Date   
Re: Alpha will not download, was Re: [nvda-devel] Extra checks forPull Requests

derek riemer
 

oh.

On Sat, Aug 3, 2019 at 1:58 AM Arnold Loubriat <datatriny@...> wrote:

Hi Derek,

Just for reference: no it was not an AppVeyor outage.

See https://github.com/nvaccess/nvda/pull/10020 that got merged by Reef yesterday!

Have a nice day.

 

Arnold

 

De : derek riemer
Envoyé le :samedi 3 août 2019 07:04
À : nvda-devel@groups.io
Objet :Re: Alpha will not download, was Re: [nvda-devel] Extra checks forPull Requests

 

I think appVeyor had an outage or something. it's working again.

 

On Thu, Aug 1, 2019 at 1:34 PM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:

For completeness.

IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'space', EndUtteranceCommand()]
IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'pressed']
IO - speech.speak (20:29:39.388):
Speaking [LangChangeCommand ('en_GB'), 'Downloading Update  dialog
Connecting']
IO - speech.speak (20:29:39.393):
Speaking [LangChangeCommand ('en_GB'), 'Cancel  button']
ERROR - updateCheck.UpdateDownloader._bg (20:29:40.938):
Error downloading
https://ci.appveyor.com/api/buildjobs/vg63itgyy5xrff9m/artifacts/output/nvda_snapshot_alpha-18266,1af87759.exe
Traceback (most recent call last):
  File "updateCheck.pyc", line 601, in _bg
  File "updateCheck.pyc", line 624, in _download
  File "urllib\request.pyc", line 222, in urlopen
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 563, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 755, in http_error_302
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 569, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 649, in http_error_default
urllib.error.HTTPError: HTTP Error 404: The specified blob does not exist.
IO - speech.speak (20:29:41.040):
Speaking [LangChangeCommand ('en_GB'), 'Error  dialog  Error downloading
update.']
IO - speech.speak (20:29:41.049):
Speaking [LangChangeCommand ('en_GB'), 'OK  button']

Brian


bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users
----- Original Message -----
From: "zvonimir stanečić, 9a5dsz" <zvonimirek222@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:02 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors



From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests



Hi everyone,



We have recently introduced new checks for Pull Requests on GitHub
<https://github.com/nvaccess/nvda/pull/9958> .



Code contributors regularly have to deal with ill-defined and inconsistently
enforced code style requirements. Code reviewers spend much of their time
reporting minor issues, time that would be better spent looking for
architectural problems / product issues / logic errors.



To improve this situation, in future PR builds the Flake8 linter will be run
checking any new code that will be introduced with a PR. This can also be
run with SCons using `scons lint base=origin/master`. Please see the
`tests/lint/readme.md` file for more information.



As part of a PR build, the Flake8 linter will be run checking any new code
that will be introduced with a PR. When Flake8 reports the does not comply
with it’s configuration, a message from Appveyor will be added as a comment
to the Pull Request, and the build will fail. Please note, that in this case
the artifacts (PR build executable) will still be available if those steps
were successful.



We have included an extension with the linter setup to allow tabs instead of
spaces, and my find it necessary to disable other warnings, or introduce
workarounds. However, in general we would like to stick as close to the
default settings as possible.



Thanks for all your contributions!

--

Reef Turner
Software Developer



 <https://www.nvaccess.org/> www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess

  <https://docs.google.com/uc?export=download&id=1ewvP2k8vO2yVx8f7qN46MLkUNt6pt8f_&revid=0B9qtAUOcqzA6dW5BQW55ZEp5UUptWjVYaXF5bnV4VC9qSGVRPQ>













--

Derek Riemer
Improving the world one byte at a time!        ⠠⠊⠍⠏⠗⠕⠧⠬ ⠸⠺ ⠐⠕ ⠃⠽⠞⠑ ⠁⠞ ⠐⠞⠖
•    Accessibility enthusiast.
•    Proud user of the NVDA screen reader.
•    Open source enthusiast.
•    Skier.

•    Personal website: https://derekriemer.com



 



--
Derek Riemer
Improving the world one byte at a time!        ⠠⠊⠍⠏⠗⠕⠧⠬ ⠮ ⠸⠺ ⠐⠕ ⠃⠽⠞⠑ ⠁⠞ ⠁ ⠐⠞⠖
•    Accessibility enthusiast.
•    Proud user of the NVDA screen reader.
•    Open source enthusiast.
•    Skier.

•    Personal website: https://derekriemer.com




Re: Alpha will not download, was Re: [nvda-devel] Extra checks forPull Requests

Arnold Loubriat
 

Hi Derek,

Just for reference: no it was not an AppVeyor outage.

See https://github.com/nvaccess/nvda/pull/10020 that got merged by Reef yesterday!

Have a nice day.

 

Arnold

 

De : derek riemer
Envoyé le :samedi 3 août 2019 07:04
À : nvda-devel@groups.io
Objet :Re: Alpha will not download, was Re: [nvda-devel] Extra checks forPull Requests

 

I think appVeyor had an outage or something. it's working again.

 

On Thu, Aug 1, 2019 at 1:34 PM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:

For completeness.

IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'space', EndUtteranceCommand()]
IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'pressed']
IO - speech.speak (20:29:39.388):
Speaking [LangChangeCommand ('en_GB'), 'Downloading Update  dialog
Connecting']
IO - speech.speak (20:29:39.393):
Speaking [LangChangeCommand ('en_GB'), 'Cancel  button']
ERROR - updateCheck.UpdateDownloader._bg (20:29:40.938):
Error downloading
https://ci.appveyor.com/api/buildjobs/vg63itgyy5xrff9m/artifacts/output/nvda_snapshot_alpha-18266,1af87759.exe
Traceback (most recent call last):
  File "updateCheck.pyc", line 601, in _bg
  File "updateCheck.pyc", line 624, in _download
  File "urllib\request.pyc", line 222, in urlopen
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 563, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 755, in http_error_302
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 569, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 649, in http_error_default
urllib.error.HTTPError: HTTP Error 404: The specified blob does not exist.
IO - speech.speak (20:29:41.040):
Speaking [LangChangeCommand ('en_GB'), 'Error  dialog  Error downloading
update.']
IO - speech.speak (20:29:41.049):
Speaking [LangChangeCommand ('en_GB'), 'OK  button']

Brian


bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users
----- Original Message -----
From: "zvonimir stanečić, 9a5dsz" <zvonimirek222@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:02 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors



From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests



Hi everyone,



We have recently introduced new checks for Pull Requests on GitHub
<https://github.com/nvaccess/nvda/pull/9958> .



Code contributors regularly have to deal with ill-defined and inconsistently
enforced code style requirements. Code reviewers spend much of their time
reporting minor issues, time that would be better spent looking for
architectural problems / product issues / logic errors.



To improve this situation, in future PR builds the Flake8 linter will be run
checking any new code that will be introduced with a PR. This can also be
run with SCons using `scons lint base=origin/master`. Please see the
`tests/lint/readme.md` file for more information.



As part of a PR build, the Flake8 linter will be run checking any new code
that will be introduced with a PR. When Flake8 reports the does not comply
with it’s configuration, a message from Appveyor will be added as a comment
to the Pull Request, and the build will fail. Please note, that in this case
the artifacts (PR build executable) will still be available if those steps
were successful.



We have included an extension with the linter setup to allow tabs instead of
spaces, and my find it necessary to disable other warnings, or introduce
workarounds. However, in general we would like to stick as close to the
default settings as possible.



Thanks for all your contributions!

--

Reef Turner
Software Developer



 <https://www.nvaccess.org/> www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess

  <https://docs.google.com/uc?export=download&id=1ewvP2k8vO2yVx8f7qN46MLkUNt6pt8f_&revid=0B9qtAUOcqzA6dW5BQW55ZEp5UUptWjVYaXF5bnV4VC9qSGVRPQ>













--

Derek Riemer
Improving the world one byte at a time!        ⠠⠊⠍⠏⠗⠕⠧⠬ ⠸⠺ ⠐⠕ ⠃⠽⠞⠑ ⠁⠞ ⠐⠞⠖
•    Accessibility enthusiast.
•    Proud user of the NVDA screen reader.
•    Open source enthusiast.
•    Skier.

•    Personal website: https://derekriemer.com



 

Re: Python 3 and Windows API functions: please standardize around Unicode (wide character) functions unless justified otherwise

Brian's Mail list account
 

That explains why keypresses in extended winamp are ignored then as effectively their is a mismatch wheree there should be a match.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "Joseph Lee" <@joslee>
To: <nvda-devel@groups.io>
Sent: Saturday, August 03, 2019 2:11 AM
Subject: [nvda-devel] Python 3 and Windows API functions: please standardize around Unicode (wide character) functions unless justified otherwise


Hi all,



The following Python 3 porting directive applies to anyone working with
Windows API functions, specifically ones requiring strings for certain
operations:



August 2019 Python 3 add-on directive no. 1: standardize around Unicode
(wide character) functions unless justified otherwise



Certain Windows API functions expect different string data format depending
on which function signature is invoked:



* *A: ANSI (bytes)
* *W: Unicode (wide characters)



Functions defined this way include FindWindow, FindWindowEx, and others.
Because Python 3 wants to use Unicode by default, you should call wide
character functions (FindWindowW, for example) unless justified otherwise.
If you absolutely need to use ANSI functions, please prefix strings with a
"b" (without quotes).



Examples:



# Yes:

HWND = winUser.user32.FindWindowW("windowClassName", None) # strictly Python
3

HWND = winUser.user32.FindWindowA(b"ANSIClassName", None) # Python 2 and 3

HWND = winUser.user32.FindWindowW(u"UnicodeClassName", None) # Python 2 and
3

HWND = winUser.user32.FindWindowA("windowClassName", None) # strictly Python
2



# No:

HWND = winUser.user32.FindWindowA("UnicodeClassName", None) # No in Python 3

HWND = winUser.user32.FindWindowW(b"ANSIClassName", None) # Wrong data type
across Python versions



Of these, I recommend going with option 3 (from first list), as it
guarantees widest compatibility.



The following add-ons are affected:



* Dropbox
* Extended Winamp
* StationPlaylist
* SystrayList
* Many more



Add-ons noted above may load under NVDA 2019.3 alpha if manifests are
modified. For add-ons hosted on community add-ons website, StationPlaylist
and SystrayList passes this check.



Next steps: please contact add-on authors regarding their thoughts on
compatibility.



Thanks.

Cheers,

Joseph



Re: Alpha will not download, was Re: [nvda-devel] Extra checks for Pull Requests

derek riemer
 

I think appVeyor had an outage or something. it's working again.

On Thu, Aug 1, 2019 at 1:34 PM Brian's Mail list account via Groups.Io <bglists=blueyonder.co.uk@groups.io> wrote:
For completeness.

IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'space', EndUtteranceCommand()]
IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'pressed']
IO - speech.speak (20:29:39.388):
Speaking [LangChangeCommand ('en_GB'), 'Downloading Update  dialog
Connecting']
IO - speech.speak (20:29:39.393):
Speaking [LangChangeCommand ('en_GB'), 'Cancel  button']
ERROR - updateCheck.UpdateDownloader._bg (20:29:40.938):
Error downloading
https://ci.appveyor.com/api/buildjobs/vg63itgyy5xrff9m/artifacts/output/nvda_snapshot_alpha-18266,1af87759.exe
Traceback (most recent call last):
  File "updateCheck.pyc", line 601, in _bg
  File "updateCheck.pyc", line 624, in _download
  File "urllib\request.pyc", line 222, in urlopen
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 563, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 755, in http_error_302
  File "urllib\request.pyc", line 531, in open
  File "urllib\request.pyc", line 641, in http_response
  File "urllib\request.pyc", line 569, in error
  File "urllib\request.pyc", line 503, in _call_chain
  File "urllib\request.pyc", line 649, in http_error_default
urllib.error.HTTPError: HTTP Error 404: The specified blob does not exist.
IO - speech.speak (20:29:41.040):
Speaking [LangChangeCommand ('en_GB'), 'Error  dialog  Error downloading
update.']
IO - speech.speak (20:29:41.049):
Speaking [LangChangeCommand ('en_GB'), 'OK  button']

Brian


bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users
----- Original Message -----
From: "zvonimir stanečić, 9a5dsz" <zvonimirek222@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:02 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors



From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests



Hi everyone,



We have recently introduced new checks for Pull Requests on GitHub
<https://github.com/nvaccess/nvda/pull/9958> .



Code contributors regularly have to deal with ill-defined and inconsistently
enforced code style requirements. Code reviewers spend much of their time
reporting minor issues, time that would be better spent looking for
architectural problems / product issues / logic errors.



To improve this situation, in future PR builds the Flake8 linter will be run
checking any new code that will be introduced with a PR. This can also be
run with SCons using `scons lint base=origin/master`. Please see the
`tests/lint/readme.md` file for more information.



As part of a PR build, the Flake8 linter will be run checking any new code
that will be introduced with a PR. When Flake8 reports the does not comply
with it’s configuration, a message from Appveyor will be added as a comment
to the Pull Request, and the build will fail. Please note, that in this case
the artifacts (PR build executable) will still be available if those steps
were successful.



We have included an extension with the linter setup to allow tabs instead of
spaces, and my find it necessary to disable other warnings, or introduce
workarounds. However, in general we would like to stick as close to the
default settings as possible.



Thanks for all your contributions!

--

Reef Turner
Software Developer



 <https://www.nvaccess.org/> www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess

  <https://docs.google.com/uc?export=download&id=1ewvP2k8vO2yVx8f7qN46MLkUNt6pt8f_&revid=0B9qtAUOcqzA6dW5BQW55ZEp5UUptWjVYaXF5bnV4VC9qSGVRPQ>














--
Derek Riemer
Improving the world one byte at a time!        ⠠⠊⠍⠏⠗⠕⠧⠬ ⠮ ⠸⠺ ⠐⠕ ⠃⠽⠞⠑ ⠁⠞ ⠁ ⠐⠞⠖
•    Accessibility enthusiast.
•    Proud user of the NVDA screen reader.
•    Open source enthusiast.
•    Skier.

•    Personal website: https://derekriemer.com




Python 3 and Windows API functions: please standardize around Unicode (wide character) functions unless justified otherwise

 

Hi all,

 

The following Python 3 porting directive applies to anyone working with Windows API functions, specifically ones requiring strings for certain operations:

 

August 2019 Python 3 add-on directive no. 1: standardize around Unicode (wide character) functions unless justified otherwise

 

Certain Windows API functions expect different string data format depending on which function signature is invoked:

 

  • *A: ANSI (bytes)
  • *W: Unicode (wide characters)

 

Functions defined this way include FindWindow, FindWindowEx, and others. Because Python 3 wants to use Unicode by default, you should call wide character functions (FindWindowW, for example) unless justified otherwise. If you absolutely need to use ANSI functions, please prefix strings with a “b” (without quotes).

 

Examples:

 

# Yes:

HWND = winUser.user32.FindWindowW(“windowClassName”, None) # strictly Python 3

HWND = winUser.user32.FindWindowA(b“ANSIClassName”, None) # Python 2 and 3

HWND = winUser.user32.FindWindowW(u”UnicodeClassName”, None) # Python 2 and 3

HWND = winUser.user32.FindWindowA(“windowClassName”, None) # strictly Python 2

 

# No:

HWND = winUser.user32.FindWindowA(“UnicodeClassName”, None) # No in Python 3

HWND = winUser.user32.FindWindowW(b”ANSIClassName”, None) # Wrong data type across Python versions

 

Of these, I recommend going with option 3 (from first list), as it guarantees widest compatibility.

 

The following add-ons are affected:

 

  • Dropbox
  • Extended Winamp
  • StationPlaylist
  • SystrayList
  • Many more

 

Add-ons noted above may load under NVDA 2019.3 alpha if manifests are modified. For add-ons hosted on community add-ons website, StationPlaylist and SystrayList passes this check.

 

Next steps: please contact add-on authors regarding their thoughts on compatibility.

 

Thanks.

Cheers,

Joseph

Re: Extra checks for Pull Requests

 

Patrick,


it might help to mention your concerns with the code style in the pr itself.


Regards,

Leonard

On 2-8-2019 12:05, Patrick ZAJDA via Groups.Io wrote:

Hello,


Finaly I fixd it, even if I am really not convinced by this kind of codestyle (PR9707).


Have a good day,


Patrick


Le 01/08/2019 à 22:33, Patrick ZAJDA via Groups.Io a écrit :

Hello Reef,

First of all, nice to see this great news.

It looks like there is something strange, or globalCommand.py must be reviewed :)

The "script" decorator make these new tests failed with this message:

"

E121 continuation line under-indented for hanging indent"

I wanted to test with my pull request as it is not a bug I though it could be good to see what happens.

I might have done something bad, even if I read and the other script with the decorator has the same style as the code I made.

What should I do to pass these tests?

Thanks,

Patrick

Le 01/08/2019 à 20:58, Reef Turner a écrit :

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess  Twitter: @NVAccess 

 

--
Patrick ZAJDA
Certification NVDA 2019
--
Patrick ZAJDA
Certification NVDA 2019

Re: Extra checks for Pull Requests

Patrick ZAJDA
 

Hello,


You can try this link:


https://ci.appveyor.com/api/buildjobs/vg63itgyy5xrff9m/artifacts/nvda_snapshot_alpha-18266%2C1af87759.exe


Downloaded it three times with success and got it following Reef's instructions.


Hop this help,


Patrick

Le 02/08/2019 à 10:22, Brian's Mail list account via Groups.Io a écrit :
Hmm, well I think I'll wait for a fix. grin.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users
----- Original Message ----- From: "Reef Turner" <reef@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 9:28 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Thanks for the heads up. This is due to the changes I made to how appveyor uploads files. I’ll push a fix for this in the morning. For now, if you want the lastest alpha, you will have to download it directly from the appveyor artifacts page.

Here is a direct link for now. Until this is fixed you can do the following:

1. Go to our snapshots page
2. Copy the link  URL for the build, eg ‘alpha-18266,1af87759’ not ‘download’
3. Paste in the address bar and add ‘/artifacts’ and press enter
4. Find the link for ‘nvda_snapshot_alpha-18266,1af87759.exe’ and press enter.

From: zvonimir stanečić, 9a5dsz
Sent: Thursday, 1 August 2019 9:03 PM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Extra checks for Pull Requests

Hi Reef,
To hijack the thread…
Why it’s not possible to download the latest alpha?
Somewhere we have 404 errors

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests

Hi everyone,

We have recently introduced new checks for Pull Requests on GitHub.

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

Thanks for all your contributions!
--
Patrick ZAJDA
Certification NVDA 2019

Re: Extra checks for Pull Requests

Patrick ZAJDA
 

Hello,


Finaly I fixd it, even if I am really not convinced by this kind of codestyle (PR9707).


Have a good day,


Patrick


Le 01/08/2019 à 22:33, Patrick ZAJDA via Groups.Io a écrit :

Hello Reef,

First of all, nice to see this great news.

It looks like there is something strange, or globalCommand.py must be reviewed :)

The "script" decorator make these new tests failed with this message:

"

E121 continuation line under-indented for hanging indent"

I wanted to test with my pull request as it is not a bug I though it could be good to see what happens.

I might have done something bad, even if I read and the other script with the decorator has the same style as the code I made.

What should I do to pass these tests?

Thanks,

Patrick

Le 01/08/2019 à 20:58, Reef Turner a écrit :

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess  Twitter: @NVAccess 

 

--
Patrick ZAJDA
Certification NVDA 2019
--
Patrick ZAJDA
Certification NVDA 2019

Re: Extra checks for Pull Requests

Brian's Mail list account
 

Hmm, well I think I'll wait for a fix. grin.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "Reef Turner" <reef@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 9:28 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Thanks for the heads up. This is due to the changes I made to how appveyor uploads files. I’ll push a fix for this in the morning. For now, if you want the lastest alpha, you will have to download it directly from the appveyor artifacts page.

Here is a direct link for now. Until this is fixed you can do the following:

1. Go to our snapshots page
2. Copy the link URL for the build, eg ‘alpha-18266,1af87759’ not ‘download’
3. Paste in the address bar and add ‘/artifacts’ and press enter
4. Find the link for ‘nvda_snapshot_alpha-18266,1af87759.exe’ and press enter.

From: zvonimir stanečić, 9a5dsz
Sent: Thursday, 1 August 2019 9:03 PM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Extra checks for Pull Requests

Hi Reef,
To hijack the thread…
Why it’s not possible to download the latest alpha?
Somewhere we have 404 errors

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests

Hi everyone,

We have recently introduced new checks for Pull Requests on GitHub.

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

Thanks for all your contributions!
--
Reef Turner
Software Developer

www.nvaccess.org
Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess










--------------------------------------------------------------------------------

Re: Extra checks for Pull Requests

Patrick ZAJDA
 

Hello Reef,

First of all, nice to see this great news.

It looks like there is something strange, or globalCommand.py must be reviewed :)

The "script" decorator make these new tests failed with this message:

"

E121 continuation line under-indented for hanging indent"

I wanted to test with my pull request as it is not a bug I though it could be good to see what happens.

I might have done something bad, even if I read and the other script with the decorator has the same style as the code I made.

What should I do to pass these tests?

Thanks,

Patrick

Le 01/08/2019 à 20:58, Reef Turner a écrit :

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess  Twitter: @NVAccess 

 

--
Patrick ZAJDA
Certification NVDA 2019

Re: Extra checks for Pull Requests

Reef Turner
 

Thanks for the heads up. This is due to the changes I made to how appveyor uploads files. I’ll push a fix for this in the morning. For now, if you want the lastest alpha, you will have to download it directly from the appveyor artifacts page.

 

Here is a direct link for now. Until this is fixed you can do the following:

 

  1. Go to our snapshots page
  2. Copy the link  URL for the build, eg ‘alpha-18266,1af87759’ not ‘download’
  3. Paste in the address bar and add ‘/artifacts’ and press enter
  4. Find the link for ‘nvda_snapshot_alpha-18266,1af87759.exe’ and press enter.

 

From: zvonimir stanečić, 9a5dsz
Sent: Thursday, 1 August 2019 9:03 PM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Extra checks for Pull Requests

 

Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests

 

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner
Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess 
Twitter: @NVAccess 

 

 

Re: Can nvda make a mouse user crash windows Explorer?

Pettyjohn, Chris G. (FTC)
 

i can make windows explorer crash every time with the steps below. It only happens with mouse movement. 
Steps to replicate.
1. With NVDA on,  click the Windows Start button
2. When the Start Menu opens move the mouse into it like you are going to select a program to run. 
3. For me, as soon as the mouse moves off the Start button and travels into the Windows Explorer window  a pop up box appears with crash error.

Alpha will not download, was Re: [nvda-devel] Extra checks for Pull Requests

Brian's Mail list account
 

For completeness.

IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'space', EndUtteranceCommand()]
IO - speech.speak (20:29:39.220):
Speaking [LangChangeCommand ('en_GB'), 'pressed']
IO - speech.speak (20:29:39.388):
Speaking [LangChangeCommand ('en_GB'), 'Downloading Update dialog Connecting']
IO - speech.speak (20:29:39.393):
Speaking [LangChangeCommand ('en_GB'), 'Cancel button']
ERROR - updateCheck.UpdateDownloader._bg (20:29:40.938):
Error downloading https://ci.appveyor.com/api/buildjobs/vg63itgyy5xrff9m/artifacts/output/nvda_snapshot_alpha-18266,1af87759.exe
Traceback (most recent call last):
File "updateCheck.pyc", line 601, in _bg
File "updateCheck.pyc", line 624, in _download
File "urllib\request.pyc", line 222, in urlopen
File "urllib\request.pyc", line 531, in open
File "urllib\request.pyc", line 641, in http_response
File "urllib\request.pyc", line 563, in error
File "urllib\request.pyc", line 503, in _call_chain
File "urllib\request.pyc", line 755, in http_error_302
File "urllib\request.pyc", line 531, in open
File "urllib\request.pyc", line 641, in http_response
File "urllib\request.pyc", line 569, in error
File "urllib\request.pyc", line 503, in _call_chain
File "urllib\request.pyc", line 649, in http_error_default
urllib.error.HTTPError: HTTP Error 404: The specified blob does not exist.
IO - speech.speak (20:29:41.040):
Speaking [LangChangeCommand ('en_GB'), 'Error dialog Error downloading update.']
IO - speech.speak (20:29:41.049):
Speaking [LangChangeCommand ('en_GB'), 'OK button']

Brian


bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "zvonimir stanečić, 9a5dsz" <zvonimirek222@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:02 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors



From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests



Hi everyone,



We have recently introduced new checks for Pull Requests on GitHub <https://github.com/nvaccess/nvda/pull/9958> .



Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.



To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.



As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.



We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.



Thanks for all your contributions!

--

Reef Turner
Software Developer



<https://www.nvaccess.org/> www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess

<https://docs.google.com/uc?export=download&id=1ewvP2k8vO2yVx8f7qN46MLkUNt6pt8f_&revid=0B9qtAUOcqzA6dW5BQW55ZEp5UUptWjVYaXF5bnV4VC9qSGVRPQ>

Re: Extra checks for Pull Requests

Brian's Mail list account
 

Yes I get a nasty dong when it tries to auto update. Probably a broken link somewhere.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "zvonimir stanečić, 9a5dsz" <zvonimirek222@...>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:02 PM
Subject: Re: [nvda-devel] Extra checks for Pull Requests


Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors



From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests



Hi everyone,



We have recently introduced new checks for Pull Requests on GitHub <https://github.com/nvaccess/nvda/pull/9958> .



Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.



To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.



As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.



We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.



Thanks for all your contributions!

--

Reef Turner
Software Developer



<https://www.nvaccess.org/> www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess
Twitter: @NVAccess

<https://docs.google.com/uc?export=download&id=1ewvP2k8vO2yVx8f7qN46MLkUNt6pt8f_&revid=0B9qtAUOcqzA6dW5BQW55ZEp5UUptWjVYaXF5bnV4VC9qSGVRPQ>

Re: Can nvda make a mouse user crash windows Explorer?

Brian's Mail list account
 

Further, I don't think it was doing it back in January, so I wonder if a windows patch may be causing a bit of code indigestion in nvda. Its annoying as I cannot get it to do it since I think you need precise mouse movements for it to happen.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "Brian's Mail list account via Groups.Io" <bglists=blueyonder.co.uk@groups.io>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 8:19 PM
Subject: Re: [nvda-devel] Can nvda make a mouse user crash windows Explorer?


I never saw that but what is happening is that my computer fixer had the problem until he turned it off and now its hit our studio machine. If I use it via keyboard I seldom get any trouble. The effect also affects inside files and folders, IE copy a track from one place to another. Sometimes it will behave sometimes not.
Its irritating to my sighted friends, so I told them to turn off nvda.
Does it do this with no add ons running by the way?

This is 2019.1.1
Windows 7 64.
Its got to be something really obsscure like it jumping out of a routine in the wrong place or something.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users
----- Original Message -----
From: "Pettyjohn, Chris G. (FTC) via Groups.Io" <Chris.Pettyjohn=va.gov@groups.io>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 3:34 PM
Subject: Re: [nvda-devel] Can nvda make a mouse user crash windows Explorer?


Here was my post. You actually responded to it.
Possible Bug with 2019.1 - crashes Windows Explorer ( https://nvda.groups.io/g/nvda/topic/possible_bug_with_2019_1/30822907?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,2,680,30822907 )

Steps to replicate.
1. With NVDA on, click the Windows Start button
2. When the Start Menu opens move the mouse into it like you are going to select a program to run.
3. For me, as soon as the mouse moves off the Start button and travels into the Windows Explorer window a pop up box appears with crash error.





Re: Can nvda make a mouse user crash windows Explorer?

Brian's Mail list account
 

I never saw that but what is happening is that my computer fixer had the problem until he turned it off and now its hit our studio machine. If I use it via keyboard I seldom get any trouble. The effect also affects inside files and folders, IE copy a track from one place to another. Sometimes it will behave sometimes not.
Its irritating to my sighted friends, so I told them to turn off nvda.
Does it do this with no add ons running by the way?

This is 2019.1.1
Windows 7 64.
Its got to be something really obsscure like it jumping out of a routine in the wrong place or something.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
Newsgroup monitored: alt.comp.blind-users

----- Original Message -----
From: "Pettyjohn, Chris G. (FTC) via Groups.Io" <Chris.Pettyjohn=va.gov@groups.io>
To: <nvda-devel@groups.io>
Sent: Thursday, August 01, 2019 3:34 PM
Subject: Re: [nvda-devel] Can nvda make a mouse user crash windows Explorer?


Here was my post. You actually responded to it.
Possible Bug with 2019.1 - crashes Windows Explorer ( https://nvda.groups.io/g/nvda/topic/possible_bug_with_2019_1/30822907?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,2,680,30822907 )

Steps to replicate.
1. With NVDA on, click the Windows Start button
2. When the Start Menu opens move the mouse into it like you are going to select a program to run.
3. For me, as soon as the mouse moves off the Start button and travels into the Windows Explorer window a pop up box appears with crash error.

Re: Extra checks for Pull Requests

zvonimir stanečić, 9a5dsz
 

Hi Reef,

To hijack the thread…

Why it’s not possible to download the latest alpha?

Somewhere we have 404 errors

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Reef Turner
Sent: Thursday, August 1, 2019 8:59 PM
To: nvda-devel@groups.io
Subject: [nvda-devel] Extra checks for Pull Requests

 

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner
Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess 
Twitter: @NVAccess 

 

Extra checks for Pull Requests

Reef Turner
 

Hi everyone,

 

We have recently introduced new checks for Pull Requests on GitHub.

 

Code contributors regularly have to deal with ill-defined and inconsistently enforced code style requirements. Code reviewers spend much of their time reporting minor issues, time that would be better spent looking for architectural problems / product issues / logic errors.

 

To improve this situation, in future PR builds the Flake8 linter will be run checking any new code that will be introduced with a PR. This can also be run with SCons using `scons lint base=origin/master`. Please see the `tests/lint/readme.md` file for more information.

 

As part of a PR build, the Flake8 linter will be run checking any new code that will be introduced with a PR. When Flake8 reports the does not comply with it’s configuration, a message from Appveyor will be added as a comment to the Pull Request, and the build will fail. Please note, that in this case the artifacts (PR build executable) will still be available if those steps were successful.

 

We have included an extension with the linter setup to allow tabs instead of spaces, and my find it necessary to disable other warnings, or introduce workarounds. However, in general we would like to stick as close to the default settings as possible.

 

Thanks for all your contributions!

--

Reef Turner
Software Developer 

 

www.nvaccess.org

Facebook: https://www.facebook.com/NVAccess 
Twitter: @NVAccess 

 

Re: Can nvda make a mouse user crash windows Explorer?

Ralf Kefferpuetz
 

This was also reported in the german NVDA mailinglist some days ago.

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Pettyjohn, Chris G. (FTC) via Groups.Io
Sent: Donnerstag, 1. August 2019 16:35
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Can nvda make a mouse user crash windows Explorer?

 

Here was my post.  You actually responded to it. 
Possible Bug with 2019.1 - crashes Windows Explorer

Steps to replicate.
1. With NVDA on,  click the Windows Start button
2. When the Start Menu opens move the mouse into it like you are going to select a program to run. 
3. For me, as soon as the mouse moves off the Start button and travels into the Windows Explorer window  a pop up box appears with crash error.

Re: Query regarding the porting of NVDA to Python3

Travis Siegel
 

Seriously folks, 64-bit NVDA is a nice to have, not a necessity.  There are multiple 3rd party libraries NVDA relies on, and unless each and every one of those are ported to 64-bit, producing a 64-bit version won't help.  Also, in general, unless a program uses large amounts of memory, or does other cpu intensive tasks, 64-bit versions will provide no benefit.  Often times, this rush to get everything to 64-bit actually hurts a program's performance, because the program is not optimized for 64-bit functionality, and the program suffers performance penalties as a result.  Joseph is doing the right thing here by working into the transition from python 2 to 3 slowly, and looking at 64-bit conversion later.  There's no benefit to producing a 64-bit version of NVDA if all of the speech synthesizers use 32-bit versions, that will at best provide no benefit, and at worst, make the program unstable.  We really don't want that, so folks, hold your horses, take a deep breath, and know that 64-bit versions of NVDA will appear when the developers determine it's appropriate to do so, and not before.  It's a voluntary effort, and nobody is going to benefit from a 64-bit version if it doesn't operate properly.

On 7/31/2019 7:45 PM, Joseph Lee wrote:

Hi,

This also means storing at least two binaries on the server, which might not be ideal if one wishes to create a portable copy that can run anywhere.

Cheers,

Joseph

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of enes saribas
Sent: Wednesday, July 31, 2019 4:44 PM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Query regarding the porting of NVDA to Python3

 

Hi Joseph,

In that case, couldn't there be a web installer, which checks operating system version, and downloads the correct version based on that information? Firefox for example does this.

 

On 7/31/2019 10:25 AM, Joseph Lee wrote:

Hi,

I think we discussed this at length a while ago, and consensus is that it is better to keep them as one for ease of portability.

Cheers,

Joseph

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of enes saribas
Sent: Wednesday, July 31, 2019 8:24 AM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Query regarding the porting of NVDA to Python3

 

Hi Joseph,

Couldn't these issues be solved by having two installers (32 and 64 bit), instead of one?

Best,

Enes

On 7/30/2019 10:22 AM, Joseph Lee wrote:

Hi,

Another downside is that x64 NVDA build won’t run on ARM64 on older Windows 10 releases because the emulator only supports x86 32-bit, although moving to 64-bit means a native ARM build becomes theoretically possible.

Cheers,

Joseph

 

From: nvda-devel@groups.io <nvda-devel@groups.io> On Behalf Of Leonard de Ruijter
Sent: Tuesday, July 30, 2019 1:07 AM
To: nvda-devel@groups.io
Subject: Re: [nvda-devel] Query regarding the porting of NVDA to Python3

 

Hello,

 

As far as I"m aware, there are currently no plans to create an x64 version of NVDA. No one stops anyone from putting some research into it, but I think the benefits it would have are mostly theoretical ones. It should be noted that we wouldn't be able to cover x86 systems with such a build of NVDA, so that's probably the most major reason not to investigate it further at this time.

 

Regards,

Leonard

 


Virus-free. www.avast.com