A received message hangs bitmessage-qt, possibly an insecurity? #900

New Issue

2016-10-20T18:37:22+02:00

kirkawolff commented

2016-10-20 18:37:22 +02:00

(Migrated from github.com)

I received the attached message, and whenever I would attempt to select it from the qt interface, the program would hang and consume 100% CPU. The only way I could recover is to kill the program, then run bitmessagemain.py with the -c option, and delete the message. It appears the contents of the program are a DOS program. Nevertheless, this could be an inseurity in bitmessage and should be looked at. I can assist in a debug session if necessary.
Re: rapidgator premium available at a discount.txt

I received the attached message, and whenever I would attempt to select it from the qt interface, the program would hang and consume 100% CPU. The only way I could recover is to kill the program, then run bitmessagemain.py with the -c option, and delete the message. It appears the contents of the program are a DOS program. Nevertheless, this could be an inseurity in bitmessage and should be looked at. I can assist in a debug session if necessary. [Re: rapidgator premium available at a discount.txt](https://github.com/Bitmessage/PyBitmessage/files/542261/Re.rapidgator.premium.available.at.a.discount.txt)

kirkawolff commented

2016-10-20 18:50:26 +02:00

(Migrated from github.com)

There is a message floating around general2 noting this problem:

Subject: Re: to new BM users: if your pyBM keeps freezing up ... remove the .EXE file

99% of bitmessage users won't know how to do this.

sadly, some dumbass posted an EXE-file on BM recently, which screws up many clients (Linux etc.).

you have to open the message.dat file with the SQLiteBrowser , go to inbox table,

and remove (open and CLEAR) the .MZ..... bullshit content.

also compact the whole database with SQLite.

else BM will freeze while navigating through the messages.

In 27 days this poisoned BM will be dropped. (so if you can still read this BM on beamstat, you're not safe yet...)

There is a message floating around general2 noting this problem: > Subject: Re: to new BM users: if your pyBM keeps freezing up ... remove the .EXE file > > 99% of bitmessage users won't know how to do this. > > --- > > sadly, some dumbass posted an EXE-file on BM recently, which screws up many clients (Linux etc.). > > you have to open the message.dat file with the SQLiteBrowser , go to inbox table, > > and remove (open and CLEAR) the .MZ..... bullshit content. > > also compact the whole database with SQLite. > > else BM will freeze while navigating through the messages. > > In 27 days this poisoned BM will be dropped. (so if you can still read this BM on beamstat, you're not safe yet...)

PeterSurda commented

2016-10-20 20:04:04 +02:00

(Migrated from github.com)

Thank you for the report, as you noticed yourself, this already has been reported, also what it triggers and a possible workaround. I'm working at a proper fix right now.

PeterSurda commented

2016-10-20 20:30:16 +02:00

(Migrated from github.com)

Can you test with 448ceaa74c (or later) if it still happens?

Can you test with 448ceaa74ca6929f626fa36962c4f3c0ca0f0659 (or later) if it still happens?

kirkawolff commented

2016-10-21 02:09:19 +02:00

(Migrated from github.com)

it has no apparent effect. problem still present.

PeterSurda commented

2016-10-21 15:59:26 +02:00

(Migrated from github.com)

It looks like there were two separate issues, one with the renderer on windows, and one with the regexp parser. The first commit fixes the first one, the second one the second one. I can't reproduce the problem anymore, can you retry?

kirkawolff commented

2016-10-21 17:36:42 +02:00

(Migrated from github.com)

It only locks up for one second. I am able to delete the message from the UI now.

PeterSurda commented

2016-10-21 19:51:14 +02:00

(Migrated from github.com)

Excellent, I just have to make another commit for Windows compatibility once I finish debugging it, and then I can finally close the issue.

PeterSurda commented

2016-10-22 01:56:51 +02:00

(Migrated from github.com)

Works for me now both on linux and windows, I'm closing the ticket.