Reasons why currently I favour this over D16882:

• slightly less complex, as it does more what one expects, adding & removing actions on show & hide
• no global static

Just a few remarks on the comments that should make them easier to understand (a priori comments should illustrate code and not require lots of different code the understand their meaning ;))

I've tested this in the 5.3 branch now, which required applying hunk #3 manually to texteditor.cpp .

It seems that it achieves its primary goal (on Linux, haven't tested OS X but context menus don't use platform menu APIs so should work the same).

However, the patch has a side-effect: I'm getting messages like this each time I switch or open documents:

kdevplatform.shell: Broken text-document:  QUrl("file:///opt/local/var/tmp/kdevelop-tmp-505-%7Bf793a513-f14e-47b9-8448-06ca72900c04%7D/kdevelop_Q27955.patch")

This example is for a file that indeed doesn't exist, but that I also do not have open (so WTH?!) but I also get it for files that open perfectly.

I fail to see an immediate reason for this, so I'm back to running my own patch.

Thanks for first review & testing.

In D22424#494980, @rjvbb wrote:

It seems that it achieves its primary goal (on Linux, haven't tested OS X but context menus don't use platform menu APIs so should work the same).

Sounds good so far.

However, the patch has a side-effect: I'm getting messages like this each time I switch or open documents:

kdevplatform.shell: Broken text-document:  QUrl("file:///opt/local/var/tmp/kdevelop-tmp-505-%7Bf793a513-f14e-47b9-8448-06ca72900c04%7D/kdevelop_Q27955.patch")

This example is for a file that indeed doesn't exist, but that I also do not have open (so WTH?!) but I also get it for files that open perfectly.

I fail to see an immediate reason for this, so I'm back to running my own patch.

Interesting. Not seen here, but giving this some stress testing now. The kdevelop_xxxxx.patch is coming from the diff review system creating text document objects for displaying the diffs, but I also see no direct relation ship to this patch here. Will poke around a bit,

In D22424#496112, @kossebau wrote:

In D22424#494980, @rjvbb wrote:

However, the patch has a side-effect: I'm getting messages like this each time I switch or open documents:

kdevplatform.shell: Broken text-document:  QUrl("file:///opt/local/var/tmp/kdevelop-tmp-505-%7Bf793a513-f14e-47b9-8448-06ca72900c04%7D/kdevelop_Q27955.patch")

This example is for a file that indeed doesn't exist, but that I also do not have open (so WTH?!) but I also get it for files that open perfectly.

I fail to see an immediate reason for this, so I'm back to running my own patch.

Interesting. Not seen here, but giving this some stress testing now. The kdevelop_xxxxx.patch is coming from the diff review system creating text document objects for displaying the diffs, but I also see no direct relation ship to this patch here. Will poke around a bit,

Playing around I finally managed to reproduce this as well. Though also for 5.3 and without this patch: seems something to do with workingsets code being eager to create TextDocument objects for any urls they have in the sessionrc file, Dumped my findings as https://bugs.kde.org/show_bug.cgi?id=409858

Friedrich W. H. Kossebau wrote on 20190717::00:12:44 re: "D22424: TextDocument: remove actions from contextmenu on hide already"

This revision was not accepted when it landed; it landed in state "Needs Review".

Hmm, since when is this acceptable?

I see the same change was backported to the 5.3 branch, understandable in a way but what about the glitch I mentioned and you confirmed? Have you solved that somehow?

FWIW I got to look at the KTE implementation of the context menu mechanism that is used here. It indeed uses and reuses a single QMenu instance (there's even a comment in the code about that).

In D22424#497204, @rjvbb wrote:

Friedrich W. H. Kossebau wrote on 20190717::00:12:44 re: "D22424: TextDocument: remove actions from contextmenu on hide already"

This revision was not accepted when it landed; it landed in state "Needs Review".

Hmm, since when is this acceptable?

It is acceptable when the person doing it is semi-maintainer (by their role of release manager). The other option would have been to release 5.3.3 without a fix for ctags plugin users, rendering it unusable for people relying on packaged kdevelop. See also https://mail.kde.org/pipermail/kdevelop-devel/2019-July/063157.html

I see the same change was backported to the 5.3 branch, understandable in a way but what about the glitch I mentioned and you confirmed? Have you solved that somehow?

If by glitch you mean the "Broken text-document:" log output appearing, please see my comment D22424#496348 above, repeating here: "Playing around I finally managed to reproduce this as well. Though also for 5.3 and without this patch: seems something to do with workingsets code being eager to create TextDocument objects for any urls they have in the sessionrc file, Dumped my findings as https://bugs.kde.org/show_bug.cgi?id=409858"

But you might mean some other glitch, which I have no idea right now?

FWIW I got to look at the KTE implementation of the context menu mechanism that is used here. It indeed uses and reuses a single QMenu instance (there's even a comment in the code about that).

Please give links into the code, as I am lost what you exactly refer to here.

The other option would have been to release 5.3.3 without a fix for ctags plugin users, rendering it unusable for people relying on packaged kdevelop.

Erm, aren't you forgetting about my patch which could have been applied to the (EOL!) 5.3 branch? That does the job just as well and would have left time to find a fix for #409858 in time for the next release.

>  But you might mean some other glitch, which I have no idea right now?

No, that glitch, which for me is just as unacceptable as the glitch your patch fixes. I get it systematically.

> FWIW I got to look at the KTE implementation of the context menu mechanism that is used here. It indeed uses and reuses a single QMenu instance (there's even a comment in the code about that).

Please give links into the code, as I am lost what you exactly refer to here.

see the source of KTextEditor::ViewPrivate::contextMenu(). The ctx menu is in fact managed by KXMLGUI.

In D22424#497271, @rjvbb wrote:

The other option would have been to release 5.3.3 without a fix for ctags plugin users, rendering it unusable for people relying on packaged kdevelop.

Erm, aren't you forgetting about my patch which could have been applied to the (EOL!) 5.3 branch? That does the job just as well and would have left time to find a fix for #409858 in time for the next release.

Please read the email behind the link I gave in my last comment (https://mail.kde.org/pipermail/kdevelop-devel/2019-July/063157.html), I did not forget about it and explicitly considered it.

>  But you might mean some other glitch, which I have no idea right now?

No, that glitch, which for me is just as unacceptable as the glitch your patch fixes. I get it systematically.

As my comment should have made clear, but perhaps was not, that very glitch is unrelated to this very patch. It can be observed without the patch applied, and the origin of the logs appearing have been roughly described in the bug created and linked.

> FWIW I got to look at the KTE implementation of the context menu mechanism that is used here. It indeed uses and reuses a single QMenu instance (there's even a comment in the code about that).

Please give links into the code, as I am lost what you exactly refer to here.

see the source of KTextEditor::ViewPrivate::contextMenu(). The ctx menu is in fact managed by KXMLGUI.

The context menu is queried every time from KXMLGUI that method is called. So if KXMLGUI internally decides to recreate the context menu, you get another object the next time. Please see again the description of this very bug.

>>   > FWIW I got to look at the KTE implementation of the context menu mechanism that is used here. It indeed uses and reuses a single QMenu instance (there's even a comment in the code about that).
>>   
>>   Please give links into the code, as I am lost what you exactly refer to here.
> 
> see the source of `KTextEditor::ViewPrivate::contextMenu()`. The ctx menu is in fact managed by KXMLGUI.

The context menu is queried every time from KXMLGUI that method is called.  So  if KXMLGUI internally decides to recreate the context menu, you get another object the next time. Please see again the description of this very bug.

Since we're dealing with "please read/see again": idem for my comment :P which only states that a single QMenu instance is being (re)used. I've never seen it change but I without trying to understand the KXMLGUI internals (or docs) I will have to assume that this is possible. That's moot though, I've been only been saying that the ctx menu instance should probably be cached in a more central location.

In D22424#497316, @rjvbb wrote:

Since we're dealing with "please read/see again": idem for my comment :P which only states that a single QMenu instance is being (re)used. I've never seen it change

Please add a line

qDebug() << "Showing context menu" << menu;

to KDevelop::TextDocument::populateContextMenu.

Rebuild & install, then run kdevelop, disable the CTags plugin, have 2 text documents open:

1. invoke context menu a few times on view of document A -> same menu instance CM_A
2. switch to view of document B, invoke context menu a few times -> same menu instance CM_B (!= CM_A)
3. switch back to view of document A, invoke context menu a few times -> same menu instance CM_A2 (!=, CM_A, != CM_B)

This is behaviour I pbserved here consistently with KF 5.60.

When enabling the CTags plugin again, suddenly it is always the same menu instance.

If we consider that every time a document is switched, that this internally triggers a remerge of all the menus defined by the kxmlguiclients, due to the xmlgui menu data from the old view being removed and the xmlgui menu data from the new view being added, it is actually not that surprising (well, once one has thought a lot about it, it only appeared to myself now while writing this :P ) that the context menu object is recreated. Because without any other kxmlgui client injecting a menu of id "ktexteditor_popup" into the system (like the ctags plugin does), when a texteditor view is removed from the shell kxmlgui, nothing else demands/defines a menu with that id, so it gets deleted. And then recreated, because the new view getting active and adding its kxmlgui data has again such a menu defined with that id.

Please add a line

  qDebug() << "Showing context menu" << menu;

to KDevelop::TextDocument::populateContextMenu.

How do you think I debugged the situation? I did the same thing in Kate too.

If memory serves me well but I did not see what you are seeing but that can have been in Kate, which I tend to consider as the reference of how all things KTE should work. I recall quite clearly thinking that KXMLGUI caches UI elements for a reason and that it's not "normal" if it recreates them.
I really don't have time right now to revisit this but there is no longer a hurry. I'll activate the debug trace on change of my lastShownMenu though; the fact it's there does suggest that I too have seen menu instance changes under certain conditions.

In D22424#497361, @rjvbb wrote:

How do you think I debugged the situation? I did the same thing in Kate too.

The idea was to show you how I did my testing exactly, by which code and test steps, so you could do the very same, so we are on the same page and can compare our observations.
If you read something else into this, don't.

In any case there isn't much else you can do; stepping through the code with a debugger is near impossible (the menu about-to-be-opened will already have grabbed mouse and keyboard focus so you'd need to display remotely).

I'll activate the debug trace on change of my lastShownMenu though

I'm doing this (in my patch from D16882 which still applies after reverting D22424):

+        if (menu != lastShownMenu.data()) {
+            if (lastShownMenu) {
+                qCWarning(SHELL) << "Added items to new contextmenu" << menu;
+            }
+            lastShownMenu = menu;
+        }

under the assumption that this will print a warning (enabled in qtlogging.ini) when lastShownMenu will be updated but not the 1st time a ctx menu is opened. Is that assumption wrong, because I'm not seeing any warnings after disabling the CTags menu (and restarting KDevelop for good measure).
Wouldn't it be great if something in my patch prevented the re-creation of the ctx menu instance? O:-)

In D22424#497415, @rjvbb wrote:

I'll activate the debug trace on change of my lastShownMenu though

I'm doing this (in my patch from D16882 which still applies after reverting D22424):

+        if (menu != lastShownMenu.data()) {
+            if (lastShownMenu) {
+                qCWarning(SHELL) << "Added items to new contextmenu" << menu;
+            }
+            lastShownMenu = menu;
+        }

under the assumption that this will print a warning (enabled in qtlogging.ini) when lastShownMenu will be updated but not the 1st time a ctx menu is opened. Is that assumption wrong, because I'm not seeing any warnings after disabling the CTags menu (and restarting KDevelop for good measure).

There is another condition here: if (lastShownMenu). So if the QMenu object instance got deleted (like there is chance this happens when kxmlgui data is removed first on view switch and nothing else has id of that context menu at that very moment), this condition will not be met and no log will be created. I would propose to add debug log also for the "not lastShownMenu" case to make this visible in the log.

Wouldn't it be great if something in my patch prevented the re-creation of the ctx menu instance? O:-)

Fear "Something" not closer defined otherwise in software has a small chace of getting the "great" attribute ;)

There is another condition here: `if (lastShownMenu)`. So if the QMenu object instance got deleted

You mean I should do something like this?

+        if (menu != lastShownMenu.data()) {
+            if (lastShownMenuSet) {
+                // don't warn about setting the cache for the 1st time
+                qCWarning(SHELL) << "Added items to new contextmenu" << menu;
+            }
+            lastShownMenu = menu;
+            lastShownMenuSet = true;
+        }

Same effect for now, no debug output.

> Wouldn't it be great if something in my patch prevented the re-creation of the ctx menu instance? O:-)

Fear "Something" not closer defined otherwise in software has a small chace of getting the "great" attribute ;)

Which is why I said "it", not referring to my patch.
I find that a perfectly analysed fix rarely qualifies for great either: with hindsight most things you understand always become trivial (and the things they fix, the opposite of great) O:-)