Patchwork Making hg 5.0 as beta release with python 3 support.

login
register
mail settings
Submitter Matt Harbison
Date March 5, 2019, 4:04 a.m.
Message ID <op.zx5uw6c39lwrgf@envy>
Download mbox | patch
Permalink /patch/39058/
State New
Headers show

Comments

Matt Harbison - March 5, 2019, 4:04 a.m.
On Mon, 04 Mar 2019 11:25:14 -0500, Pulkit Goyal <7895pulkit@gmail.com>  
wrote:

> Hey everyone,
>
> I hope everything is going well.
>
> After years of work on porting mercurial to Python 3 by everyone, we are
> close, very close. Right now, only 5 tests fail on python 3 and there  
> are 4
> regressions (tests which were passing earlier and started failing
> recently). Few days ago, I installed mercurial using Python 3 as default
> mercurial on my personal system. Things are working good. Things which I
> have noticed not working are:

This is great news!  As an additional data point, Windows is lagging  
behind with 41 failures (604 pass).  Some aren't important (test-gendoc*  
complaining titles are too short), but some are (some log commands fail  
because of bytes vs str in nturl2path.pathname2url()[1]).  A couple new  
failures seemed to have crept into the list in the last ~150 commits, and  
I think pager tests might be flaky.  Some of the protocol testing seems to  
be picking up an extra trailing '\r':

    write limit reached; closing socket
    $LOCALIP - - [$ERRDATE$] Exception happened during processing request  
'/?cmd=getbundle': (glob)
    Traceback (most recent call last):

If there's interest, I should be able to set up a py3 buildbot on a much  
faster Windows 10 system.  The total output is ~130K, so I don't want to  
just post that.

[1]  
https://www.mercurial-scm.org/pipermail/mercurial-devel/2018-October/123923.html
Pulkit Goyal - March 5, 2019, 1:25 p.m.
On Tue, Mar 5, 2019 at 7:04 AM Matt Harbison <mharbison72@gmail.com> wrote:

> On Mon, 04 Mar 2019 11:25:14 -0500, Pulkit Goyal <7895pulkit@gmail.com>
> wrote:
>
> > Hey everyone,
> >
> > I hope everything is going well.
> >
> > After years of work on porting mercurial to Python 3 by everyone, we are
> > close, very close. Right now, only 5 tests fail on python 3 and there
> > are 4
> > regressions (tests which were passing earlier and started failing
> > recently). Few days ago, I installed mercurial using Python 3 as default
> > mercurial on my personal system. Things are working good. Things which I
> > have noticed not working are:
>
> This is great news!  As an additional data point, Windows is lagging
> behind with 41 failures (604 pass).  Some aren't important (test-gendoc*
> complaining titles are too short), but some are (some log commands fail
> because of bytes vs str in nturl2path.pathname2url()[1]).  A couple new
> failures seemed to have crept into the list in the last ~150 commits, and
> I think pager tests might be flaky.  Some of the protocol testing seems
> to
> be picking up an extra trailing '\r':
>
> --- c:/Users/Matt/hg/tests/test-http-bad-server.t
> +++ c:/Users/Matt/hg/tests/test-http-bad-server.t.err
> @@ -995,8 +995,8 @@
>     sendall(9 from 9) -> (77) 4\\r\\n\x00\x00\x00\x00\\r\\n (esc)
>     sendall(9 from 9) -> (68) 4\\r\\n\x00\x00\x00)\\r\\n (esc)
>     sendall(47 from 47) -> (21)
> 29\\r\\n\x0bCHANGEGROUP\x00\x00\x00\x00\x01\x01\x07\x02
> \x01version02nbchanges1\\r\\n
> (esc)
> -  sendall(9 from 9) -> (12) 4\\r\\n\x00\x00\x01\xd2\\r\\n (esc)
> -  sendall(12 from 473) -> (0) 1d2\\r\\n\x00\x00\x00\xb2\x96\xee\x1d (esc)
> +  sendall(9 from 9) -> (12) 4\\r\\n\x00\x00\x01\xd2\\r\\n\r (esc)
> +  sendall(12 from 473) -> (0) 1d2\\r\\n\x00\x00\x00\xb2\x96\xee\x1d\r
> (esc)
>     write limit reached; closing socket
>     $LOCALIP - - [$ERRDATE$] Exception happened during processing request
> '/?cmd=getbundle': (glob)
>     Traceback (most recent call last):
>
> If there's interest, I should be able to set up a py3 buildbot on a much
> faster Windows 10 system.  The total output is ~130K, so I don't want to
> just post that.
>

There is definitely interest in having a py3 buildbot on windows. This will
help people who don't have windows to look into failures and try to fix
them or atleast reason why they are failing. Also it will help to catch
regressions.
It will be great if before the release, we can improve the status on
windows and maybe make it as same for linux.

>
> [1]
>
> https://www.mercurial-scm.org/pipermail/mercurial-devel/2018-October/123923.html
>
Matt Harbison - March 5, 2019, 2:26 p.m.
> On Mar 5, 2019, at 8:25 AM, Pulkit Goyal <7895pulkit@gmail.com> wrote:
> 
> 
> 
>> On Tue, Mar 5, 2019 at 7:04 AM Matt Harbison <mharbison72@gmail.com> wrote:
>> On Mon, 04 Mar 2019 11:25:14 -0500, Pulkit Goyal <7895pulkit@gmail.com>  
>> wrote:
>> 
>> > Hey everyone,
>> >
>> > I hope everything is going well.
>> >
>> > After years of work on porting mercurial to Python 3 by everyone, we are
>> > close, very close. Right now, only 5 tests fail on python 3 and there  
>> > are 4
>> > regressions (tests which were passing earlier and started failing
>> > recently). Few days ago, I installed mercurial using Python 3 as default
>> > mercurial on my personal system. Things are working good. Things which I
>> > have noticed not working are:
>> 
>> This is great news!  As an additional data point, Windows is lagging  
>> behind with 41 failures (604 pass). 

>> If there's interest, I should be able to set up a py3 buildbot on a much  
>> faster Windows 10 system.  The total output is ~130K, so I don't want to  
>> just post that.
> 
> There is definitely interest in having a py3 buildbot on windows. This will help people who don't have windows to look into failures and try to fix them or atleast reason why they are failing. Also it will help to catch regressions.
> It will be great if before the release, we can improve the status on windows and maybe make it as same for linux.
>> 
>> [1]  
>> https://www.mercurial-scm.org/pipermail/mercurial-devel/2018-October/123923.html

Ok. Augie set me up with the account info for the other 2 bots.

What I’d like to do here is figure out how to run it as a service, instead of having to manually fire up an MSYS shell, change directories, and then spin up buildbot.  I haven’t figured out how to do that.  WSL might make it a little easier, but I don’t have that nailed down yet either.

Patch

--- c:/Users/Matt/hg/tests/test-http-bad-server.t
+++ c:/Users/Matt/hg/tests/test-http-bad-server.t.err
@@ -995,8 +995,8 @@ 
    sendall(9 from 9) -> (77) 4\\r\\n\x00\x00\x00\x00\\r\\n (esc)
    sendall(9 from 9) -> (68) 4\\r\\n\x00\x00\x00)\\r\\n (esc)
    sendall(47 from 47) -> (21)  
29\\r\\n\x0bCHANGEGROUP\x00\x00\x00\x00\x01\x01\x07\x02	\x01version02nbchanges1\\r\\n  
(esc)
-  sendall(9 from 9) -> (12) 4\\r\\n\x00\x00\x01\xd2\\r\\n (esc)
-  sendall(12 from 473) -> (0) 1d2\\r\\n\x00\x00\x00\xb2\x96\xee\x1d (esc)
+  sendall(9 from 9) -> (12) 4\\r\\n\x00\x00\x01\xd2\\r\\n\r (esc)
+  sendall(12 from 473) -> (0) 1d2\\r\\n\x00\x00\x00\xb2\x96\xee\x1d\r  
(esc)