vbauer / manet

Website screenshot service powered by Node.js, SlimerJS and PhantomJS
MIT License
575 stars 102 forks source link

problem with some weird redirected URLs (BBC) #23

Closed jfix closed 8 years ago

jfix commented 9 years ago

Hello, I'm noticing some weird redirects with URLs from the BBC:

At least from my location, in France, when I attempt to go here http://www.bbc.co.uk/news/business-31599492 I get redirected to http://www.bbc.com/news/business-31599492. However it's not a classic redirect (301 or 302), the web inspector in Chrome doesn't show this redirect, but ends up with the .com page.

Now, how does this concern manet? Well using manet, I don't get a screenshot, but a 404. Not sure that something can be done on manet's level, but I would still be interested?

Thanks, Jakob.

vbauer commented 9 years ago

It looks like there is an open bug in PhantomJS: https://github.com/ariya/phantomjs/issues/10389 Unfortunately, I couldn't check it now with SlimerJS. I'll try it asap.

jfix commented 9 years ago

@vbauer thanks for following this up. As suggested over there, adding a delay (I used 5000 msec, just to be sure) actually works in this particular case. As I'm not after real-time synchronous screenshots, that's a feasible workaround for me.

vbauer commented 9 years ago

I've just tried to investigate this problem and found a possible solution: --ignore-ssl-errors=true.

If it doesn't help, could you please try to add "--ssl-protocol=tlsv1" in the config/default.json

jfix commented 9 years ago

Sorry for the delay of my response. I just tried with the new configuration option (I tried both actually, first --ignore-ssl-errors=true then in addition also --ssl-protocol=tlsv1), but the results are non-conclusive:

info: Capture site screenshot: www.bbc.co.uk/news/business-31599492
error: Error while sending file: ENOENT, stat '/tmp/eyJ1cmwiOiJodHRwOi8vd3d3LmJiYy5jby51ay9uZXdzL2J1c2luZXNzLTMxNTk5NDkyIn0=.png'

I tried with both the protocol http:// and without, just to make sure I also tried just google.com which confirmed that the manet installation itself is not broken...

Here is what Chrome sees, apparently it's a 404 ... hmmmm ....

screen shot 2015-09-10 at 22 07 44

vbauer commented 9 years ago

@jfix Thank you for information. I've just fixed wrong HTTP status codes, but I still do not understand why it doesn't work..

jfix commented 9 years ago

@vbauer does it work for you?

cheers, Jakob.

On Fri, Sep 11, 2015 at 2:22 AM, Vladislav Bauer notifications@github.com wrote:

@jfix https://github.com/jfix Thank you for information. I've just fixed wrong HTTP status codes, but I still do not understand why it doesn't work..

— Reply to this email directly or view it on GitHub https://github.com/vbauer/manet/issues/23#issuecomment-139417141.

vbauer commented 9 years ago

@jfix Yesterday I committed possible workaround https://github.com/vbauer/manet/commit/a4dbc36bf1f36dbd9f96a4e2cdc101172f1122b5 and it worked, but today as I see it doesn't work. Strange..

jfix commented 9 years ago

Hmm, I just tried this additional option and it didn't change a thing, unfortunately. :-( Thanks for trying!

vbauer commented 8 years ago

Could you please try to test it with last commit (https://github.com/vbauer/manet/commit/5853aa6d526a555659891b76d8772a48aa03ae10)?

vbauer commented 8 years ago

JFYI: Now, it is possible to install the release version (0.4.4) to check it.

jfix commented 8 years ago

Thanks @vbauer, unfortunately I cannot currently test. If it works for you, then please close the issue. Once I'm able to pick this project up again, I will test and if necessary come back to this issue. Thanks a lot for this!