- User Since
- Mar 20 2016, 11:37 PM (49 w, 1 d)
My big apologies, Wladimir, you were right that kdesu could not be found on my Arch linux because of the path differences You and Alex talked about. I've tested this patch on useractions instead of invoking root-mode, that's why it previously worked for me, stupid of me...
For example me and a few friends of mine use it to keep some tabs in their locked path. When you select such locked tab and start browsing in it, you expect the tab to be duplicated and preserve the original selected tab. This is of course similar to selecting a tab, duplicating it and then browse. But with locked tab you can be sure you don't loose the tabs location. Of course it would be nice to have a visual difference between these two tab states (total commander appends * to such tab name) and a confirmation dialog whether I really want to close a locked tab. But I believe current state is good enough.
So far I cannot find any regression. And since I'm many-tab-user I can really feel the speedup of Krusader startup, awesome, thanks Alex! :)
This works as expected. Thanks!
Perfect, Alex, thanks! :)
Oh, good point, Toni! Thanks for making it through the long comments. I believe this will get eventually fixed in Kate but lets use kwrite again. At least until it is fixed.
I agree with Wladimir and Alex on this. I'm also fond of the third option: making root-mode unnecessary by providing support for elevating privileges for specific operations operations.
Fri, Feb 24
Perfect! Although it would be nice (in future) to make sure this askpass program is installed (e.g. in INSTALL file?). I'll try to integrate the internal editor support for sudoedit and at the same time add the dependency.
Wed, Feb 22
I agree, security is now better with this patch. It works on my Arch Linux. Thanks for taking care of this!
Tue, Feb 21
Thanks! I didn't test it in Kubuntu but it still works fine on Arch.
This is perfect, thanks!
Tue, Feb 14
Yes, the new behaviour makes definitely more sense. Thanks, Wladimir! :)
Working very nicely, thanks! :-)
That's a lot of work, thanks Alex! So far it is working nicely, I couldn't find any regressions.
Fri, Feb 10
Yes, that must have been it. It now works nicely, thanks for the patch! :)
Thu, Feb 9
Alex have good points as well as You Wladimir about the simplicity of the previous code. I also agree now with the previous approach and sorry about the previous suggestion then.
Wed, Feb 8
I find these arguments good enough to set the default to case insensitive, like You proposed :).
Sat, Feb 4
The arguments about default case insensitivity in editors are convincing. But I'm heavy user of this search dialog and like the default case-sensitive setting which is on par with grep (like Alex just said). Maybe the best solution would be to make the setting persistent over Krusader restarts?
I'd also say I don't like it that much. And maybe we don't want to duplicate every setting but just some of it - e.g. the one regarding KrView? My proposal is to create alternative KrView::restoreSettings method which would take another KrView reference as a parameter and load the settings directly from this given view. Any other suggestions? :)
Agreed. Thanks a lot for your work! :)
Sun, Jan 29
About that F11 subject: Is it a change in the user interface of Krusader? :-?
I was also wondering that for a moment :-) but I believe Alex was talking about his own key bindings. It's still F8 by default in the code.
Jan 28 2017
Good improvement, thanks, Alex!
Ship it! (Sorry for my late response)
Jan 16 2017
This look nice, almost perfect :)).
Jan 15 2017
I'm definitely for this change! I don't like loosing the path when e.g. external disk is temporarily removed. Thanks!
I understand the motivation behind the additional css now, so I'm also fine with your proposed changes. We can always adjust it later if we change our minds :).
Jan 10 2017
Unfortunately I don't have this area studied enough yet to consult the direction. But I certainly can test your code and report issues:). So far I found some regarding refreshing:
- external changes aren't reflected (I have to manual refresh Ctrl+R to see them)
- undo action will not refresh view after it is done
I like the idea about the installation note! But is it necessary to pull another css? How about leaving a plain paragraph to visually align with another paragraphs under rest of headings? Something like:
I like these changes / fixes very much:). Thanks!
Jan 3 2017
Working nicely, thanks for the rework!
It works, thanks! :)
Dec 18 2016
Dec 7 2016
Very nice, thanks for taking care of the KIO patch!
I like it, thanks! The start mode enum makes the code more understandable.
Agreed. I'm also not happy about the complexity. Now I realize, that reversing the mode is not such a good idea. So I'd at least remove F2 In Parallel option - this probably won't have much uses. I'd only be happy if F2 Queue was there when queueMode is off (in matter of additional feature to the original JobMan experience).
Dec 5 2016
Actually that was intentional :). You started the jobs in queue mode, that's remembered, so they are performed in queue regardless of changed queueMode option. And it should work vice versa.
Dec 4 2016
comments dont belong into variable names :)
Agreed & fixed.
Dec 3 2016
Just added another mime (application/vnd.rar) + code readability improved.
Tested another archives like deb, rpm and rar but none of them is write-supported by 7-zip. In fact 7-zip websites claim only unpack support for the rest of proprietary archive types I didn't test.
Dec 1 2016
Changes in this diff:
- added back Start Paused checkbox
- renamed "queue" parameter to "reverseQueueMode"
- renamed "queueBox" to "pauseBox"
- added tooltips for F2 buttons
Nov 30 2016
- compatibility after merging the manuallyPaused check
- Run Immediately button textation
- bool isQueueMode = _queueMode != reverseQueueMode;
If I see this right, the behavior is exactly the same as if you would enable/disable the "queue mode" before creating the job.
Yes, this is exactly what I intended. And I agree I'd have to add a tooltip. But I don't think it is complicated. It simply allows user to selectively choose to run new job in the other mode. I'll explain the intention later.
"F2 in parallel" doesn't do anything different if no jobs are currently running
That's true. Maybe it should be renamed to "F2 Run Immediately"?
this is less ambiguous
Yes, this is better, thanks!
It is impossible to create a paused job now if no other jobs are currently running. I think this was the main point of the old queue manager.
Yes, you are right about the impossibility to create the paused job. And maybe it can have some use-cases so I should add the checkbox back?
Nov 29 2016
Great work. thanks! Also thanks for fixing the focus in search modes etc.
Nov 26 2016
Here on Arch (KF5.28) and Kubuntu 16.04 (KF5.18) the compilation now works with no warnings, thanks! Debian Stretch has KF5.22 so I suspect no problems there.
Nov 25 2016
Thanks for clarification! Now I see what the "Enqueue Operation" checkbox is really about. So every job added with the checkbox checked has to be unpaused manually to perform. The checkbox should probably be renamed to "Start Paused" but I'm unsure, too.
Nov 21 2016
Thanks for fixing the krarc protocol!
Great work, Alex. Queuing is working perfectly, refreshing is fixed, and we have a nice job overview:). Many thanks for Your work!
Nov 20 2016
people have a lot to thank to you :-)
Exactly. Thanks for doing it the right way! :)
Nov 16 2016
I politely disagree. It is true that "unmanaged" corresponds to the the current behaviour- just start all jobs right away. But during testing i often had the situation that my system slows down a lot when more than two copy operations are working in parallel. And I believe most users will expect the queue behaviour.
Nov 15 2016
Perfect! Thanks for all the fixes!
Nov 14 2016
Works well, thanks for this Alex!
Nov 12 2016
Hi Alex! Thanks a lot for your fixes and advices!
Nov 7 2016
Very good, Alex thanks!
Nov 2 2016
We better keep VFS clean.
Here is a new patch for the view :). Should address the external rename/remove. Again, don't hesitate to throw it away.
Nov 1 2016
But I really think these are problems with the view, not the VFS.
You're probably right. Here are my thoughts that led me to my "solution". I think that the selective events for file creation and deletion should actually be provided by KDirWatch. Since this is not quite possible and we have to rely only on a dirty signal, I suspected the next layer (VFS) should handle that and emit granular changes to view. So view can only redraw what really changed.
Oct 30 2016
I've corrected a crash bug in my code and created a proper revision:
So please forgot the attachment in my previous post.
OK, here is my first attempt for fixing the scroll jump and cursor move.
Oct 26 2016
Confirmed, free space info is now correctly refreshed, thanks Alex!
Oct 23 2016
Thanks a lot for making the new release a reality, Alex! You are doing a lot of work including bugzilla, so thanks again!
Oct 19 2016
I'm all for it! Will You take care of the release or should I? :)
I forgot to say that loosing current selection after rename is now truly fixed. Thanks!
Oct 18 2016
An example for a good branching model is here: http://nvie.com/posts/a-successful-git-branching-model/
Thanks for the useful link!
Oct 17 2016
It should be fixed now.
Now I can see no delays, thanks Alex!
Oct 16 2016
Oct 13 2016
You're right, now I'm now noticing that, too. It looks like it is slow even with e.g. 2 folders, so it looks like some kind of "fixed" delay when refreshing.
Oct 12 2016
Thanks a lot for this extensive work, Alex! I'll be using this branch as my default now.
Oct 10 2016
I just added:
- checks if rename command is not empty and if exit code is a success
- support for 7z archives
Thanks Davide, that (ERR_CANNOT_RENAME) works perfectly. I'll try to upload new diff with more checks.
Oct 9 2016
Closing. (Automatic close by commit message didn't work this time)