GitHub-Proxy/jina_reader
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/jina-ai/reader.git synced 2025-08-19 04:16:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: fail early on special cookie redirects

Browse Source
This commit is contained in:
Yanlong Wang 2025-03-08 18:43:10 +08:00
parent 26f6202f79
commit e551695d17
No known key found for this signature in database
GPG Key ID: C0A623C0BADF9F37
1 changed files with 10 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
src/services/curl.ts
View File

@ -294,7 +294,8 @@ export class CurlControl extends AsyncService {
} }
async urlToFile(urlToCrawl: URL, crawlOpts?: CURLScrappingOptions) { async urlToFile(urlToCrawl: URL, crawlOpts?: CURLScrappingOptions) {
let leftRedirection = 10; let leftRedirection = 6;
let cookieRedirects = 0;
let opts = { ...crawlOpts }; let opts = { ...crawlOpts };
let nextHopUrl = urlToCrawl; let nextHopUrl = urlToCrawl;
const fakeHeaderInfos: HeaderInfo[] = []; const fakeHeaderInfos: HeaderInfo[] = [];
@ -312,10 +313,16 @@ export class CurlControl extends AsyncService {
if (parsed.length) { if (parsed.length) {
opts.cookies = [...(opts.cookies || []), ...parsed]; opts.cookies = [...(opts.cookies || []), ...parsed];
} }
if (!location) {
cookieRedirects += 1;
}
} }
if (!location && !setCookieHeader) { if (!location && !setCookieHeader) {
throw new AssertionFailureError(`Failed to access ${urlToCrawl}: Bad redirection from ${nextHopUrl}`); throw new ServiceBadAttemptError(`Failed to access ${urlToCrawl}: Bad redirection from ${nextHopUrl}`);
}
if (!location && cookieRedirects > 1) {
throw new ServiceBadAttemptError(`Failed to access ${urlToCrawl}: Browser required to solve complex cookie preconditions.`);
} }
nextHopUrl = new URL(location || '', nextHopUrl); nextHopUrl = new URL(location || '', nextHopUrl);
@ -331,7 +338,7 @@ export class CurlControl extends AsyncService {
}; };
} while (leftRedirection > 0); } while (leftRedirection > 0);
throw new AssertionFailureError(`Failed to access ${urlToCrawl}: Too many redirections.`); throw new ServiceBadAttemptError(`Failed to access ${urlToCrawl}: Too many redirections.`);
} }
async sideLoad(targetUrl: URL, crawlOpts?: CURLScrappingOptions) { async sideLoad(targetUrl: URL, crawlOpts?: CURLScrappingOptions) {