mirror of
https://github.com/quasar/Quasar.git
synced 2026-04-25 15:25:59 +03:00
[GH-ISSUE #209] Release Candidate for RELEASE4 #88
Labels
No labels
bug
bug
cant-reproduce
discussion
duplicate
easy
enhancement
help wanted
improvement
invalid
need more info
pull-request
question
wont-add
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
starred/Quasar#88
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @MaxXor on GitHub (May 26, 2015).
Original GitHub issue: https://github.com/quasar/Quasar/issues/209
Please download the artifcats from the latest AppVeyor builds in Release mode and help testing all features for the next RELEASE4. Please report all bugs you find.
I think the current status of the
master-branch is stable, some minor bugs only I hope.@ghost commented on GitHub (May 26, 2015):
Other than the little bugs that were recently fixed, Reverse Proxy in context menu has no icon! XD
I'll continue to try and break things, but everything seems stable!
@MaxXor commented on GitHub (May 26, 2015):
Gonna add one. :D
@yankejustin commented on GitHub (May 26, 2015):
We need to also test these features for multiple clients. I doubt most of us are testing many of these features using more than one client.
@yankejustin commented on GitHub (May 26, 2015):
@MaxXor :)
Moved to: Issue 216
@rabbitsmith commented on GitHub (May 27, 2015):
afaik xrat is stable but in win10 the remote desktop does not start sometimes when start button is clicked. (at least for me). probability of not starting is 20 of 100 . and file manager download isnt working for me
@ghost commented on GitHub (May 27, 2015):
@rabbitsmith, i'll be doing some testing with win10. What is your build of win10 is you don't mind me asking?
@rabbitsmith commented on GitHub (May 27, 2015):
9879 64bit
@ghost commented on GitHub (May 27, 2015):
@rabbitsmith, Thank you. I'm using the Windows 10 Pro Insider Preview (Build 10074 32bit) and everything seems to be working fine for me. If I recall correctly, you had mentioned your uplink was slow. Perhaps this could be a factor?
@ghost commented on GitHub (May 27, 2015):
Do you have any other OS's you could test this on?
@rabbitsmith commented on GitHub (May 27, 2015):
my isp bans uploading of files. i have a data plan only on my phone. i test xrat only on lan with two pcs running win10 9879
@rabbitsmith commented on GitHub (May 27, 2015):
the file manager transfer is really fast but on my pc the integrity of files isnt assured. i transferred a 7mb file with no errors and hashes matching. the problem still persists for files larger than 15 mb.
i mentioned my uplink was slow because i couldnt upload the xrat binaries and the screenshot.
@ghost commented on GitHub (May 27, 2015):
Thank you for the information. I will try and emulate this problem you are having.
@yankejustin commented on GitHub (May 27, 2015):
@MaxXor I apologize if this is a foolish question, but would it be possible to add me as a contributor?
@MaxXor commented on GitHub (May 27, 2015):
@yankejustin You are doing a great job here for the project, but I would add only reluctantly somebody. I hope you understand this. :)
@rabbitsmith You have a pretty scary ISP if it bans uploading of files. :o Is this restricted to all kinds of files? You should upgrade your Windows 10 to the latest preview version, maybe there's the problem.
@yankejustin commented on GitHub (May 27, 2015):
@MaxXor Yes I understand. Unfortunately adding contributors is a huge jump in power since it is a public repository.
I think we really should fix Remote Shell before we release the next version of xRAT. The input it accepts is still very lenient, such as
exit(simple trimming of the command, then checking). It also doesn't provide error output of the client's shell yet. I almost have it working, though. :)@MaxXor commented on GitHub (May 27, 2015):
Ok. :)
I have just added the trimming.
@MaxXor commented on GitHub (May 29, 2015):
Please do some more tests, the last commits changed a lot of things.
@ghost commented on GitHub (May 29, 2015):
Nice changes! I'm having trouble breaking things now! :'(
@ghost commented on GitHub (May 31, 2015):
There is a noticeably large memory leak in the server.
To reproduce: Launch 50-100 clients. Select all -> Send reconnect command -> watch xRAT.exe in taskmgr -> repeat sending reconnect commands when clients reconnect
@MaxXor commented on GitHub (May 31, 2015):
Updated the client & server a bit again, added some more locks & fixed the documentation. The Server.cs documentation is complete, the Client.cs 50% each.
Gonna work on this memory leak now.
@MaxXor commented on GitHub (Jun 2, 2015):
Does the memory leak fix work reliable for you?
@ghost commented on GitHub (Jun 2, 2015):
It is working fine for me. Memory is being reallocated, and I haven't noticed too much of a performance issue.
@yankejustin commented on GitHub (Jun 2, 2015):
Would you be able to provide a rough estimate of the cpu cost? It is nice to know it isn't as big of a hit as I thought. :)
Not quite sure how many times the garbage collector would try to compress the LOH...
@yankejustin commented on GitHub (Jun 2, 2015):
@MaxXor Also, thanks for the "green apple"! x)
@ghost commented on GitHub (Jun 2, 2015):
Surprisingly, the only significant CPU usage that occurs is when the server is allocating the resources for each client. My i7 2600k jumps from 0% idle to ~25-35% when the clients are connecting. When I sent a command to reconnect 100 clients, my CPU usage jumped up to ~7%
@yankejustin commented on GitHub (Jun 2, 2015):
@d3agle I would appreciate if you would test the performance change on my experimental branch to see if what I did helps out with the CPU usage.
@MaxXor commented on GitHub (Jun 8, 2015):
Could anyone help me test the branch
dev-buffer? I think I'm done improving the buffer management.Changes: https://github.com/MaxXor/xRAT/pull/263#issuecomment-109762339
@yankejustin commented on GitHub (Jun 8, 2015):
@MaxXor Think we could add an
#if DEBUGat the builder to allow the creation of multiple clients at one time for easier testing of many clients? It would certainly save some time!Perhaps add a little TextBox for numeric input at initialization of the Builder Form in
DEBUGmode to generate that many clients. :)I will help you test it tonight.
@MaxXor commented on GitHub (Jun 8, 2015):
What I do while testing is replacing the setting
MUTEXwithpublic static string MUTEX = Guid.NewGuid().ToString();So you can just start the client multiple times for tests.@yankejustin commented on GitHub (Jun 8, 2015):
Hmmm. Perhaps I could do a bit of changes to make a new Configuration for testing multiple clients by imitating the
RELEASEsetting. Would you suppose such an addition would be beneficial?@yankejustin commented on GitHub (Jun 8, 2015):
That is a nice point about the
MUTEX. Thank you for the reminder.@yankejustin commented on GitHub (Jun 8, 2015):
@d3agle @MaxXor Would you please help me test my new configuration that I made for testing multiple clients please? The only part that does not work is the client. The client just crashes and I am not quite sure why. After starting all the clients, they just stop working.
Perhaps I set a setting for the client incorrectly...
@yankejustin commented on GitHub (Jun 9, 2015):
Oh wait. I take that back. The client doesn't work for me on the
MaxXor/dev-bufferbranch at all. I'm going to open an issue for it: it likely is not the fault of my code on the new configuration. We shall see...@ghost commented on GitHub (Jun 9, 2015):
@yankejustin, @MaxXor changed the namespace for the networking part of the client/server to
xCore.Client.Networkingbut forgot to add it in theBuild/Renamer.cs.That should fix it
@ghost commented on GitHub (Jun 9, 2015):
Sorry, I meant
xClient.Core.Networking@yankejustin commented on GitHub (Jun 9, 2015):
Oh, good. Thank you! :)
@yankejustin commented on GitHub (Jun 9, 2015):
@d3agle Would you mind sending me a pull request please? I would like to test my new configuration. :)
@yankejustin commented on GitHub (Jun 9, 2015):
I think it would help you out quite a great deal.
@ghost commented on GitHub (Jun 9, 2015):
I'm sorry I don't understand, what would I be sending a pull request for? I'll pull your branch and test it out. :D
@yankejustin commented on GitHub (Jun 9, 2015):
I did notice I forgot to exclude the
Labelfor configurations other thanMULTI_CLIENT... I will fix it in a minute. :)@ghost commented on GitHub (Jun 9, 2015):
How would I go about creating multiple clients? I changed both Build options for Client/Server but I don't see anything where I can select how many Clients to create.
@yankejustin commented on GitHub (Jun 9, 2015):
If you select the "Multi-Client Release" Configuration, re-build the server and run it. You should see it on the bottom-right corner of the Builder Form.
@yankejustin commented on GitHub (Jun 9, 2015):
Make sure you restart Visual Studio because adding the new configuration meant I had to modify the project files.
@yankejustin commented on GitHub (Jun 9, 2015):
@d3agle Sorry to bug you again but I am unable to see the reason why my
yankejustin/NewConfigurationbranch still won't build functional clients... Is there any change you can see that could be the cause of this?@ghost commented on GitHub (Jun 9, 2015):
@yankejustin, No worries, I'm hardly ever bugged. :) I can take a look tonight when I'm free.
@MaxXor commented on GitHub (Jun 9, 2015):
@d3agle Would you like to also test the new client & server from
dev-bufferin Release mode? :)@ghost commented on GitHub (Jun 9, 2015):
@MaxXor, Sure thing. I've already played with it a little bit, and everything seemed good. Memory usage was down dramatically, as well as CPU usage. I believe that there was one instance where only some of the clients connected out of all 100, however I'm unsure if it was a fault on my part with testing or if there is a bug somewhere. I can look into it as well! :)
@MaxXor commented on GitHub (Jun 9, 2015):
@d3agle Yea, that would be great if you can find the cause for it mayb. :)
@yankejustin commented on GitHub (Jun 10, 2015):
@d3agle In my case, I can get no clients to connect correctly. Any that I try to start will crash on this configuration...
More Information
1. Operating System - Windows 7 Home Premium - Service Pack 1 2. System - Processor: Intel(R) Core(TM) i7 CPU - 920 @ 2.67 GHz 2.67 GHz - RAM: 16 GB - System type: 64-bit Operating System 3. Connection - Local Connection to the server -(127.0.0.1)- - Default port -(4282)- 4. Client - Default Settings - No administrator privileges (I have tried, still doesn't work) - Keylogger is not enabledStarting the Client
1. Double-Click the client 2. Nothing happens for (~15)-(~30) seconds, then a message from Windows will pop up stating that the program has stopped working. 3. As one can guess, by then the program can not recover and it will cause the client to crash.Final Notes
1. This is likely an issue with the Settings for the client. I have tried debugging to see what is causing this issue but it will function correctly. 2. There are no problems. When I use theRELEASEconfiguration, the clients built only crash.@ghost commented on GitHub (Jun 10, 2015):
@yankejustin, when the client crashes it should leave a metadata file in
%temp%and it should show what exception is being thrown runtime. Could you try to get the client to crash again? Instead of closing the dialog that pops up, allow it to collect the information and it will display another dialog with a path to the.dmpfile and an.xmlfile (the xml file will contain the info)@yankejustin commented on GitHub (Jun 10, 2015):
@d3agle A client from earlier this morning was used, so I tried the more-current client a bit ago and it works to some extent. My current issue is having the client show up in the GUI using the latest dev branch. While the client does connect successfully (and is stored in
_clients), the client is not shown in theListView. I cannot also do things like show the Statistics (states that there are no connected clients). Again, they are in_clientsandObjectDisposedExceptionsare thrown when I attempt to stop listening on the server for each of the clients, but they are not fully connected. I do know that theObjectDisposedExceptionis thrown when we disconnect the client and call to close the socket (if (_handle != null) _handle.Close(); etc...... Not sure what is preventing the client from callingProcess... It is not getting past here and it must for the server to update the GUI and do further processing...@MaxXor commented on GitHub (Jun 10, 2015):
The
ObjectDisposedExceptionis normal when clients are connected and the server stops listening, because at this line it can't receive anymore from the socket. :)Well, I've tested this now with multiple clients using the RELEASE mode and it's working fine, with enabled/disabled startup and installtion...
Please completely rebuild your solution in RELEASE mode and re-build your client and test it again.
Also try disabling your windows firewall for this. Be sure to use the same password for client & server!
Try this with the latest commit from
dev-buffer.//edit: the problem I'm experiencing now is that clients won't be added to the listview after some reconnects and start/stop listening.. :(
@MaxXor commented on GitHub (Jun 11, 2015):
This issue should be gone now, can you confirm this with the latest commit? :)
//edit: Tested it now on Windows 7, 8.1 and 10. Seems to working! Finally! 💃
@ghost commented on GitHub (Jun 11, 2015):
@MaxXor, good work! I apologize but I won't be able to participate for ~3 weeks. I have a lot of work to do in school, and I need to concentrate on this work instead. Keep up the good work guys! 👍
@MaxXor commented on GitHub (Jun 11, 2015):
@d3agle Thanks. No problem! =)
@yankejustin commented on GitHub (Jun 12, 2015):
Yeah I have been pretty busy these last few days. I will test tonight if I can.