Re: FireFox password management 2

Paul wrote:
> On 1/30/2022 4:26 AM, Robert Baer wrote:
>> Paul wrote:
>>> On 1/29/2022 12:31 AM, Robert Baer wrote:
>>>> Paul wrote:
>>>>> On 1/19/2022 12:12 AM
>>>>>
>>>>>>>    Maybe because Linux is involved?
>>>>>>>
>>>>>>>    Thanks.
>>>>>
>>>>> The same test results in a Win10 VM.
>>>>>
>>>>>     [Picture]  Right-click empty frame and select "Reload picture" 
>>>>> to see picture
>>>>>
>>>>>     https://i.postimg.cc/qRnm0S5m/Win10-FF96-Password-Save-Test.gif
>>>>>
>>>>>    Paul
>>>>    FireFox 96.0 in Win7.1 does not save passwords, and for some 
>>>> sites does not even save user names.
>>>>    It also seems that the online "help" avoids that fact.
>>>>
>>>>    So, how far back does one need to go (ie: what version) to get a 
>>>> FireFox that does work properly?
>>>>
>>>>     Thanks.
>>>
>>> In your view, how would the OS influence the outcome
>>> of an event which is private to the software (Firefox) ?
>>> The passwords don't pass through the OS. The SChannel is
>>> occasionally used for crypto, but if it were used (which it
>>> probably isn't), it would only be used for datapath (https
>>> transmission). The password level is too far removed
>>> from the OS, for the OS to be the issue. If the passwords
>>> were "coming from User Accounts", again it would be a different
>>> matter, but the passwords have nothing to do with User Accounts
>>> in the OS.
>>>
>>> The only thing I can suggest, is a disagreement about
>>> the file contents of the Profile folder that Firefox is
>>> currently using.
>>>
>>> Profiles are "migrated", each time that Firefox is upgraded.
>>> This includes (very occasionally) adding a new database
>>> file for things like passwords. Any time the API changes,
>>> the version number at the end of the filename changes.
>>> Like "cert8" or similar. The files used consist of a name
>>> with a number on the end, and the number is a hint as to
>>> which API level it is at.
>>>
>>> You also claim a level of inconsistency, almost as if the
>>> password file is corrupted somehow.
>>>
>>> I would create a new profile, and retest.
>>>
>>> Firefox has a Profile Manager. Create a new profile and retest.
>>> Switch back to your old profile, to access your Bookmarks.
>>> You can Export Bookmarks and Import them, if you need some
>>> information preservation.
>>>
>>> https://support.mozilla.org/en-US/kb/profile-manager-create-remove-switch-firefox-profiles 
>>>
>>>
>>>     Paul
>>    Well, in essence, nothing works.
>>    Tried it all (restart, refresh, profile).
>>    Installed FireFox 80. Same problems.
>>
>>    Now get:
>> ** COPY **
>> Secure Connection Failed
>> ** END COPY **
>>    for almost every URL that used to work.
>>    For one site, i get there and try to login, and get:
>> ** COPY **
>> Access Denied
>>
>> You are unauthorized to access this resource.
>>
>> Reference ID is: d14b2c2729
>> ** END COPY **
>>
>>    In fact, a URL that worked an hour ago now gives me the URL message.
>>
>>    This funnybusiness is a bunch of crap.
>>
>>    Is my Win7.1 getting corrupted?
>>
>>     Thanks.
>>
>>    Now
> 
> So you're having a LetsEncrypt certificate issue ?
> 
> What you describe is not a password issue.
> 
> You can't have passwords, without "Transport" first.
> There has to be a "pipe" to pass the password to the host.
> 
> The https encrypted transport, would help protect the
> password in flight. To otherwise set up a secure path
> to pass a password, would be a duplication of effort.
> Might as well set up the "good" transport first, then
> pass the password.
> 
> You can't have https: transport, without crypto (CHACHA20),
> TLS1.2, and a certificate that testifies the web site is
> what it says it is.
> 
> This URL tests a browser. The design of this page has apparently
> been stripped, so it's no longer as informative.
> 
>     https://www.ssllabs.com/ssltest/viewMyClient.html
* Basically sez it is OK. So i did not go further, but i will keep this
posting and not delete it.

> 
> To test a website:
> 
>     https://www.ssllabs.com/ssltest/analyze.html?d=walmart.com
> 
> Some browsers might use the platform certificates, but the
> others use their own internal certificate stores.
> 
> https://support.microsoft.com/en-us/topic/support-for-urgent-trusted-root-updates-for-windows-root-certificate-program-in-windows-a4ac4d6c-7c62-3b6e-dfd2-377982bf3ea5 
> 
> 
> Checking or repairing browser certificates manually, is no
> fun at all.
> 
> When you open an https: page (or get redirected to an https:
> page when you typed http: instead), look in the lower
> right corner of the web page for the "lock symbol". If the
> lock symbol is closed, then the transport is likely working.
> If the lock symbol is open, you can always click on the
> lock symbol and get details.
> 
> I'm no expert on this stuff. My browser breaks just
> like your browser breaks, and is likely to stay broken
> for days at a time. Finding a good recipe for repairs
> can take a *lot* of searches.
> 
> HTH,
>     Paul

*  Thanks!