update 1.1.9.3

[Rudi De Vos]

Test builds, please notify if installer is not working.

1.1.9.2
mirror driver and left secondary screen
u2 encoder missing updates
1.1.9.3
factory reset when temp is not writable
manual refresh when driver is selected but not activated

DOWNLOAD test builds
http://www.uvnc.com/downloads/ultravnc.html

[Rudi De Vos]

msi installers added

DOWNLOAD test builds
http://www.uvnc.com/downloads/ultravnc.html

[Rudi De Vos]

Digital signature updated
All exe dll and installers are now signed with a VeriSign Class 3 code Signing CA ( valid till 10/2016).

DOWNLOAD test builds
http://www.uvnc.com/downloads/ultravnc.html

[itsmee]

Hi,

Just installed, the factory re-set of settings was the big one for me, will feed back when I have ran it for a bit...

A note on the installer, if you just select for the viewer to be installed you still get asked if you want to install as a service and use CAD. Could be confusing for noobs...

Ta,

[Prisma]

Rudi, would you please update the zipped binaries? Especially Win2k ZIP includes only 1.1.9.1

[Rudi De Vos]

The testbuilds are always on uvnc.eu for easy access.

The final version, is placed onn our download server..
http://www.uvnc.com/downloads/ultravnc/ ... -1190.html
Else we can't make any $ with advertising...

I still need to remove this old folder to avoid confusing.

[Prisma]

No question, you have to make money, because we need you absolutely, now and in future.

Anyway, the w2k binaries on .com are also 1.1.9.1. Signed on ‎Wednesday, ‎14. ‎August ‎2013 15:26:18. Or is there only a problem with the program informations?

[Rudi De Vos]

thanks for the feedback, i forgot to update the w2k build.
With the signing ( that change the date) i don't know if the w2K version is correct with wrong version info or not.
Gonna update all zip/installer that contain the w2k server.

[Rudi De Vos]

updated the zip w2k and setup x86 ( on www.uvnc.com)

[plants]

Hi Rudi,
Could you confirm whether the Win8 binaries should include the following additional dlls?
w8hook.dll
schook.dll
(and their 64-bit equivalents)

And more generally, are vnchooks.dll and schook.dll required in all cases for correct functioning?
Thx in advance
Tiziano

[Rudi De Vos]

All the dll's are optional.

w8hook; use the new duplication engine on win8. I's 10x faster then without . I always would install this one.
schook/vnchook: both hint the system about possible screen changes ( using a different method), vnc detect faster the with pure polling
When both are installed only schook is used.

You can use pure polling without any hookdll. It use more cpu but if you only need to connect a few minutes
to verify something it doesn't gonna make a big difference.


Short: You can run without, but the dll's make vnc faster. The speed gain depend on the system
and on the application. You need to try what works the best for you.

[plants]

Awesome. I feel (more) in control now!

Having noticed that schook does not feature as frequently in your releases, are you in favour of vnchooks?

I ask because from a debugging point of view, would it not be better if we all tested using exact same set-up, or does it not make any difference to core function (i.e.: cannot influence crashing or screen update issues)

FYI: previous to this latest release (1193), the schook(64).dlls I had dated back to 10/09/2012 (the last version I have used was 1191) - I always download all your releases (zipped binaries, setup exes and now the msi's) in the hope I get all the latest code, as I understand it must be complicated to keep a handle all all these releases, especially at the frequency you're producing!

Again, thanks in advance, and apologies if you'd read my previous version of this post; I've obtained the latest schook.dlls from the Addons packages, both dated 14/08/2013 @ 15:27

[Rudi De Vos]

vnchook: gpl, opensource this use pure hooking
schook:freeware this use the accessiblity engine

Both files didn't changed for years
last source changes (vnchook 2007/schook 2012)
http://sourceforge.net/p/ultravnc/code/ ... /vnchooks/

The different dates you see are caused by the signing or a rebuild of the same source.

[plants]

Got it. Thanks Rudi. Much appreciated, as always!
Feedback on 1193 to follow...

[Prisma]

I've obtained the latest schook.dlls from the Addons packages, both dated 14/08/2013 @ 15:27

Be aware: The addon setup installs dlls dependent from you OS. Not only differencing between 32- and 64-bit. 2000, XP-7 and Win8 also. If you plan to pick out the schook.dll for repackaging, be shure to have the universal one (build for w2k and above). A fault I made myself.

[plants]

I've obtained the latest schook.dlls from the Addons packages, both dated 14/08/2013 @ 15:27

Be aware: The addon setup installs dlls dependent from you OS. Not only differencing between 32- and 64-bit. 2000, XP-7 and Win8 also. If you plan to pick out the schook.dll for repackaging, be shure to have the universal one (build for w2k and above). A fault I made myself.

Thanks for the reminder. I always repackage to simplify / customise deployment, so I'm well aware of that behaviour (it's a bit annoying, frankly speaking!)
T

[Octopuss]

Is it possible to make the msi installer equal to the Inno version, options-wise? I just noticed the msi one lacks the "server silent" option, which is very important, because the configuration popup is really annoying, and goes in the way of unattended installation.

[plants]

With regard to 1.1.9.3 - it's by far the most stable version of the recent builds. No unusual behaviour, no crashes or lock-ups for both LAN and WAN. The viewer switches to the u2 encoding more often than not, which is useful under most conditions

However, I have similar / identical problems to report regarding stability under the reverse connection scenario, whereby if either the initiator or listener computers have a weak or inconsistent internet flow, the listening vnc viewer regularly produces CPU-spikes, the process becomes deadlocked, and crashes. It's not unusual for those three events to follow in sequence, but they don't always occur in that order (not reproducible mainly because of the unpredictability of internet fluctuation)

I'll provide more detail / conclusions following some other live sessions

[Futurepower(R)]

We have been unable to get any version of UltraVNC to work properly since v1.0.9.6.2.

Is the 1.1.9.3 (test) viewer not compatible with the old server?

Problems with the web site:
v1.1.9.3 is not listed as "TEST" on the download page. It is apparently a test version.
v1.1.9.3 is listed as v1.1.9.0 on the download page.

[Rudi De Vos]

1.1.9.3 is no test version.

Website updated as it seems confusing.
1.1.9.0 replaced by 1.1.9.X
download page now show 1.1.9.X with title 1.1.9.3
Possible people now gonna ask for the X version


Mixing different server/viewer versions.
We try to keep them compatible, but sometimes we forgot something a few years ago
to make to server behave 100% with a new viewer.

Code: Select all

We have been unable to get any version of UltraVNC to work properly

Sorry, but his doesn't give any info to help. What isn't working is it a viewer/server issue
does it happen with special encoders etc... To fix something we need to be able to repeat it.

[Futurepower(R)]

Rudi,

Thanks for your quick reply!

You said not to mix server and viewer versions. That was our problem!

UltraVNC 1.0.8.2 SERVER can be accessed using any of the newer VIEWER versions up to 1.0.9.6.2. UltraVNC 1.0.8.2 VIEWER can access any SERVER version up to 1.0.9.6.2.

Later VIEWERS, after 1.0.9.6.2, cannot access earlier SERVERS. Later SERVERS cannot be accessed using earlier VIEWERS.

Writing help? Rudi, Do you want some help with writing your web site? English is my native language. Also, I have plenty of experience with professional writing.

See below for some proposed changes:

_____________________________
_____________________________

NOW -- The web site says now:
_____________________________

Page title: UltraVNC 1.1.9.3
Page heading: UltraVNC 1.1.9.3
Product version: Product version: 1.1.8 Submission date: 11/26/2012
Reuploaded as 1.1.9.3 on 02/09/2013
Download link: setup win32 119X
Download page: setup win32 119X 1.1.9.0
Downloaded file name: UltraVNC_1_1_9_X86_Setup.exe

_____________________________
_____________________________

SHOULD BE -- The web site should say:
_____________________________

Page title: UltraVNC 1.1.9.3
Page heading: UltraVNC 1.1.9.3
Product version: Product version: 1.1.9.3 Submission date: 02/09/2013
Reuploaded as 1.1.9.3 on 02/09/2013
Download link: setup win32 1.1.9.3
Download page: setup win32 1.1.9.3
Downloaded file name: UltraVNC_1.1.9.3_X86_Setup.exe

[lwc]

Each individual download link (source or binary) still displays "119X 1.1.9.0 License: GNU/GPL Size: ...".
And overall the download page is called ...download-ultravnc-1190.html
Please fix this.

[plants]

Hi Rudi,
I've had several scenarios that all behave consistently and I have a suggestion for debugging. First, the evidence:

Variables and Parameters:
The server initiates a reverse connection to a single listener. Autoreconnect is off. No encryption used. Firewall exception set for server executable. Using 1.1.9.3 x86 and x64 and Win8 variant wherever applicable
Viewer uses the following command line parameters:
/listen 5511 /quickoption 10 /autoscaling /reconnectcounter 0 /autoreconnect 0 /autoacceptnodsm /autoacceptincoming /notoolbar /disablesponsor
Viewer tested on Windows XP and Windows 7 and reports as being version 1.1.9.1 (although it was obtained via the 1.1.9.3 package)

Cases:
1) Server [Poor internet quality (from ISP), LAN connection, WinXP x86] -> Viewer [LAN, good internet, XP&7 x86]
- Frequent hangs, but no crashes. Screen is often covered in grey blocks and loss of control regularly occurs for up to 10 seconds or so. Following loss of control viewer either resumes connection automatically or hangs indefinitely, in the latter case with no CPU usage and nominal memory usage (=deadlocked). However, in all cases, viewer can be terminated normally and the server detects this break in connection and terminates normally

2) Two servers simultaneously [Poor Wifi, Win8 x64] + [LAN, good internet, Win7 x86] -> Viewer [Good Wifi, Good internet, XP&7 x86]
- Initially a stable connection with both servers, maybe lasting 30 seconds. Viewer then hangs indefinitely, no significant CPU or mem load. Viewer terminated normally, status of servers unknown, although presumed exited normally because both attempt to connect subsequently. However, from this point onwards, whenever the servers connect together (two viewer instances appear in the taskbar) the viewer crashes immediately. This situation persisted, until I asked the second client (LAN, Win7x86) to end their session

3) Server [Poor Wifi, Win8 x64 (as above, but different occasion)] -> Viewer [Good Wifi, Good internet, XP&7 x86]
- The viewer often hangs but the server rarely / never detects that the connection has dropped (once the viewer is forcibly terminated). The server must also be forcibly terminated as it appears to be under the impression that all is well...
- Regardless of good or bad connection, the mouse pointer is never drawn (never more than a small square / dot) - as though the viewer has been instructed to not show the remote cursor. When inspecting the viewer connection options during such a session, it is set to 'Track Remote Cursor locally' (the default)

I have a suggestion to emulate the conditions of poor / inconsistent internet. On most laptops, the wifi adapter's properties page should have a setting (or more) for power usage. If you can set this to a maximise battery life / use minimum power you will recreate a similar scenario, but the device has to be quite far away and it must be a reverse connection!

Tiziano

[Rudi De Vos]

Need to make time for this...

There existed some virtual image that act as a router but with the option to
add some internet lag/ dropped packets etc...but kan find it anymore.
Perhaps someone know the name of this packet.

[Rudi De Vos]

Found some app that seems to do the trick
NEWT ( Network Emulator Windows Toolkit), selecting GPRS stop the viewer, but's it is actual the server that close the connection.
Testing...

[Rudi De Vos]

First test results:

On connection with a slow network ( GPRS) the viewer stay black and after a while the cursor change in a dot, even if you
hover over the buttons.

Cause:
Om slow connections, zrle is used by the auto settings.

The zrle encoder lock the screenbuffer while updating, but unlike other encoders it download data in the encoder.
As a result the screenbuffer stay locked for minutes. But in some other part of the code the screenbuffer is used
to show mouse updates. In normal cases you only have a update lock for a few ms, but now the lock is minutes
and the full viewer start locking all his threads as all are waiting for the unlock. Viewer become unresponsive and die.

The difference with the other encoders is that they only lockscreen before changing the screen buffer en unlock after.
The download part isn't in the lock.

Solution:
Zrle lock mechanism changed to exclude the network part.
For the initial screen, we now show each individual rect, instead of a black screen for minutes you now see the screen being build.
It take the same time, but you have a feeling it's faster because you see what is happening.

Still testing, the network tool realy is a nice addon for testing uvnc.
It also has other settings that even can simulate network losses/disconnects/packets reorders etc...
I hope to upload some test viewer this weekend

[Rudi De Vos]

test viewer
http://www.uvnc.eu/1194/vncviewer_test_lowband.zip

[plants]

May I be the first to express my excitement about these new developments!! Woo hoo!
Erm... where was I?

I'm back from (very fricken windy) holiday now and I have a remote session in the morning... I will use the test viewer
I presume from what you've said, that I allow it to choose its own encoding? The client will be connected wirelessly and I would say their signal strength is typically 50-75%

Many thanks Rudi, much appreciated
T

[plants]

As posted, I've used the test version viewer with excellent results on XP and Win7 x86. Stability retained over a period of over six hours
The only major anomaly to report was a file-transfer hang that resulted in a viewer crash. The server detected a loss of connection and resumed normally on reconnect
Partial file-transfer resumed without problems.
Many many thanks! I guess I'll get on with 1194... seeing as you appear to be blitzing the bugs
T

[mommafrye]

What version of GPL is Ultra VNC licensed under (i.e.; Version 2.1, 3.0, etc.)? Please advise.