What regex can be used to filter out dalvikvm AND dalvikvm-heap messages from the logcat

Question

Using this link I was able to create a filter using the regex (?!dalvikvm\b)\b\w+ to filter out messages with the tag dalvikvm, but I have tried several variations of the regex such as (?!dalvikvm-heap\b)\b\w+, (?!dalvikvm\\-heap\b)\b\w+, (?!dalvikvm[-]heap\b)\b\w+, and many others and I can't seem to get rid of the dalvikvm-heap messages. Ideally I would like to filter them both, but I haven't figured that part out yet either.

Any help would be appreciated.

Answer 1

Use ^(?!dalvikvm) in the tag field instead. That will show only messages whose tag doesn't start with "dalvikvm".

The below is notes about how this works; you can skip them if you're not interested. To start, you have to remember that the question, "Does this string match the regex?" really means, "Is there any position in this string where the regex matches?"

The tricky thing about (?!x) negative assertions is that they match wherever the next part of the string doesn't match x: but that's true of every place in the string "dalvikvm" except the start. The blog post you linked to adds a \b at the end so that the expression matches only at a place that's not just before "dalvikvm" and is a word boundary. But this would still match, because the end of the string is a word boundary, and it doesn't have "dalvikvm" after it. So the blog post adds the \w+ after it, to say that after the word boundary there have to be more word characters.

It works for exactly that case, but it's a bit of an odd way of making a regex, and it's relatively expensive to evaluate. And as you've noticed, you can't adapt it to (?!dalvikvm-heap\b)\b\w+. "-" is a non-word character, so immediately after it, there is a word boundary, followed by word characters, and not followed by "dalvikvm-heap", so the regex matches at that point.

Instead, I use ^, which only matches at the start of the string, along with the negative assertion. Overall, the regex only matches at the start of the string, and only then if the start of the string isn't followed by "dalvikvm". That means it won't match "dalvikvm" or "dalvikvm-heap". It's also cheaper to evaluate, because the regex engine knows it can only possibly match at the start.

Making the regex this way, you can filter out multiple tags by just putting them together. For example, ^(?!dalvikvm)(?!IInputConnectionWrapper) will filter out tags that start with "dalvikvm" or "IInputConnectionWrapper", because the start of the string has to not be followed by the first and not be followed by the second.

BTW, thanks for your link. I didn't realise that you could use the logcat filters that way, so I wouldn't have come up with my answer without it.