用于从 Content-Disposition 标头中提取文件名的 javascript 正则表达式

Question

Content-disposition header contains filename which can be easily extracted, but sometimes it contains double quotes, sometimes no quotes and there are probably some other variants too. Can someone write a regex which works in all the cases.

Content-Disposition: attachment; filename=content.txt

Here are some of the possible target strings:

attachment; filename=content.txt
attachment; filename*=UTF-8''filename.txt
attachment; filename="EURO rates"; filename*=utf-8''%e2%82%ac%20rates
attachment; filename="omáèka.jpg"
and some other combinations might also be there

Answer 1

You could try something in this spirit:

filename[^;=\n]*=((['"]).*?\2|[^;\n]*)

filename      # match filename, followed by
[^;=\n]*      # anything but a ;, a = or a newline
=
(             # first capturing group
    (['"])    # either single or double quote, put it in capturing group 2
    .*?       # anything up until the first...
    \2        # matching quote (single if we found single, double if we find double)
|             # OR
    [^;\n]*   # anything but a ; or a newline
)

Your filename is in the first capturing group: http://regex101.com/r/hJ7tS6

Answer 2

Slightly modified to match my use case (strips all quotes and UTF tags)

filename\\*?=['"]?(?:UTF-\\d['"]*)?([^;\\r\\n"']*)['"]?;?

https://regex101.com/r/UhCzyI/3

Answer 3

/filename[^;=\n]*=(?:(\\?['"])(.*?)\1|(?:[^\s]+'.*?')?([^;\n]*))/i

https://regex101.com/r/hJ7tS6/51

Edit : You can also use this parser: https://github.com/Rob--W/open-in-browser/blob/master/extension/content-disposition.js

Answer 4

Disclaimer: the following answer only works with PCRE (eg Python / PHP), if you have to use javascript, use Robin's answer.

---

This modified version of Robin's regex strips the quotes:

filename[^;\n=]*=(['\"])*(.*)(?(1)\1|)

filename        # match filename, followed by
[^;=\n]*        # anything but a ;, a = or a newline
=
(['"])*         # either single or double quote, put it in capturing group 1
(?:utf-8\'\')?  # removes the utf-8 part from the match
(.*)            # second capturing group, will contain the filename
(?(1)\1|)       # if clause: if first capturing group is not empty,
                # match it again (the quotes), else match nothing

https://regex101.com/r/hJ7tS6/28

The filename is in the second capturing group.

Answer 5

Here is my regular expression. It works on Javascript.

filename\*?=((['"])[\s\S]*?\2|[^;\n]*)

I used this in my project.

Answer 6

filename[^;\n]*=(UTF-\d['"]*)?((['"]).*?[.]$\2|[^;\n]*)?

I have upgraded Robin's solution to do two more things:

1. Capture filename even if it has escaped double quotes.

2. Capture UTF-8'' part as a separate group.

This is an ECMAScript solution.

https://regex101.com/r/7Csdp4/3/

Answer 7

I made a regex that finds these names using a group filename

/(?<=filename(?:=|\*=(?:[\w\-]+'')))["']?(?<filename>[^"';\n]+)["']?/g

 const regex = /(?<=filename(?:=|\\*=(?:[\\w\\-]+'')))["']?(?<filename>[^"';\\n]+)["']?/g const filenames = ` attachment; filename=content.txt attachment; filename*=UTF-8''filename.txt attachment; filename="EURO rates"; filename*=utf-8''%e2%82%ac%20rates attachment; filename="omáèka.jpg" ` function logMatches(){ const array = new Array filenames.split("\\n").forEach(line => { if(!line.trim()) return const matches = line.matchAll(regex) const groups = Array.from(matches).map(match => match?.groups?.filename) array.push(groups.length === 1 ? groups[0] : groups) }) console.log(array) } logMatches()