Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 16:23
    kdcis synchronize #1864
  • 13:41
    shekhar-cis opened #1871
  • 08:24
    shekhar-cis synchronize #1866
  • 06:46
    PeterSurda closed #1860
  • 06:46
    PeterSurda closed #1867
  • 06:46

    PeterSurda on v0.6

    removed assertTrue and added fo… Update Myaddress test removed create adddress test and 2 more (compare)

  • Oct 22 13:20
    shekhar-cis review_requested #1861
  • Oct 22 12:19
    shekhar-cis opened #1870
  • Oct 22 12:10
    shekhar-cis review_requested #1866
  • Oct 22 12:10
    shekhar-cis synchronize #1866
  • Oct 22 11:40
    shekhar-cis synchronize #1861
  • Oct 22 11:34
    shekhar-cis opened #1869
  • Oct 22 10:56
    shekhar-cis review_requested #1863
  • Oct 22 10:23
    cis-muzahid synchronize #1794
  • Oct 22 09:57
    shekhar-cis closed #1868
  • Oct 22 09:55
    shekhar-cis opened #1868
  • Oct 22 09:55
    shekhar-cis synchronize #1863
  • Oct 22 09:28
    shekhar-cis review_requested #1860
  • Oct 22 09:20
    shekhar-cis synchronize #1867
  • Oct 22 09:15
    shekhar-cis opened #1867
Peter Šurda
@PeterSurda
i.e. it skips more tests
on python2
Dmitri Bogomolov
@g1itch
A bit late... the relavant travis doc about _with_system_site_packages is https://docs.travis-ci.com/user/languages/python/#travis-ci-uses-isolated-virtualenvs
I'm reading the unittest source to find where does they make second pass
Dmitri Bogomolov
@g1itch
I run bitmessage from my testing branch which includes qt5-wip and lately found an issue with tray icon: it becomes a lamp or disappears than new message is received ):
Peter Šurda
@PeterSurda
@g1itch I assinged one of the junior developers to do exhaustive testing of #1389 and document it, while I do the code review. Since we don't have automated testing for QT yet
Dmitri Bogomolov
@g1itch
OK, lets count it as beta testing. I'm constantly promoting my qt5-wip branch, counting it an alpha testing (:
Dmitri Bogomolov
@g1itch
I'm a fool ): the second pass is from pybitmessage symlink
John Doe
@FucktheUSA_gitlab

@g1itch "I ... don't know what is BitMessage. This chat is for Bitmessage project, not BitMessage."

I don't know where is or what is Ukraine. But I do know every 3rd person speak Russian. I think "Ukraine" is part of Russia.

You are complaining that someone called "Bitmessage" BitMessage? It is absurd. Why you who has nothing to do with BitMessage complaining in the first place. Did you invented BitMessage? Is the license forbids people to call Bitmessage as BitMessage?
The License is a MIT license which even grant people all right to SUBLICENSE BitMessage:
"without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software"
Sublicensing mean they can change the license to anything they like for example to GPLv3+
Ranting about that someone changed one character in Bitmessage is absurd.
John Doe
@FucktheUSA_gitlab
I'm a communist and I'm all for Russia, I hope they beat the shit out of a few Ukrainian neo-nazis which are financed by the U.S. Department of State.. You know exactly what "battalion" I talk about.. So better stop ranting about if someone calls Bitmessage "BitMessage". Even AARON SWARTZ called Bitcoin as "BitCoin" ..
@PeterSurda @g1itch Travis is totally unreliable, more problems it creates than solves and returns a bunch of false positives...
John Doe
@FucktheUSA_gitlab
How dare the person who wrote these instructions calling Bitmessage "BitMessage" how dare you... how dare is Russia to defend itself from NATO terrorists and Ukrainian neo-nazis playing war games at the Russian borders... You being an asshole, I can be asshole too my dear Russian comrade. How dare you..
Peter Šurda
@PeterSurda
@FucktheUSA_gitlab Please keep the conversation on topic. This isn't an appropriate place to discuss politics.
While I am puzzled why a self-proclaimed communist is posting about Bitcoin and Bitmessage, it isn't relevant in this chan.
Peter Šurda
@PeterSurda
Regarding naming, g1itch was right to point out the capitalisation issue, I think that if people can't spell, they just expose their own unprofesionalism. This is how we got professor Bitcorn. But I tend to be more focused on my own professionalism, not other peoples'.
I don't think it's strange for a software developer to complain about single character.
Peter Šurda
@PeterSurda
Regarding travis being unreliable, I'm not sure if you're referring to travis-ci's web service, or PyBitmessage tests in general. I have had technical problems with travis-ci's website before but they were resolved after raising tickets. I'm more worried about being able to run many jobs for each commit (not enough parallel jobs even for expensive plans) and other minor things. Regarding tests in general not being reliable/deterministic, we're aware of it, it will probably take a while to fix these issues, it needs a lot of refactoring.
I'm studying the concepts of continuous delivery and refactoring so that it can be applied to the development process, and am trying to guide other developers to follow these practices too.
John Doe
@FucktheUSA_gitlab
I referring to the website
I would rather test PyBitmessage with different versions of python by myself than use any automated thing...
Also, I don't think that PyBitmessage need to support any version of python below version 3.5
Python 3.9.1+ (default, Feb 5 2021, 13:46:56)
[GCC 10.2.1 20210110] on linux
John Doe
@FucktheUSA_gitlab
In the other hand as I stated Aaron Swartz who was a well known programmer killed by the U.S. government called Bitcoin BitCoin. I don't think is problematic to change Bitmessage to BitMessage
John Doe
@FucktheUSA_gitlab

Can we replace ECDSA with EdDSA? I think the use of ECDSA in the case of Bitmessage would pose a serious threat, sending the same message two times actually will allow the recovery of the private key of the sender.

I'm not sure if Bitmessage signs the message with if does than calculation of the private key is certainly possible.

right there, Bitmessage signs the message with ECDSA without adding any timestamp to the message whatsoever
so if you resend the same message two times the recipient can calculate your private key in seconds
John Doe
@FucktheUSA_gitlab
meh,
signing the same message should produce the same signature
so it is secure.
except if ECDSA weak randomness is happening, if the RNG on the client fails than the calculation of the private key is 100% possible
What is the source for RNG used by Bitmessage? Is some python library? If the random number generator fails at encryption from the signatures the recipient can recover the private key of the sender. This is a fundamental issue with ECDSA
John Doe
@FucktheUSA_gitlab
replacing ECDSA with EdDSA (ed25519) would be beneficial for Bitmessage (more secure as well)
John Doe
@FucktheUSA_gitlab
Did someone noticed that PyElliptic is DEPRECATED? https://github.com/yann2192/pyelliptic
John Doe
@FucktheUSA_gitlab
@g1itch aparantly aware of it based on his pull requests from yesterday
Peter Šurda
@PeterSurda
@FucktheUSA_gitlab you can run the tests now locally using './run-tests-in-docker.sh'. It run the travis tests.
Peter Šurda
@PeterSurda
I'm not sure how sending the same message twice will allow key recovery. If I see correctly, the signature is the same. Encryption uses a hybrid scheme (ECIES), which encrypts a (random) AES key, so the ciphertext is different even if plaintext isn't.
pyelliptic development continues inside PyBitmessage, we're aware upstream was abandoned, we've been patching it to keep up with changing environments.
randomness for keys is retrieved from operating system, if that fails (e.g. OS doesn't support it), from openssl.
@g1itch buildbot workers will be off for a while, maybe until Monday, I had to change the internet provider and there are some problems they have to solve, until then the workers can't be accessed.
John Doe
@FucktheUSA_gitlab
Thank you @PeterSurda you are a great man!
Peter Šurda
@PeterSurda
@g1itch I managed to put together some sort of tunnel to allow communication between the buildbot master and the workers, so the jobs can now run again. However I'm not sure the tunnel is persistent, I'll have to monitor it over weekend.
hemengtingyaodi
@hemengtingyaodi
yes
i
am a great man