bpo-24241: Preferred X browser #85

davesteele · 2017-02-14T02:37:52Z

When calling webbrowser.open*(), the module goes through a list of installed browsers, and uses the first one that succeeds, to process the request.

The first 'browsers' in the 'X' list are 'xdg-open' and others of that ilk. The problem is that they only have one 'open' behavior - the 'new' parameter is ignored ('same window', 'new window', 'new tab').

These commits move the default web browser to the front of the tryorder list, allowing get() behavior to match the documentation.

This same problem was recently fixed in Windows via Issue #8232

the-knights-who-say-ni · 2017-02-14T02:37:53Z

Hello, and thanks for your contribution!

I'm a bot set up to make sure that the project can legally accept your contribution by verifying you have signed the PSF contributor agreement (CLA).

Unfortunately we couldn't find an account corresponding to your GitHub username on bugs.python.org (b.p.o) to verify you have signed the CLA. This is necessary for legal reasons before we can look at your contribution. Please follow these steps to help rectify the issue:

If you don't have an account on b.p.o, please create one
Make sure your GitHub username is listed in "Your Details" at b.p.o
If you have not already done so, please sign the PSF contributor agreement
If you just signed the CLA, please wait at least one US business day and then check "Your Details" on bugs.python.org to see if your account has been marked as having signed the CLA (the delay is due to a person having to manually check your signed CLA)
Reply here saying you have completed the above steps

Thanks again to your contribution and we look forward to looking at it!

davesteele · 2017-02-14T02:43:28Z

I've tied my bpo 'daves' account to github 'davesteele'. The contributor form was received 2015-07-17.00:00:00.

codecov · 2017-02-14T03:12:48Z

Codecov Report

Merging #85 into master will decrease coverage by -0.01%.
The diff coverage is 14.28%.

@@            Coverage Diff             @@
##           master      #85      +/-   ##
==========================================
- Coverage   82.38%   82.38%   -0.01%     
==========================================
  Files        1428     1428              
  Lines      351138   351143       +5     
==========================================
- Hits       289291   289275      -16     
- Misses      61847    61868      +21

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b692dc8...602ba9f. Read the comment docs.

ultimatecoder

Found improvements to the code.

ultimatecoder · 2017-02-18T12:54:45Z

Lib/webbrowser.py

+        cmd = "xdg-settings get default-web-browser".split()
+        _preferred_browser = subprocess.check_output(cmd).decode().strip()
+    except (FileNotFoundError, subprocess.CalledProcessError):
+        pass


I am not sure you should pass the exceptions silently.

I'm not sure what to report, otherwise. Passing on the FileNotFoundError feels right to me.

I could drop catching CalledProcessError?

You could use shutil.which to determine if xdg-settings exists before trying to call it (LBYL). But here ignoring the exception (EAFP) seems right to me.

With a _preferred_browsers set, this would become:

try: cmd = "xdg-settings get default-web-browser".split() _xdg_default_browser = subprocess.check_output(cmd).decode().strip() except (FileNotFoundError, subprocess.CalledProcessError): pass else: _preferred_browsers.add(_xdg_default_browser)

I believe that would be more self-explanatory as to the purpose of ignoring the exceptions here (if we fail to find the xdg default browser, we simply don't add it to the preferred browser set)

ultimatecoder

I will suggest updating the documentation for method register at Doc/library/webbrowser.rst

merwok · 2017-02-18T18:45:02Z

Lib/webbrowser.py


-def register(name, klass, instance=None, update_tryorder=1):
-    """Register a browser connector and, optionally, connection."""
+def register(name, klass, instance=None, priority=False):


I wonder if it’s good to change the signature here. Is it impossible to give high priority to the browser found with xdg-settings, if any, with the update_tryorder param?

I took the opportunity to make what I thought was an improvement. The tri-state definition of the old parameter seemed a bit clumsy and unnecessary (note that it was left out of the documentation).

I can back out the change to update_tryorder if it is deemed appropriate.

I'd suggest making the parameter preferred rather than priority and making it keyword-only.

davesteele · 2017-02-18T21:41:11Z

I've added documentation for the register() priority argument.

ncoghlan

I like the general idea of this patch, but think it would be more self-explanatory with some changes to the specific implementation details.

ncoghlan · 2017-02-19T10:50:14Z

Lib/webbrowser.py

@@ -15,14 +15,15 @@ class Error(Exception):

 _browsers = {}          # Dictionary of available browser controllers
 _tryorder = []          # Preference order of available browsers
+_preferred_browser = None # The preferred browser for the current environment


Rather than only allowing a single preferred browser, perhaps it would make sense to make this a _preferred_browsers set? That would have a ripple effect on the other updates below.

My original implementation did this, but I didn't find the enhancement useful. There weren't any ripples.

That said, I'll make the change if this is considered a blocker.

ncoghlan · 2017-02-19T10:51:00Z

Lib/webbrowser.py

-    if update_tryorder > 0:
-        _tryorder.append(name)
-    elif update_tryorder < 0:
+    if priority or (_preferred_browser and name in _preferred_browser):


With a set, this condition would become if priority or name in _preferred_browsers:

'preferred' is a better name. I'll make the change.

I had managed to make it to this point without encountering the keyword-only concept in the library (or elsewhere, for that matter). Would this cause unnecessary confusion?

Note that 'name in _preferred_browsers' is not enough. xdg-settings returns e.g. 'firefox.desktop'.

keyword-only is our preferred approach to handling boolean arguments in Python 3, as otherwise you end up with code that's ambiguous at the point of calling (as here, where the True conveys no info to a reader that doesn't already know the signature, while preferred=True is a lot more descriptive)

Regarding _preferred_browser potentially needing to match things like firefox against firefox.desktop, that belongs in an explanatory comment in the code itself. I'm fine with dropping the idea of switching to a set, though.

A slight name tweak may help convey the difference between this field and the new parameter to register though: what do you think about renaming it to _os_preferred_browser?

ncoghlan · 2017-02-19T10:54:22Z

Lib/webbrowser.py

+        cmd = "xdg-settings get default-web-browser".split()
+        _preferred_browser = subprocess.check_output(cmd).decode().strip()
+    except (FileNotFoundError, subprocess.CalledProcessError):
+        pass


With a _preferred_browsers set, this would become:

try: cmd = "xdg-settings get default-web-browser".split() _xdg_default_browser = subprocess.check_output(cmd).decode().strip() except (FileNotFoundError, subprocess.CalledProcessError): pass else: _preferred_browsers.add(_xdg_default_browser)

I believe that would be more self-explanatory as to the purpose of ignoring the exceptions here (if we fail to find the xdg default browser, we simply don't add it to the preferred browser set)

ncoghlan · 2017-02-19T10:55:36Z

Lib/webbrowser.py


-def register(name, klass, instance=None, update_tryorder=1):
-    """Register a browser connector and, optionally, connection."""
+def register(name, klass, instance=None, priority=False):


I'd suggest making the parameter preferred rather than priority and making it keyword-only.

ncoghlan · 2017-02-19T10:56:38Z

Lib/webbrowser.py

-    register("firefox", None, MacOSXOSAScript('firefox'), -1)
-    register("chrome", None, MacOSXOSAScript('chrome'), -1)
-    register("MacOSX", None, MacOSXOSAScript('default'), -1)
+    register("safari", None, MacOSXOSAScript('safari'), True)


With the above suggested change to the signature of register, this and the other options below would become:

register("safari", None, MacOSXOSAScript('safari'), preferred=True)

I'll add the 'else' construct.

codecov-io · 2017-02-19T17:36:55Z

Codecov Report

Merging #85 into master will decrease coverage by -0.01%.
The diff coverage is 13.33%.

@@            Coverage Diff             @@
##           master      #85      +/-   ##
==========================================
- Coverage   82.38%   82.37%   -0.01%     
==========================================
  Files        1428     1428              
  Lines      351138   351206      +68     
==========================================
+ Hits       289291   289317      +26     
- Misses      61847    61889      +42

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b692dc8...795ee09. Read the comment docs.

davesteele · 2017-02-19T18:08:20Z

@ncoghlan, I've made some changes based on your review, and pushed back on a few others. You be the final arbiter on the issues. Let me know.

ncoghlan · 2017-02-23T14:19:38Z

+1 on leaving the helper variable as a single string - I've suggested a name change and an additional explanatory comment to cover some of the subtleties for future maintainers.

For the preferred change, I still think keyword-only arguments will be clearer, as preferred=True conveys a lot more information at the call sites than a bare True.

The commit message also still needs adjustment to better cover the changes being made (i.e. both the preferred parameter being added to webbrowser.register, and the module being updated to use xdg-settings to find the appropriate preferred browser on Linux).

If you're so inclined, you could also add an entry to the 3.7 What's New for this change (otherwise I'll just add it along with the NEWS entry when pushing the change)

Replace the existing undocumented tri-state 'try_order' parameter with the boolean keyword-only 'preferred' parameter. Setting it to True places the browser at the front of the list, preferring it as the return to a subsequent get() call. 'preferred' is added to the documentation.

The traditional first entry in the tryorder queue is xdg-open, which doesn't support new window. Make the default browser first in the try list, to properly support these options.

davesteele · 2017-02-24T22:55:05Z

@ncoghlan, I've reworked, reworded, and rebased the commits based on your suggestions.

ncoghlan · 2017-02-25T04:42:10Z

The codecov complaints are due to CI not running on Mac OS X, so this looks good to me now - merging!

davesteele · 2017-02-25T04:48:34Z

Thanks for your help on this.

ncoghlan · 2017-05-28T05:08:13Z

Added the sprint label, as this PR was submitted at the PyCon Pune 2017 core development sprint.

…WatchdogEx Fix issue python#85: don't schedule already scheduled tasklets. Ensure that the internal watchdog list is empty before and after all test cases of class TestNewWatchdog. Fix test case TestNewWatchdog.test_watchdog_priority_{hard|soft}. The timeout value was much to short to build up a list of watchdogs. The new test cases TestNewWatchdog.test_schedule_deeper_{hard|soft} triggers the assertion failure reliably. They are skipped for now. https://bitbucket.org/stackless-dev/stackless/issues/85 (grafted from 2235702eb90cb0709dd40544b3952a956f2032a9 and 3fb55ce5b2d4)

Sometimes we need to acquire locks before the current PyThreadState is set. This moves the waiter data to it's own struct. It's shared between PyThreadStates on the same thread, such as in multi-interpreter settings. Fixes #85

Add github workflow to build and publish docker image

the-knights-who-say-ni added the CLA not signed label Feb 14, 2017

davesteele changed the title ~~Issue #24241: Preferred X browser~~ bpo-24241: Preferred X browser Feb 14, 2017

zware removed the CLA not signed label Feb 14, 2017

the-knights-who-say-ni added the CLA signed label Feb 14, 2017

ultimatecoder reviewed Feb 18, 2017

View reviewed changes

ultimatecoder suggested changes Feb 18, 2017

View reviewed changes

merwok reviewed Feb 18, 2017

View reviewed changes

ncoghlan requested changes Feb 19, 2017

View reviewed changes

davesteele added 2 commits February 24, 2017 16:27

bpo-24241: xdg-settings for preferred X browser

871455f

The traditional first entry in the tryorder queue is xdg-open, which doesn't support new window. Make the default browser first in the try list, to properly support these options.

davesteele force-pushed the preferred_browser branch from 795ee09 to 871455f Compare February 24, 2017 21:37

ncoghlan approved these changes Feb 25, 2017

View reviewed changes

ncoghlan merged commit e3ce695 into python:master Feb 25, 2017

ncoghlan mentioned this pull request Feb 25, 2017

Require code coverage not drop in a PR python/core-workflow#38

Closed

ncoghlan added the sprint label May 28, 2017

akruis pushed a commit to akruis/cpython that referenced this pull request Sep 9, 2017

Enable a skiped test case for issue python#85

f76353a

akruis pushed a commit to akruis/cpython that referenced this pull request Sep 9, 2017

merge 3.4-slp (Stackless python#81, python#85)

4717ee3

jaraco pushed a commit that referenced this pull request Dec 2, 2022

Update aiohttp from 3.1.1 to 3.2.0 (#85)

443d0dd

markshannon pushed a commit to faster-cpython/cpython that referenced this pull request Jan 10, 2023

Merge pull request python#85 from lesteve/docker-build-push

1761801

Add github workflow to build and publish docker image

gvanrossum mentioned this pull request Aug 22, 2023

heap-use-after-free in _PyFunction_LookupByVersion #108253

Closed

ngoldbaum mentioned this pull request Sep 23, 2024

Crash running PyO3 tests with --test-threads=1000 #124375

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

bpo-24241: Preferred X browser #85

bpo-24241: Preferred X browser #85

davesteele commented Feb 14, 2017

the-knights-who-say-ni commented Feb 14, 2017

davesteele commented Feb 14, 2017

codecov bot commented Feb 14, 2017

ultimatecoder left a comment

ultimatecoder Feb 18, 2017

davesteele Feb 18, 2017

merwok Feb 18, 2017

ncoghlan Feb 19, 2017

ultimatecoder left a comment

merwok Feb 18, 2017

davesteele Feb 18, 2017

ncoghlan Feb 19, 2017

davesteele commented Feb 18, 2017

ncoghlan left a comment

ncoghlan Feb 19, 2017

davesteele Feb 19, 2017

ncoghlan Feb 19, 2017

davesteele Feb 19, 2017

davesteele Feb 19, 2017

davesteele Feb 19, 2017

ncoghlan Feb 23, 2017

ncoghlan Feb 23, 2017

ncoghlan Feb 19, 2017

ncoghlan Feb 19, 2017

ncoghlan Feb 19, 2017

davesteele Feb 19, 2017

codecov-io commented Feb 19, 2017

davesteele commented Feb 19, 2017

ncoghlan commented Feb 23, 2017

davesteele commented Feb 24, 2017

ncoghlan commented Feb 25, 2017

davesteele commented Feb 25, 2017

ncoghlan commented May 28, 2017

bpo-24241: Preferred X browser #85

bpo-24241: Preferred X browser #85

Conversation

davesteele commented Feb 14, 2017

the-knights-who-say-ni commented Feb 14, 2017

davesteele commented Feb 14, 2017

codecov bot commented Feb 14, 2017

Codecov Report

ultimatecoder left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ultimatecoder left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davesteele commented Feb 18, 2017

ncoghlan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov-io commented Feb 19, 2017

Codecov Report

davesteele commented Feb 19, 2017

ncoghlan commented Feb 23, 2017

davesteele commented Feb 24, 2017

ncoghlan commented Feb 25, 2017

davesteele commented Feb 25, 2017

ncoghlan commented May 28, 2017