pyotr filipivich wrote:

Every so often, say once or twice a week, Mailwasher.exe hangs up at
the splash screen. Just, sits there.

I've right clicked on the taskbar icon and chosen "exit", clicked on
the red x to close the window. Used Task Manager and Process Hacker,
separately, and sequentially, to attempt to kill the process. No
luck.I've logged out or selected "Restart" as the shutdown option,
and windows has hung up on the "logging out" screen. Normally a
reboot takes long enough I can go warm up a cup of coffee on the
stove, sometimes make a fresh pot. But in this case, I have time to
brew a Fresh Cup from scratch. As in go to Columbia, pick the beans
myself, dry, roast, and grind them. I'm assuming that, I haven't
waited two weeks, I just kill the power when I get back from getting
coffee.

Windows issues a request for processes to exit. If they don't respond,
Windows should timeout and give you an option to kill the hanging
process(es). You don't see that timeout prompt asking if your want to
forcibly kill the unresponsive process(es)?

Could be your installation of Mailwasher has become corrupted. You
never identified which edition of Mailwasher you have: free or paid.
They have a different feature set. For example, the free version only
support *one* e-mail account. It has been decades since I had a single
e-mail account. If it is a paid edition, well, support is included so
call them, e-mail them ), or use their forums.

Because until it is dead, I can't start it again.

Hold down the Power button for 4 seconds to force a cold reboot.

What I am wondering is, is there a way to determine what
services/modules/handles one might be able to track, hunt down and
terminate, unload, kill, in order to closed down the process, process
tree, or whatever is holding the splashscreen / program open?

I've attempted to close modules through Process Hacker, as well as
terminate specific threads, but no change.

Maybe it uses 2 processes. Each checks for the presence of the other.
If the other process is missing, it gets started. Meanwhile you kill
one process and the other restarts it.

Review all processes to see to which programs they belong. If you find
more than one process is linked back to a Mailwasher program, create a
shortcut that runs:

taskkill.exe /im process1 /f & taskkill.exe /im process2 /f

The /f forces a kill on the process. The & means to run multiple
commands in one line (without waiting for the prior command to complete
and regardless of the return status or errorlevel of the other command).
/im means to kill the process with THAT image name. So if the two
processes were mailwasher.exe and mw.exe, you would substitute those
process names (including the .exe extension) as process1 and
process2.

The kills should happen very quickly to prevent the remaining program
yet to get killed from having enough time to notice its mate program has
been killed.

Did you check Windows services for Mailwasher? If it runs as a service,
it can restart.

Did you disable your anti-virus program (permanently - until you
reenable it yourself), reboot, and retest if Mailwasher still hangs?

Did you check Task Scheduler to make sure there are no events defined
there to load Mailwasher (and perhaps repeatedly reload it)?

msconfig.exe only gives some startup locations. It doesn't show events
defined in Task Scheduler. It doesn't show logon events (defined in the
registry). It doesn't show other events assigned to objects (e.g.,
folder open and your desktop is a folder). If you want to see them all,
use SysInternals' AutoRuns.

Reboot Windows into its safe mode. Then uninstall Mailwasher. You
might want to use Revo Uninstaller (free) to both run Mailwasher's
uninstaller followed by a medium cleanup. Just be sure about which
remnant file and registry keys to delete: they should be pathed back to
the Mailwasher install path, not simply referenced by another program
(like your AV having an entry regarding its heuristic analysis of
Mailwasher). Then boot Windows into its normal mode. Check there is no
hanging on login or logout.

Reinstall Mailwasher if you still want it. Personally I prefer
server-side anti-spam solutions. Chaining together more clients (e-mail
client to local proxy to server) means a more fragile chain. I've often
seen the local proxy become unresponsive or crash (and not just for
e-mail but also some AVs) and the user then reports they cannot get any
e-mail (or all web traffic is dead for the AV proxy). If whomever you
use for e-mail service is poor regarding spam filtering, you could chain
together multiple e-mail providers to make use of each one's spam
filtering. You aggregate their spam filters. For example, you could
create and use a Gmail account to poll your other e-mail account. The
e-mail arrives at your other account, gets processed through that spam
filter (which presumably is enabled there but apparently you believe is
insufficient), Gmail polls your other account, and Google's spam
filtering gets employed on those e-mails. While your local e-mail
client would be polling Gmail for inbound e-mails, you can configure
your local e-mail client to send new or reply e-mails using either
account. So, by default, you would receive via Gmail but could send via
your other account (so recipients don't see a change in your sending
e-mail account). If you add the other account to Gmail as a sending
account, you can receive and send using Gmail. That doesn't mean you
have to change the From header added by your local e-mail client to
identify your e-mail address. Even when sending using Gmail, your From
header can still show your old e-mail address.

I don't know which public DNSBL (DNS blacklist) uses for their spam
filtering. I knew it when I trialed Mailwasher Free many years ago.
Fire Trust provides no donations to the blacklist provider despite they
incorporate it into their commercial version of Mailwasher. With other
local anti-spam solutions, you can specify which DNSBLs are employed in
the spam filtering. Back when I used those, I used Spamhaus Zen and
SpamCop, in that order. Spamhaus Zen caught almost everything that
Spamcop caught but once in a while Spamcop caught something earlier than
Spamhaus. That's because Spamcop has a userbase of spam reporters that
help update their database.

Another solution (used in addition to DNSBLs) is the Rhyolite DCC; see
https://en.wikipedia.org/wiki/Distri..._Clearinghouse. It's
really not a blacklist but instead a database updated by users to rate
the spaminess of an e-mail based on how many users received the nearly
same e-mail (fuzzy logic is used to determine if an e-mail is very
similar to another). It really is a automatic "bulk mail" voting scheme
where each user of DCC creates a hash of a received e-mail and adds the
hash to the DCC database. When a user gets an e-mail, they can see how
many other received the same e-mail. This lets users see if an e-mail
was sent in bulk. However, many confirmation messages, like an e-mail
from registering at a forum within which you need to click a link to
complete the registration, will look the same or close enough to qualify
as bulk mailing.

Other local anti-spam programs (ran as clients or proxies) can make use
of multiple DNSBLs and even incorporate DCC. Mailwasher uses only one
DNSBL (to which they don't donate nor contribute their discoveries).
From:

http://help.firetrust.com/en/product...nformation.htm
The cache subfolder
MWP.db3 - This is an SQLite database which contains your Friends List,
Blacklist, cached results from FirstAlert and DNSBLs, hash tables for
deleted and cached emails and other miscellaneous settings.

it appears FireTrust has added the option to specify more than one
DNSBL, plus they have their own. Bayesian filtering is also mentioned.
LOTS of anti-spam programs and even e-mail clients (e.g., Outlook
[Express], Thunderbird) have their own Bayes filter. That is a
statistically based guessing scheme. Keywords are added to a Bayes
database to weight words as spam or ham which is then used in later
e-mails to rate them as spam or ham. Some Bayes filters get
pre-populated with a keywords database (e.g., Outlook gets a monthly
update) while some start empty and require a learning period within
which the user has to keep voting e-mails as spam or ham. Eventually
the database gets large enough to make better guesses but they are still
guesses so it is still the responsibility of the user to review the Spam
folder (or Pending folder in the anti-spam proxy) to determine if there
were false positives (ham marked as spam). Bayesian filtering has been
around a long time. It should be the LAST method used to detect spam.
DNSBLs are more accurate. Some DNSBLs are better than others. Spamhaus
Zen and Spamcop are good. They will expire entries after awhile because
spammers move. A good Bayesian filter will have a "floor function" to
do the same expiration of old keywords that have not been present in any
received e-mails for a while. Alas, many anti-spam programs or client
using Bayes do not have a floor or expiration feature so ancient spam no
longer relevant to your current e-mail traffic can improperly flag ham
as spam.