime time.Time
		dateLayout          string = "2 Jan 2006 15:04"
		currentDate         string = "19 May 2024 09:00"
	)

	parsedStartDatetime, _ = timeLocaleDetector.Parse(
		dateLayout,
		currentDate)
	fmt.Println(parsedStartDatetime)
}

The output is the following

<input>:1:13: invalid digit '9' in octal literal
2024-05-19 09:00:00 +0000 UTC

It appears more as a warning, because the datetime is parsed smootly and the Parse method of monday is not returning any error at all.

[arran4]

Yep looks like it, I am not sure if my PR addresses this or not or other things but try the version in that to see if it's any better.