From 93e6bd68477dd34a48bc066a1ece07e6a84879af Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Mike=20F=C3=A4hrmann?= Date: Sat, 5 Nov 2022 18:55:38 +0100 Subject: [PATCH] [uploadir] use utf-8 filenames (#3162) --- gallery_dl/extractor/uploadir.py | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/gallery_dl/extractor/uploadir.py b/gallery_dl/extractor/uploadir.py index eeafa127..bd18c0ac 100644 --- a/gallery_dl/extractor/uploadir.py +++ b/gallery_dl/extractor/uploadir.py @@ -41,6 +41,16 @@ class UploadirFileExtractor(Extractor): "id": "gxe8ti9v", }, }), + # utf-8 filename + ("https://uploadir.com/u/fllda6xl", { + "pattern": r"https://uploadir\.com/u/fllda6xl", + "count": 1, + "keyword": { + "extension": "png", + "filename": "_圖片_🖼_image_", + "id": "fllda6xl", + }, + }), ("https://uploadir.com/uploads/rd3t46ry"), ("https://uploadir.com/user/uploads/rd3t46ry"), ) @@ -71,7 +81,9 @@ class UploadirFileExtractor(Extractor): else: hcd = response.headers.get("Content-Disposition") - data = text.nameext_from_url(text.extr(hcd, 'filename="', '"')) + name = (hcd.partition("filename*=UTF-8''")[2] or + text.extr(hcd, 'filename="', '"')) + data = text.nameext_from_url(name) data["id"] = self.file_id yield Message.Directory, data