So I think I have a small subset of this patch ready that encompassed
everything from the original path but loader.py and trace related files.
But I just realized a problem with updates setup.py. It now depends on
distro package - https://pypi.org/project/distro/ - which requires pip to
install
Indeed, in hindsight, I was as good as vanished. I will make sure to be
more active on the mailing list from now on! Thank you for the
clarifications on the communication channels.
Τη Τρίτη, 18 Φεβρουαρίου 2020 - 6:40:33 π.μ. UTC+2, ο χρήστης Waldek
Kozaczuk έγραψε:
>
> Hi,
>
> Good to hear fro
Thank you! Examining the patch and the relevant discussion reveals some
useful paths in the code (especially around the implementation of mmap).
Also, it's a good example of code interacting with the MMU, which will be
necessary for DAX also, but indeed this optimization in ROFS's page cache
is
Let me split this patch into some smaller ones and take your approach of
first checking if anything breaks with python3.
I think that loader.py needs to be treated separately. Is it true that gdb
comes in with its own python runtime and these days it should be the
version 3 (we are not relying
On Tue, Feb 18, 2020 at 3:54 PM Waldek Kozaczuk
wrote:
> This -
> https://stackoverflow.com/questions/27476079/why-does-2to3-change-mydict-keys-to-listmydict-keys
> -
> might explain this "list" related changes by 2to3.
>
Yet it doesn't explain the change - it explains why it is *not* actually
On Tuesday, February 18, 2020 at 10:12:56 AM UTC-5, Nadav Har'El wrote:
>
>
> On Mon, Feb 17, 2020 at 11:58 PM Waldemar Kozaczuk > wrote:
>
>> This bug was discovered when running httpserver unit tests with python
>> scripts upgraded to version 3. The new version of the requests module
>> chunk
From: Waldemar Kozaczuk
Committer: Nadav Har'El
Branch: master
tests: improve regex rules to catch failures in tst-dlfcn
Some of the asserts in tst-dlfcn.cc were failing but the regex
rules in testing.py would not catch them. This patch fixes
the rules and also temporarily disables the portion
Committed. Thanks, and good riddance :-)
However, one unfortunate issue which is caused by the fact you put the
files in exactly the same place as the previous makefile downloaded them is
that it that it may confuse git who may refuse people from checking out the
new version on a non-clean working
From: Waldemar Kozaczuk
Committer: Nadav Har'El
Branch: master
tests: add bionic tests to the source tree
Instead of cloning full bionic source repo (~70MB) everytime we need to
build tests image, this patch simply adds seven small source files
to the OSv repo. The files are copies from the di
On Mon, Feb 17, 2020 at 11:58 PM Waldemar Kozaczuk
wrote:
> This bug was discovered when running httpserver unit tests with python
> scripts upgraded to version 3. The new version of the requests module
> chunks
> POST requests with body so that they are sent over socket in two parts -
> the requ
This -
https://stackoverflow.com/questions/27476079/why-does-2to3-change-mydict-keys-to-listmydict-keys
-
might explain this "list" related changes by 2to3.
On Tuesday, February 18, 2020 at 8:42:47 AM UTC-5, Waldek Kozaczuk wrote:
>
> Hey,
>
> Thanks for the review.
>
> Let me first explain wh
Hey,
Thanks for the review.
Let me first explain what my process was. At first, I just tried to submit
Matt's changes with proper signoff. But then I decided to test it a bit and
discovered some things were breaking. Now I do not think it was because of
any problems in his original script, but
Thanks. I commented with a few concerns below. I'm only really worried
about loader.py - the gdb script, which was already supposed to be working
for both Python 2 and Python 3 (although we probably didn't test it
recently), and I'm worried these changes are breaking it, rather than
improving it -
13 matches
Mail list logo