bpo-32839: Add after_info to tkinter #5664

csabella · 2018-02-13T18:27:53Z

https://bugs.python.org/issue32839

Lib/tkinter/__init__.py

Lib/tkinter/test/test_tkinter/test_misc.py

serhiy-storchaka · 2018-02-13T19:14:19Z

Lib/tkinter/__init__.py

@@ -764,14 +764,32 @@ def after_cancel(self, id):
        Identifier returned by after or after_idle must be
        given as first parameter."""
        try:
-            data = self.tk.call('after', 'info', id)
+            data = self.after_info(id)
            # In Tk 8.3, splitlist returns: (script, type)
            # In Tk 8.4, splitlist may return (script, type) or (script,)


Check that these comments still are related.

I couldn't find any Tk/Tcl documentation that referenced this. Here is the commit that added the comment:
3c0f2c9

Maybe after_cancel had been called with None, which wouldn't have returned the 2-tuple?

Lib/tkinter/test/test_tkinter/test_misc.py

terryjreedy · 2018-02-13T22:36:21Z

Lib/tkinter/__init__.py

@@ -763,15 +763,31 @@ def after_cancel(self, id):

        Identifier returned by after or after_idle must be
        given as first parameter."""
+        if id is None:
+            return


This makes no sense to me. id is required and AFAKI should not be passed as None. Do after or after_idle ever return None?

Please see my following comment with the Tcl code. I added this because I'm not sure if the original comments about the tuple changing were accurate. I think it was a result of id being None vs it having a value.

terryjreedy · 2018-02-13T22:36:21Z

Lib/tkinter/__init__.py

-            data = self.tk.call('after', 'info', id)
-            # In Tk 8.3, splitlist returns: (script, type)
-            # In Tk 8.4, splitlist may return (script, type) or (script,)
+            data = self.after_info(id)


This change requires that tkinter after_info return a reference to the tcl wrapper, (which I think is wrong), so that the wrapper can be deleted (which it should be). I think after_cancel should continue using the tk.call here. The only change here might deleting the 8.3 note, which I believe is obsolete, but this is up to Serhiy.

It looks to me, that after info $id returned a list of 2 elements not in all versions. We can now ignore old unsupported versions, but we should check sources of supported versions. We shouldn't believe the documentation, it contains errors.

I've looked at the 8.3 Tcl code (from 2001) and the current (8.6) code. I'm not sure if the comment about 8.3 and 8.4 was accurate. My thought is that somehow, id was None on a call, thus causing info to return the list of event_handlers instead of a tuple, as the previous code was expecting.

Here is the 8.3 Tcl code:

case AFTER_INFO: { Tcl_Obj *resultListPtr; if (objc == 2) { for (afterPtr = assocPtr->firstAfterPtr; afterPtr != NULL; afterPtr = afterPtr->nextPtr) { if (assocPtr->interp == interp) { sprintf(buf, "after#%d", afterPtr->id); Tcl_AppendElement(interp, buf); } } return TCL_OK; } if (objc != 3) { Tcl_WrongNumArgs(interp, 2, objv, "?id?"); return TCL_ERROR; } afterPtr = GetAfterEvent(assocPtr, objv[2]); if (afterPtr == NULL) { Tcl_AppendResult(interp, "event \"", Tcl_GetString(objv[2]), "\" doesn't exist", (char *) NULL); return TCL_ERROR; } resultListPtr = Tcl_GetObjResult(interp); Tcl_ListObjAppendElement(interp, resultListPtr, afterPtr->commandPtr); Tcl_ListObjAppendElement(interp, resultListPtr, Tcl_NewStringObj( (afterPtr->token == NULL) ? "idle" : "timer", -1)); Tcl_SetObjResult(interp, resultListPtr); break; }

Here is the current Tcl code:

case AFTER_INFO: if (objc == 2) { Tcl_Obj *resultObj = Tcl_NewObj(); for (afterPtr = assocPtr->firstAfterPtr; afterPtr != NULL; afterPtr = afterPtr->nextPtr) { if (assocPtr->interp == interp) { Tcl_ListObjAppendElement(NULL, resultObj, Tcl_ObjPrintf( "after#%d", afterPtr->id)); } } Tcl_SetObjResult(interp, resultObj); return TCL_OK; } if (objc != 3) { Tcl_WrongNumArgs(interp, 2, objv, "?id?"); return TCL_ERROR; } afterPtr = GetAfterEvent(assocPtr, objv[2]); if (afterPtr == NULL) { const char *eventStr = TclGetString(objv[2]); Tcl_SetObjResult(interp, Tcl_ObjPrintf( "event \"%s\" doesn't exist", eventStr)); Tcl_SetErrorCode(interp, "TCL","LOOKUP","EVENT", eventStr, NULL); return TCL_ERROR; } else { Tcl_Obj *resultListPtr = Tcl_NewObj(); Tcl_ListObjAppendElement(interp, resultListPtr, afterPtr->commandPtr); Tcl_ListObjAppendElement(interp, resultListPtr, Tcl_NewStringObj( (afterPtr->token == NULL) ? "idle" : "timer", -1)); Tcl_SetObjResult(interp, resultListPtr); } break;

objc == 2 is the difference between calling with an id and not calling with an id.

Here's the commit where the comment was added: 3c0f2c9 which changed the code:

- (script, type) = self.tk.splitlist( - self.tk.call('after', 'info', id)) + data = self.tk.call('after', 'info', id) + # In Tk 8.3, splitlist returns: (script, type) + # In Tk 8.4, splitlist may return (script, type) or (script,) + script = self.tk.splitlist(data)[0]

So, my thought is that somehow self.tk.call('after', 'info', id)) was called with an id = None, which caused the info call to not return a 2-tuple. I don't know how to look at the original issue report, so I don't know if it was able to be recreated or someone just patched the unpacking.

As I remember, tkinter now requires at least 8.4, (while IDLE requires 8.5), so 8.3 is not relevant.

Lib/tkinter/__init__.py

bedevere-bot · 2018-02-13T22:36:23Z

A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment on this pull request containing the phrase I have made the requested changes; please review again. I will then notify any core developers who have left a review that you're ready for them to take another look at this pull request.

Lib/tkinter/test/test_tkinter/test_misc.py

csabella · 2018-03-04T15:37:26Z

I pushed a rebase and got the Appveyor error. I'm working on the other concerns about the after_info functionality, so I'll push a fix once those are worked out. Although, this issue on Appveyor does answer the previous question about the after_info() event list containing other events than the ones added in this test. It seems that the other new after* tests are leaving something behind.

Lib/tkinter/__init__.py

ZackerySpytz · 2019-12-27T04:01:40Z

Misc/NEWS.d/next/Library/2018-02-13-10-02-54.bpo-32839.McbVz3.rst

@@ -0,0 +1 @@
+Add after_info to tkinter.


I think the news entry could/should be a bit more descriptive.

If this were reopened, I would be willing to expand this (after the intial test rerun finishes).

Lib/tkinter/__init__.py

terryjreedy · 2019-12-31T18:02:14Z

Cheryl, can you explain the closing? I still think adding after_info would be good.

terryjreedy · 2019-12-31T23:22:15Z

The only reported error, only on Travis, on line 191 is an extra item, 'after#8' (the 'timer' of line 180, I presume), in the tuple returned from .after_info. (The Travis job page somehow disables copying with current Firefox.) On this system, with the tcl/tk used, the preceding root.update() did not remove the 'timer' callback from the list.

I restarted the failing job to see if the failure is repeatable on Travis.

Answer: yes.

ZackerySpytz · 2020-01-04T12:32:31Z

FWIW, I also think adding after_info() would be good.

csabella · 2020-01-26T17:48:12Z

Sorry about closing this prematurely. I had thought there hadn't been much interest in adding the functionality when I originally proposed it, so I was just doing some housekeeping. Good to know there is interest. Having said that, the tests pass locally, so it may be tricky to get them to pass. I have a root.update() to clear the timer event from a previous test, so it seems to work locally but not on Travis. I'll look into it more.

serhiy-storchaka

Please expand the NEWS entry, add a What's New entry apply and test the code suggestion.

The PR looks mostly good to me and I want to land it in 3.9.

serhiy-storchaka · 2020-03-22T08:46:45Z

Lib/tkinter/__init__.py

+        function to be called by the event handler and type is either 'idle'
+        or 'timer' to indicate what kind of event handler it is.
+        """
+        return self.tk.splitlist(self.tk.call('after', 'info', id))


Suggested change

return self.tk.splitlist(self.tk.call('after', 'info', id))

result = self.tk.splitlist(self.tk.call('after', 'info', id))

if id is not None:

result = tuple(map(self.tk.splitlist, result))

return result

We need to split nested level tuples. Otherwise we sould get a tuple of strings when wantobject=0.

Please test your changes manually by temporary setting wantobject=0 in the tkinter module (we need to add a CLI test option for this, but it is a different issue).

serhiy-storchaka · 2020-03-22T08:46:45Z

Misc/NEWS.d/next/Library/2018-02-13-10-02-54.bpo-32839.McbVz3.rst

@@ -0,0 +1 @@
+Add after_info to tkinter.


github-actions · 2022-08-15T00:12:19Z

This PR is stale because it has been open for 30 days with no activity.

the-knights-who-say-ni added the CLA signed label Feb 13, 2018

bedevere-bot added the awaiting review label Feb 13, 2018

serhiy-storchaka reviewed Feb 13, 2018

View changes

terryjreedy requested changes Feb 13, 2018

View changes

bedevere-bot added awaiting changes and removed awaiting review labels Feb 13, 2018

terryjreedy reviewed Feb 14, 2018

View changes

Lib/tkinter/test/test_tkinter/test_misc.py Show resolved Hide resolved

csabella added 2 commits March 4, 2018 08:40

Add after_info to tkinter

f3598e1

Changes per review.

51bfe2b

csabella force-pushed the bpo32839 branch from 90c6408 to 51bfe2b Compare March 4, 2018 13:43

ZackerySpytz reviewed Dec 27, 2019

View changes

csabella closed this Dec 31, 2019

serhiy-storchaka reopened this Dec 31, 2019

Merge branch 'master' into bpo32839

57806b7

Cancel specific timer in test

ff9e774

serhiy-storchaka requested changes Mar 22, 2020

View changes

ezio-melotti removed the CLA signed label Jul 13, 2022

github-actions bot added the stale Stale PR or inactive for long period of time. label Aug 15, 2022

bpo-32839: Add after_info to tkinter #5664

bpo-32839: Add after_info to tkinter #5664

csabella commented Feb 13, 2018 •

edited by bedevere-bot

serhiy-storchaka Feb 13, 2018

csabella Feb 13, 2018

terryjreedy Feb 13, 2018

csabella Feb 14, 2018

terryjreedy Feb 13, 2018

serhiy-storchaka Feb 14, 2018

csabella Feb 14, 2018

terryjreedy Feb 15, 2018

bedevere-bot commented Feb 13, 2018

csabella commented Mar 4, 2018

ZackerySpytz Dec 27, 2019

terryjreedy Dec 31, 2019 •

edited

serhiy-storchaka Mar 22, 2020

terryjreedy commented Dec 31, 2019

terryjreedy commented Dec 31, 2019 •

edited

ZackerySpytz commented Jan 4, 2020

csabella commented Jan 26, 2020

serhiy-storchaka left a comment

serhiy-storchaka Mar 22, 2020

serhiy-storchaka Mar 22, 2020

github-actions bot commented Aug 15, 2022

-        return self.tk.splitlist(self.tk.call('after', 'info', id))
+        result = self.tk.splitlist(self.tk.call('after', 'info', id))
+        if id is not None:
+            result = tuple(map(self.tk.splitlist, result))
+        return result

bpo-32839: Add after_info to tkinter #5664

Are you sure you want to change the base?

bpo-32839: Add after_info to tkinter #5664

Conversation

csabella commented Feb 13, 2018 • edited by bedevere-bot

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

bedevere-bot commented Feb 13, 2018

csabella commented Mar 4, 2018

Choose a reason for hiding this comment

terryjreedy Dec 31, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

terryjreedy commented Dec 31, 2019

terryjreedy commented Dec 31, 2019 • edited

ZackerySpytz commented Jan 4, 2020

csabella commented Jan 26, 2020

serhiy-storchaka left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Aug 15, 2022

csabella commented Feb 13, 2018 •

edited by bedevere-bot

terryjreedy Dec 31, 2019 •

edited

terryjreedy commented Dec 31, 2019 •

edited