How to escape \ before tag

[yanyongyu]

In my server application, i'm using loguru logger to log messages from client. But in some case, the log failed with the error: nonebot/adapter-onebot#74 ValueError: Closing tag "</le>" has no corresponding opening tag

Here is the example of the case:

message = r"\\<le>test</le>"
logger.opt(colors=True).info(message)

It seems double \ cannot escape the escape before tag according to the ansi parser. how can i escape this message currectly?

loguru/loguru/_colorizer.py

Lines 223 to 232 in e1f48c9

	for match in self._regex_tag.finditer(text):
	markup, tag = match.group(0), match.group(1)
	self._tokens.append((TokenType.TEXT, text[position : match.start()]))
	position = match.end()
	if markup[0] == "\\":
	self._tokens.append((TokenType.TEXT, markup[1:]))
	continue

[AlTosterino]

Hey,

Hmm, are you able to add space after \\? I just checked and adding a space is fixing the issue you are facing. Note that this is just workaround.

from loguru import logger

if __name__ == "__main__":
    logger.remove()
    logger.add(sys.stdout)
    logger.opt(colors=True).info("\\ <le>test</le>")
    # 2023-09-19 17:40:56.966 | INFO     | __main__:<module>:8 - \ test

[Delgan]

You need to escape both opening and closing tags.

message = "\<le>test\</le>"
logger.opt(colors=True).info(message)
# Output: "<le>test</le>"

Note that you shouldn't generally combine colors=True with messages that may contain arbitrary data. It's better to pass them as arguments, to avoid false positives of this type.

logger.opt(colors=True).info("<green>Parsed data:</green> <blue>{}</blue>", data)

[yanyongyu]

The message is a markup text with a \ before tag, the <le>test</le> is highlighted and should not be escaped.

Expected: TEXT \, OPEN tag <le>, TEXT test, CLOSING tag </le>

[Delgan]

In this case, you need to escape the two tags in addition to the \ already present.

message = r"\\<le>test\</le>"
logger.opt(colors=True).info(message)

Loguru uses tags such as <le> to colorize the message. To avoid the tag being interpreted as a colorization command, it must be preceded by \. In such case, "\<le>" is converted to "<le>". Because you want to preserve the \ from your original message, you need two \: the one to escape colors, and the one from your original message.

[yanyongyu]

May be i should provide detailed info about the case.:joy:

We have a tokenizer to parse the input string, and highlight special tokens. Here is an example with detailed steps:

1. Input: This is a message with [special_token]
2. tokenize result: This a message with , [special_token]
3. output: This a message with <le>[special_token]</le>

Once there is a \ before special token, the highlight is failed to display:

1. Input: This is a message with \[special_token]
2. tokenize result: This a message with \, [special_token]
3. output: This a message with \<le>[special_token]</le>

The output is not a valid loguru markup message.

[Delgan]

Thanks for the explanation. I'm sorry, this is indeed a bug. It's not possible to display a literal "\" immediately before a text to be colored. Thanks for the report.

The workaround provided by @AlTosterino seems to be the simplest solution for now.

[Delgan]

Fixed on master branch. It is now possible to use double backslash to display a literal \ without escaping the color markup.

For example, the following code will now print Message: \test with test being blue colored:

from loguru import logger

logger.opt(colors=True).info(r"Message: \\<le>test</le>")