With Firefox 45.0, with or without add-ons (safe mode), I can connect to
that site okay. Have you purged all Firefox caches (clear on exit, or
use a cleaner, like CCleaner) and retried a connection to that site?
They may have had a momentarily problem.
Do you use the HTTPS-Everywhere extension? If so, has it been recently
updated? Sometimes that extension will break a site so you have to wait
until they fix their rules (or the site owner to change something on
their end). For example,
http://www.virubtn.com/ uses the Location
header to redirect visitors to their new site at
https://www.virusbulletin.com. They redirect from HTTP (old site) to
HTTPS (new site). The problem is that the rule in HTTPS-Everywhere only
converted http: to https: but the old
virusbtn.com site doesn't support
HTTPS. So when HTTPS-Everywhere changed
http://www.virubtn.com to
https://www.virusbtn.com then I would get a "server not found" error. I
contacted both Virus Bulletin and HTTPS-Everywhere about the failed
redirect by HTTPS-Everywhere. I do not see that HTTPS-Everywhere
changed their rule. It still simply changes http: to https: (converts
http://www.virusbtn.com to
https://www.virusbtn.com) but it looks like
the site owner made some change to his old site. Using the Location
header on his old HTTP site that pointed to his new HTTPS site worked if
HTTPS-Everywhere was involved. See my discussion at:
https://github.com/EFForg/https-everywhere/issues/4273
continued at:
https://github.com/EFForg/https-everywhere/pull/4280
Seems the rule set for HTTPS-Everywhere has to keep getting updated via
user reports where this extension breaks a web site. In the virusbtn
case, their rule was for the old site (when it did support or have a
valid cert for the HTTPS connect to that site). Then the site went to a
different domain and the old rule (still the current rule) was no longer
valid (until the site owner made a change). I've hit way too many sites
where HTTPS-Everywhere causes problems (usually error pages) that I will
probably discard it. One, it obviously only works at the limited number
of sites for which it has rules. It does not blanket switch all http:
requests to https: requests. No matter how many rules they have, they
will never approach the number of web sites that exist even if only for
those that support HTTPS. So it really is misnamed as HTTPS-Everywhere
and should really be named HTTPS-WhereWeKnowAbout.
Way over a decade ago, Internet Explorer had options to determine if any
mixed (active and image) content was allowed in a supposedly HTTPS
secure web page. Mixed content means HTTP content delivered in an HTTPS
web page: you think the page is secure, see the lock icon, but some
content is not secure. A decade later Mozilla added user-configurable
options for mixed content (HTTP content delivered with a supposedly
HTTPS-secured web page), by default Firefox only blocks *active* mixed
content (security.mixed_content.block_active_content) and not images
(security.mixed_content.block_display_content). That is because LOTS of
sites have insecure images included in their secure web page. For
example, when looking at offers at
craigslist.org, their web pages
appear secured but you won't see any images if you also block insecure
content (i.e., if blocking mixed content includes both active and image
content). To see the images at Craiglist, you need to have Firefox (or
any web browser) configured to block mixed (insecure) active content but
allow mixed (insecure) image content. Mozilla doesn't want to break
lots of site that pretend to have secure (HTTPS) content but instead
deliver mixed content.
Mixed content means the secure page is not secure. A page is secure or
it is not, not somewhere between. Because Firefox, by default, allows
some insecure content (images), you'll see a new lock icon at the left
end of the address bar in version 45. On sites, like Craigslist, that
deliver mixed content, the lock icon appears not as green (meaning fully
secure - no mixed content, including no insecure images) but as green
with a yellow hazard overlay. Looking at the details of the partially
secure lock icon doesn't tell you want content was secure. You get an
indication that insecure images are at fault in the message "Parts of
this page are not secure (such as images)". That doesn't explicitly
state that insecure images is the culprit of mixed content (which means
the secure page is not secure).
If you decide to configure Firefox to also block mixed content for
images (security.mixed_content.block_display_content = True) then you
will find lots of sites where images are missing and replace by blank
placeholders.
You should check if whatever security software (anti-virus/malware) that
you use has the ability to interrogate HTTPS traffic. For example,
Avast can do that if the HTTPS scan option is enabled. They install a
cert used in a MITM (Man In The Middle) attack scenario that lets them
intercept the HTTPS traffic to inspect for nefarious content. For some
reason, Mozilla decided to use their NSS tools to manage a private
certificate store used by Firefox instead of using the Windows cert
store (as do IE and Google Chrome). If the antimalware with HTTPS
scanning doesn't insert its cert into Firefox's private cert store than
all HTTPS connects will fail. However, you only mentioned a problem at
a single HTTPS site, not that you had problems at all HTTPS sites. You
gave one site as an example. Was that an example showing what happens
when you visit any HTTPS web site?