Sebastian Nagel commented on NUTCH-2550:

Confirmed, thanks [~HansBrende]! Sorry, that this issue was overseen in the 
changes for NUTCH-2375. After a look I've found another issue (NUTCH-2553) and 
the way URLs are hold in FetcherThread.run() needs a general review: there are 
now 4 variables (fit.url, fit.u, url, urlString) which makes it difficult to 
avoid such issues.

> Redirects are broken
> --------------------
>                 Key: NUTCH-2550
>                 URL: https://issues.apache.org/jira/browse/NUTCH-2550
>             Project: Nutch
>          Issue Type: Bug
>          Components: fetcher
>    Affects Versions: 1.15
>            Reporter: Hans Brende
>            Priority: Blocker
>             Fix For: 1.15
> As I detailed in this github 
> [comment|https://github.com/apache/nutch/commit/c93d908bb635d3c5b59f8c8a22e0584ebf588794#r28470348],
>  it appears that PR #221 broke redirects. The fetcher will repeatedly fetch 
> the *original url* rather than the one it's supposed to be redirecting to 
> until {{http.redirect.max}} is exceeded, and then end with 
> I noticed this issue when I was trying to crawl a site with a 301 MOVED 
> PERMANENTLY status code.
> Should be pretty easy to fix though: I was able to get redirects working 
> again simply by inserting the code {code:java}url = fit.url{code} 
> [here|https://github.com/apache/nutch/blob/8682b96c3b84018f187eabaadc096ceded34f250/src/java/org/apache/nutch/fetcher/FetcherThread.java#L388]
>  and 
> [here|https://github.com/apache/nutch/blob/8682b96c3b84018f187eabaadc096ceded34f250/src/java/org/apache/nutch/fetcher/FetcherThread.java#L409].

This message was sent by Atlassian JIRA

Reply via email to