01-10-2017, 03:49 PM | #16 |
null operator (he/him)
Posts: 20,936
Karma: 27620688
Join Date: Mar 2012
Location: Sydney Australia
Device: none
|
Not sure if this problem belongs to MR, Kovid Goyal or me. If I click the word 'Calibre' at the start of the Calibre Forum description I get security related errors (see attachment). As well as getting it in Firefox, I get it in Chrome, Edge, Qupzilla and IE.
I clicked on it inadvertently, so it might not be a new thing. Although I think my mouse has clicked just under the forum name before and I landed at ==>> http://calibre-ebook.com Similar links for Calibre Companion and Sigil are fine. Added - IMO putting links behind plain text is poor UI design, they should be differentiated. In this instance a link at the end of the text would be preferable - eg Calibre Home Page. BR Last edited by BetterRed; 01-10-2017 at 04:02 PM. |
01-10-2017, 04:07 PM | #17 | |
Fully Converged
Posts: 18,170
Karma: 14021202
Join Date: Oct 2002
Location: Switzerland
Device: Too many to count here.
|
Quote:
Just a guess since I didn't get this error when following the old link... is it possible that you are using some kind of security software like Avast, Kaspersky, or BitDefender? These like to intercept connections and downgrade HTTPS connections into HTTP (so that they can scan the traffic). Many sites, including ours, don't allow a protocol downgrade for security purposes (the mechanism we use is referred to as HSTS). Security vendors usually offer root certificates to be installed locally to get around such mechanisms so you can still use them to scan traffic. Read further below. |
|
01-10-2017, 04:17 PM | #18 |
Fully Converged
Posts: 18,170
Karma: 14021202
Join Date: Oct 2002
Location: Switzerland
Device: Too many to count here.
|
Actually a more likely scenario is:
The old link used to be http://calibre.kovidgoyal.net/. Today when you clicked on that link it should have redirected you to http://calibre-ebook.com. The issue is that calibre.kovidgoyal.net is a subdomain of kovidgoyal.net. https://kovidgoyal.net exists and uses headers to tell your browser that it must follow only HTTPS connection of its domain and subdomains: Code:
$ curl -svo /dev/null https://kovidgoyal.net/ * Trying 166.78.104.224... * Connected to kovidgoyal.net (166.78.104.224) port 443 (#0) * found 173 certificates in /etc/ssl/certs/ca-certificates.crt * found 697 certificates in /etc/ssl/certs * ALPN, offering http/1.1 * SSL connection using TLS1.2 / ECDHE_RSA_AES_128_GCM_SHA256 * server certificate verification OK * server certificate status verification SKIPPED * common name: kovidgoyal.net (matched) * server certificate expiration date OK * server certificate activation date OK * certificate public key: RSA * certificate version: #3 * subject: CN=kovidgoyal.net * start date: Sun, 08 Jan 2017 06:45:00 GMT * expire date: Sat, 08 Apr 2017 06:45:00 GMT * issuer: C=US,O=Let's Encrypt,CN=Let's Encrypt Authority X3 * compression: NULL * ALPN, server accepted to use http/1.1 > GET / HTTP/1.1 > Host: kovidgoyal.net > User-Agent: curl/7.47.0 > Accept: */* > < HTTP/1.1 200 OK < Server: nginx < Date: Tue, 10 Jan 2017 21:12:23 GMT < Content-Type: text/html; charset=utf-8 < Content-Length: 2456 < Last-Modified: Thu, 25 Aug 2016 23:34:10 GMT < Connection: keep-alive < ETag: "57bf8072-998" < Strict-Transport-Security: max-age=10886400; includeSubdomains < Accept-Ranges: bytes < { [2456 bytes data] * Connection #0 to host kovidgoyal.net left intact Because there is no HTTPS version of the old link (try: https://calibre.kovidgoyal.net/) you get the error you're seeing. Anyways... thanks again, replacing the link fixed the issue. |
01-10-2017, 06:04 PM | #19 | |
Wizard
Posts: 2,657
Karma: 73864785
Join Date: Dec 2010
Location: PDXish
Device: Kindle Voyage, various Android devices
|
Quote:
These are generated from Google Docs using the "Publish Chart" as "image", I believe it publishes them as .PNG files. They do update automatically though, which is why I have used them, so that might be a problem with a proxy/cache. I will leave them there for now (at least until you are done testing). If they don't work, they don't work. Not a big deal, just something I have done for my own amusement. I could easily update them manually as fixed images a couple times a year instead. |
|
01-10-2017, 08:04 PM | #20 |
null operator (he/him)
Posts: 20,936
Karma: 27620688
Join Date: Mar 2012
Location: Sydney Australia
Device: none
|
|
01-11-2017, 03:15 AM | #21 |
Fully Converged
Posts: 18,170
Karma: 14021202
Join Date: Oct 2002
Location: Switzerland
Device: Too many to count here.
|
Generally the problem with these spreadsheet links is that Google occasionally times out loading them or just slows down a lot. When it's too slow (the proxy client has a timeout value of 3s - we could set that to a higher number but that could lead to other negative side effects), the proxy considers the connection to have timed out until the next time you try loading it. Try this. Go to
https://www.mobileread.com/forums/sh...78&postcount=9 And click on reload while holding your CTRL/STRG button. This will force reloading the images from the origin instead of loading from the browser cache. In other words, these specific images work and occassionally don't when the response from Google is too slow. My guess is that Google recalculates these graphs based on your spreadsheet every you load the image and that they only briefly cache them before recalcuating them. I can imagine this to be quite a hit on their servers. Is there a way to generate those images statically and only refresh them when new data is entered in the spredsheet? |
01-11-2017, 10:41 AM | #22 | |
Wizard
Posts: 2,657
Karma: 73864785
Join Date: Dec 2010
Location: PDXish
Device: Kindle Voyage, various Android devices
|
Quote:
I will check on getting a static image or I may just drop them entirely if they don't work reliably. When they first rolled out the feature it behaved similar to this but had been reliable for the last year plus. |
|
01-11-2017, 10:51 AM | #23 |
Fully Converged
Posts: 18,170
Karma: 14021202
Join Date: Oct 2002
Location: Switzerland
Device: Too many to count here.
|
Did you try to refresh before or after posting in this thread? Because initially there was indeed a bug where the proxy received the image URL with escaped HTML characters which wasn't supposed to be. The other issue with the image not always showing is related to the irregular reponse time from Google.
|
01-11-2017, 12:32 PM | #24 | |
Wizard
Posts: 2,657
Karma: 73864785
Join Date: Dec 2010
Location: PDXish
Device: Kindle Voyage, various Android devices
|
Quote:
|
|
01-11-2017, 12:35 PM | #25 |
Fully Converged
Posts: 18,170
Karma: 14021202
Join Date: Oct 2002
Location: Switzerland
Device: Too many to count here.
|
Working fine here... Did you try a few times? It seems like the first time one requests the image from Google, they will create it from the spreadsheet. Once it's created, it should load quickly for as long as they have it cached. After that the same thing... first request will timeout, eventually the request will work when the image is created, and so on.
|
01-11-2017, 03:05 PM | #26 |
Grand Sorcerer
Posts: 27,903
Karma: 198500000
Join Date: Jan 2010
Device: Nexus 7, Kindle Fire HD
|
Loading Dazrin's post images is kind of hit or miss on my machines, but refreshing while holding the crtl key always brings up the images for me.
|
01-14-2017, 07:07 AM | #27 |
Chocolate Grasshopper ...
Posts: 27,599
Karma: 20821184
Join Date: Mar 2008
Location: Scotland
Device: Muse HD , Cybook Gen3 , Pocketbook 302 (Black) , Nexus 10: wife has PW
|
Good to see https working successfully - good job well done all round .
Don't envy your workload to change to xenforo though will be interested to see it go through |
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
M92 stopped loading HTTPS pages | CJY | Onyx Boox | 32 | 01-22-2015 10:12 AM |
Classic HTTPS/SSL with Classic Nook | cdstech69 | Barnes & Noble NOOK | 1 | 02-16-2011 08:40 AM |