Do you want to use CSD? No please!!

geekland · 2020-01-17 07:30:12

I've been using XFCE for 10 years and I've been really happy till today.

After realizing that you want to use CSD I really believe that 4.14 will be the last version of XFCE that I will use. I really don't understand why you need to start using CSD. From the user point of view in my opinion doesn't have any advantage.

1- Break the looking consistency between apps. Some apps will look different than others.
2- Maybe in some cases helps to save desktop screen. But in some other it's opposite. If you really want to optimize the space that apps use in the screen just make a good global menu.
3- The thick bar just looks terryble horrible. And Gnome looks horrible too. Why to do the same things that made people run away from gnome?

I hope you forget about the idea you have or at least provide a good fallback mode for the people that hate CSD. There is already the software gtk3-nocsd but final result is not satisfactory as you see traditicional title bar window + the ugly CSD together at the same time....

In my opinion Focus on what gnome is doing is not a good idea. I don't like their concept, I don't like their looking and I don't like the performance of their deskop. Follow their steps will kill XFCE. Is better you focus with Plasma Desktop.

alcornoqui · 2020-01-17 09:39:49

Welcome to the forum, or are you on your way out already? wink

I don't like it much either, but I use gtk3-nocsd on recent Xfce and I'm quite happy, no plans of changing (though I've also heard that Plasma is very lightweight nowadays).

Does people here use gtk3-nocsd or is going along with CSD?

Aravisian · 2020-01-17 10:01:44

geekland wrote:

I've been using XFCE for 10 years and I've been really happy till today.
After realizing that you want to use CSD I really believe that 4.14 will be the last version of XFCE that I will use. I really don't understand why you need to start using CSD. From the user point of view in my opinion doesn't have any advantage.
1- Break the looking consistency between apps. Some apps will look different than others.
2- Maybe in some cases helps to save desktop screen. But in some other it's opposite. If you really want to optimize the space that apps use in the screen just make a good global menu.
3- The thick bar just looks terryble horrible. And Gnome looks horrible too. Why to do the same things that made people run away from gnome?
I hope you forget about the idea you have or at least provide a good fallback mode for the people that hate CSD. There is already the software gtk3-nocsd but final result is not satisfactory as you see traditicional title bar window + the ugly CSD together at the same time....

I posted a rant on this, as well:
https://forum.xfce.org/viewtopic.php?pid=56135#p56135
To me, what I am seeing is utterly Unbelievable.
And I agree with you 100%: There is NO advantage for the user for CSD. The only advantage is for the developers and the apologetics in Magazine articles trying to convince us that it is actually good for us are as convincing as Adolph Hitlers compassionate side.
I am still using XFCE 4.12. I have used 4.14, quite a bit since it came out. And it just doesn't work right.
In 4.12, images and styling is Crisp and Clear. Modifications and theming are straightforward. It is true that some things, .css does better. But in the images, no. The images are sloppy and blurry.
But I also felt that 4.14 was a slippery slope. It was a method of getting the users to begrudgingly go along with switching to a Gnome-like desktop. 4.16 suggests that my fears were well founded.

geekland wrote:

The thick bar just looks terryble horrible. And Gnome looks horrible too. Why to do the same things that made people run away from gnome?

Nailed It.
The thick header bar looks awful, is hard to theme, serves no function and wastes space. Completely wastes space. But the thing is...
On most peoples screens, the difference between the thick Gnome Headerbar and a standard titlebar is... Millimeters.
Seriously.
Look, I don't like that big clunky header bar for 4.16 for many reasons but screen space does not concern me. The window I am working on usually takes up 80% - 99% of the screen. Three millimeters--- if that is bothering people like Gnome claims it does, you couldn't get a needle up their anus with a sledge hammer.
So this whole argument about "Precious screen Real Estate" has always been complete BUNK anyway. It's funny how anyone in 2009 ever got any work done on their tiny clogged screens, right?
Think about it- for all these years, the end user has been using Conky and Docks and Taskbars and menus and activities overview... Yet, suddenly, three Millimeters of space on a File Manager window is intolerable corruption of Precious Screen Real Estate? Bumpkus.Total hogwash. Whoever believed that claim?
And the CSD solution to this supposed problem: Make it bigger! Brilliant! Anyone smell the rat?

geekland wrote:

In my opinion Focus on what gnome is doing is not a good idea. I don't like their concept, I don't like their looking and I don't like the performance of their deskop. Follow their steps will kill XFCE. Is better you focus with Plasma Desktop.

Your opinion is a clear and educated one.
This is not an issue about what the Users want, I can assure you.
And why follow and mimic Gnome? When XFCE was a course to get AWAY from Gnome?
One thing about helping others on the forums (For me, more on the Zorin Forum than here), I can tell you Gnome Desktop is riddled with Bugs. Big Bugs, too, not minor ones. Net connection issues. And minor ones, too- Like the update manager claiming to fail to connect even when the update went through causing lots of confusion. Gnome is Bug Central. It is slow, clunky and ugly.
That XFCE and Mate both seem to have this strange and senseless desire to become like Gnome is Very suggestive. Especially with Gnome trying so hard to be like Microsoft.
What they want is to show us that they are in control.
And it is up to us to remind them that they are not.
They are not the boss- We are.
They do not tell us what we want. We tell them what we want.
Not with threats or by going to H.R. and filling out a pink slip---- But simply by whether we like or dislike their product.

EDIT:

alcornoqui wrote:

I don't like it much either, but I use gtk3-nocsd on recent Xfce and I'm quite happy

THANKS!!!
I never heard of this...
I installed the Debian pacakge, dependencies and after a couple of small tweaks to the theme I made, it is Rockin'! (The result was bumping the headerbar buttons down but used the same background- So I switched the Titlebar background in my theme out with the Toolbar background. I saved the originals as backup.)
Thanks for teaching me (and other readers) that this exists.

It's only a matter of time before Gnome writes a patch to break it.
I've got package holds set using

apt-mark <PACKAGE NAME>

(which you had also taught me some time back) and all automatic Updates turned off.

And... Geekland- Thank you for joining and Standing Up to Be Heard.

Last edited by Aravisian (2020-01-17 10:40:00)

ToZ · 2020-01-17 12:03:04

Moderator note: This is going to be a controversial topic. Lets remember to be respectful.

Here is the original roadmap design plan:
- https://wiki.xfce.org/releng/4.16/roadm … ral_ui/csd

Here are a couple of mailing list discussions about this topic:
- https://mail.xfce.org/pipermail/xfce/20 … 36684.html
- https://mail.xfce.org/pipermail/xfce/20 … 36689.html

Here is a current bug report to make CSD optional:
- https://bugzilla.xfce.org/show_bug.cgi?id=16375

Providing rational, constructive feedback to the developers may help change this plan, if enough people do so. The mailing lists or bug reports are good ways of getting their attention.

geekland · 2020-01-17 15:59:40

The aim is not be rude. I just pretend to expose my opinion. Then the developers of course can do what they want. And of course if I don't like the changes I also can change the desktop or use older versions of XFCE.

If the CSD goes ahead at least do in a way everyone is happy. A good example to follow is Catfish. Despite Catfish doesn't have menu and this point I don't like it... at least let you chose between traditional tittle bar or CSD.

In my opinion most of the XFCE users doesn't like the CSD implemented like Gnome. But it's just my opinion... If decision you took is correct or not you will see it quick. If decision is wrong users will run away like when gnome changed from 2.0 to 3.0. And In my case will be the one that runaway. For instance I don't use Evince because only offers you CSD option. Solution is Use Atril. If XFCE uses CSD then solution is move to older version or use Plasma.

Thks

Aravisian · 2020-01-17 19:49:25

ToZ wrote:

Moderator note: This is going to be a controversial topic. Lets remember to be respectful.
Here is the original roadmap design plan:
- https://wiki.xfce.org/releng/4.16/roadm … ral_ui/csd
Here are a couple of mailing list discussions about this topic:
- https://mail.xfce.org/pipermail/xfce/20 … 36684.html
- https://mail.xfce.org/pipermail/xfce/20 … 36689.html
Here is a current bug report to make CSD optional:
- https://bugzilla.xfce.org/show_bug.cgi?id=16375
Providing rational, constructive feedback to the developers may help change this plan, if enough people do so. The mailing lists or bug reports are good ways of getting their attention.

I understand your point of view. But let me be very blunt (As I often am):
Polite and constructive feedback is great when someone asks you for their opinion on something or you notice something that could stand improvement.
I am often polite and apparently rational, as 99% of my posts reflect.
There are some ideas that as presented, demonstrate that the idea is to take advantage of people, provide subpar service or institute changes for the worse. These ideas more often deserve ridicule and derision. Politeness can be confused with complacency or weakness. It encourages the presenter that they can get away with pushing what they want on the user.
When confronting an irrational idea, it suggests that a rational rebuttal will reach deaf ears.
XFCE 4.16 is not a rational construction.

It is unthinkable that the developers of XFCE, given its current structure, are unable to fathom the nature of its differences with Gnome and the wishes of its user base.
It is unthinkable that the plan for XFCE 4.16 was made rationally and constructively and it is unthinkable that the report to its disclosure could be soft-spoken. Let's not act like there is something wrong with Speaking UP.

Last edited by Aravisian (2020-01-17 19:50:56)

ToZ · 2020-01-17 21:19:58

Aravisian wrote:

Let's not act like there is something wrong with Speaking UP.

I'm not saying that at all. By all means, speak up, make your thoughts known (above, I linked to sources where one might be able to get the ear of the developers easier than at these forums). We are the users, we have skin in the game. My comments really are around being thoughtful, respectful and constructive about what we say when controversial topics are discussed. Otherwise our opinions may be dismissed.

There are two sides to every story and we need to better understand the motivations of the core developers. They are not bad people. This current group of developers have made great strides in breathing life back into Xfce. Lets hear them out. And where we disagree, lets do so in a way that encourages positive discourse and the exchange of ideas and opinions.

Just my 2 cents.

Aravisian · 2020-01-17 22:33:30

ToZ wrote:

Aravisian wrote:
Let's not act like there is something wrong with Speaking UP.
I'm not saying that at all. By all means, speak up, make your thoughts known (above, I linked to sources where one might be able to get the ear of the developers easier than at these forums). We are the users, we have skin in the game. My comments really are around being thoughtful, respectful and constructive about what we say when controversial topics are discussed. Otherwise our opinions may be dismissed.
There are two sides to every story and we need to better understand the motivations of the core developers. They are not bad people. This current group of developers have made great strides in breathing life back into Xfce. Lets hear them out. And where we disagree, lets do so in a way that encourages positive discourse and the exchange of ideas and opinions.
Just my 2 cents.

Toz,psychology is not a topic on this particular forum, but it is relevant, all the same.
While this forum is not the "ear of the developers," it is the ear of the users. And the more users made aware of the topic and of the sides, the more of them that may use the links you provided above.
I can be thoughtful. I can be constructive.
But respectful? No. That must be earned. XFCE up until 4.12 Earned respect.
However, just because you paid your rent last month does not mean that you can forgo paying this months rent since you paid it last month.

The current description of XFCE 4.16 not only does not deserve respect, it loses respect. It shows a trend of development that goes a bit beyond merely assuming the motives of the developers and this is not easily dismissed with a hand wave.
Toz, the respect I have toward you is current. I respect you and I respect what you do and I respect the help you give to many on the forum. But I can respectfully disagree with you as to how I engage others.
Sometimes, being polite, calm and smooth-talking works. And sometimes, a hard-headed person that Wants What They Want, users opinions notwithstanding will hear nothing short of Harsh, Sharp, Critical Judgment. Nothing else will carry enough weight. Positive discourse with the exchange of ideas is a Great Thing in Science. But when dealing with Stubborn, it results in them telling you all about how you are going to accept what they want to do. And the article I linked to in the thread I also started on this topic Shows That In Action Clearly.

It may not be your preferred way but... Remember that nothing is ever always best done One Persons Way.
I say hammer them.
Just my two cents.

ozjd · 2020-01-17 23:19:54

ToZ wrote:

Providing rational, constructive feedback to the developers may help change this plan, if enough people do so.

Seems that is unlikely given the response on the release mailing list to one of the developers who questioned the change. That may not have been the best place to raise it but the message was that it is happening.

Aravisian · 2020-01-18 00:37:36

ozjd wrote:

ToZ wrote:
Providing rational, constructive feedback to the developers may help change this plan, if enough people do so.
Seems that is unlikely given the response on the release mailing list to one of the developers who questioned the change. That may not have been the best place to raise it but the message was that it is happening.

God, I hate being right...

https://mail.xfce.org/pipermail/xfce/20 … 36689.html

We knew this would cause some stir(snip)

Translation: We knew you, the users, won't like what we are doing...

As for why GTK CSD are beneficial...

Translation: But we are going to do it, anyway. Like it. Like it hard.

Let's read:

1. They are natively supported by the toolkit. Any GTK application can
take advantage of them, and more will over time. This part of the code
is maintained by GNOME and will be supported and less prone to breakage.

While kind of true- it is also not necessary. This is simply a way of saying, "We want to dump responsibility for the software onto someone else."

It also means scaling just works.

This statement is simply misleading. Scaling currently works and even if improvement is viable, it is not necessary by any means to scrap the current configuration on favor of something the users do not want.

3. Consistency is *increased* since the decoration themes will always
follow the GTK theme.

What themes? Has anyone noticed how Gnome completely cracked themes into being Gnome Clones that just change colors?
It's all about the Gnome Brand Image.

This also opens the door for scenario-based
theming, such as dark mode or dark applications (think Parole as an
overlay / picture-in-picture video player). I've honestly been holding
back on this feature because of the expected backlash.

It shuts more doors than it opens and very few people like overlay. Instructions for turning the thing off is rampant on the internet. Mine is turned off and has been from the beginning.
When they- AGAIN- admit that they expect backlash, then they are admitting that they are forcing their wants on the users.

4. Consistency is increased for GTK applications. The fact is, there are
already CSD-based GTK applications, and there are applications that have
a fallback appearance for Xfce, etc. This increases the effort for
developers that want to support newer DEs and Xfce, and we're honestly a
much smaller user group.

Consistency is already present. XFWM4 is present- they plan to Drop it, not plan to increase consistency. Instead of Consistency, did he mean Consistently conforming to Gnome 3?
And consistency with Gnome is what we DO NOT WANT.
I mean, why bother with having XFCE at all if their plan is to just become Gnome? And we already know Gnomes plan.
Stick to Gnomes controlled Brand Image.

As for some of the concerns...

Huh boy...

1. Yes, *some* of our users are strongly against CSD layouts.

MOST Sean, Most. Try to be honest, ok?

This is definitely known.

Translation: But we don't care.

At the same time, this is not *all*

Translation: Screw' em.

and probably not even *most* of our users. In general, folks just want applications that
work and don't care about the window decorations.

Wrong. Did I not just suggest you be honest? "Probably?:" You don't know? Where does this assumption come from... "Probably...Ah, we just don't CARE. Who knows how many. Who cares."
This claim is demonstrably wrong given the activity and demand on themes. On DeviantArt, Openbox and the rest. Such an easily observed falsehood in his denial is Absurd. Reminder: Success of Linux Mint largely rides on User Controlled Window Management that it still offers through Marco and through XFWM.
If Sean admits, repeatedly, that he expected backlash, expect4ed it would cause a stir- why is he then claiming that MOST users are not going to be upset? Which is it? Speaking of Consistency....

Personally, I've only ever accidentally rolled my
windows up, then rolled them back down and disabled the setting.

Well, your argument above is that since you are "not all" or "maybe even not most" your own personal experiences don't mean jack. By your own argument above.

Can we fix or improve on some CSD pain points? Certainly. However, if we
continue pushing back against them fundamentally, we're never going to
and nobody else is either. Instead of outright rejecting the client-side
decorations, let's figure out how to effectively utilize them or fix the
remaining issues around them.

Translation: "We should adopt and join Gnome Standard and try to improve it instead of enjoying the improvements we already have and have been using."
Of course, as already pointed out-- This Claim is Very Misleading. It implies that CSD are "progress" and a step forward.
They are not. They are merely one way of doing something and it is a way that can be done better by using a different way. Again, Microsoft Windows has been using the same Base Window Management system for over a decade, with many improvements over the years, but not a replacement with something ugly that does not work well and never will. So there is your evidence that the claim that it Must be a Progressive Upgrade is complete and total Bunk.

What it IS, however, is the integration of all desktops into ONE desktop plan. It discourages independent development and creativity and is counter to what FOSS is all about.

Speaking of Microsoft... anyone notice how they own github, lately? You know... Just a random thought...

Here's to more constructive communication :-)

Admit it: That is not what you want. You want us to accept your long winded speech about how what You Want For US is Good For Us.

NO.

I said above... before this... That the Stubborn won't listen. They will only push us to accept what they want for us.
This post and Seans response above Demonstrate That Without Doubt.

Last edited by Aravisian (2020-01-18 01:08:13)

alcornoqui · 2020-01-18 15:04:43

After reading the whole thread (thanks for the links!), my trust in the developers has increased.

Sean Davis wrote:

We're still usability focused, highly configurable, and user-first. I think everybody still has a sour taste from the GNOME 3 CSD implementation, where features were seemingly dropped to make headerbars work. We're doing the opposite: taking advantage of the new features and making them work for Xfce.

I think it's going to be all right.

Aravisian · 2020-01-18 20:34:19

alcornoqui wrote:

After reading the whole thread (thanks for the links!), my trust in the developers has increased.
Sean Davis wrote:
We're still usability focused, highly configurable, and user-first. I think everybody still has a sour taste from the GNOME 3 CSD implementation, where features were seemingly dropped to make headerbars work. We're doing the opposite: taking advantage of the new features and making them work for Xfce.
I think it's going to be all right.

You and I are different in this, for the reasons I spelled out, above. My trust has decreased because instead of giving solid valid reasons, the developers simply give soothing words, excuses and repeated comments about how they plan to go ahead, in spite of being fully aware of resistance. That is not how you usually build trust.
Especially when one moment, a person says they knew they would get backlash and then the next moment, claim few people are worried about it.
The above links didn't show a discussion and exchange of ideas as much as it showed "HEre's what is going to happen and why you should just accept it." It was very one sided.

denyer · 2020-01-22 23:07:44

Also sceptical and have made some meandering notes here;
https://virtualdebris.co.uk/blog/050286 … after-xfce

Most of the feedback seems to be negative. If the plan is to increase use of CSDs rather than resolve issues with Xfwm4, and involves being a desktop environment that doesn't consistently support window manager functionality I'll probably switch to one that does, personally. No sense in fighting the environment when there are some good alternatives.

Realistically things like DialogsUseHeader and gtk3-nocsd will probably keep CSD mostly at bay for a while, or that gtk-mushrooms project that also seems to address it at the toolkit level.

geekland · 2020-01-23 07:06:07

denyer wrote:

Also sceptical and have made some meandering notes here;
https://virtualdebris.co.uk/blog/050286 … after-xfce
Most of the feedback seems to be negative. If the plan is to increase use of CSDs rather than resolve issues with Xfwm4, and involves being a desktop environment that doesn't consistently support window manager functionality I'll probably switch to one that does, personally. No sense in fighting the environment when there are some good alternatives.
Realistically things like DialogsUseHeader and gtk3-nocsd will probably keep CSD mostly at bay for a while, or that gtk-mushrooms project that also seems to address it at the toolkit level.

I agree with you. And for me doesn't matter if they implement it as Scenario 1, 2 or 3. in all the scenarios they will make the looking not consistent between programs. I don't like to have borders and to avoid them the best solution for me is use the global menu. With the CSD the global menu is useless... because all the CSD programs tested doesn't have implemented menus to appear in the xfce4 panel.

I agree no need to argue. Seems the developers have clear they want to implement it. So wait and see. And if I don't like it there are other desktop alternatives.

Aravisian · 2020-01-23 07:18:22

denyer wrote:

Realistically things like DialogsUseHeader and gtk3-nocsd will probably keep CSD mostly at bay for a while, or that gtk-mushrooms project that also seems to address it at the toolkit level.

Correct me if my assumption is wrong, but doesn't using gtk-nocsd require a window manager to be present on the system? If it is removed...
Besides, I suspect the gtk-nocsd hack will get targeted before long to break it.

geekland wrote:

I agree no need to argue. Seems the developers have clear they want to implement it. So wait and see. And if I don't like it there are other desktop alternatives.

As some may observe, I can get very passionate. I am not a lay down kinda guy. I'm not even stand up guy.
More like hang from the chandelier challenging any opponent kinda guy.
If we do not argue, our voices go unheard. And that doesn't sit right with me.

Perhaps our voices will be raised and still go unheard due to deaf ears. then, yes... an alternative to XFCE is necessary. However, I will ask the obvious question:
What alternative?
Mate?
Cinnamon?
Admittedly, I have more respect for Cinnamon/Mate than I do Mate. I suspect Mate D.E. will probably go CSD, too.
Either way, it feels like we are being forced into a box until we give in and just take it.

I am a novice. I have been using Linux for about a year, now... give or take a few days.
In that time, I have made my own themes, ported others. I have written a couple extensions and a piece of software. I have been learning Python, and C and Java and whatever else I can scrounge bits of information up on that will help me expand. Yet, I remain a wholly ignorant novice.
Perhaps a fork of XFCE D.E. is what is to come. I cannot sit idly by hoping someone else will do it. I lack the knowledge to do it, myself. But that- at least as stated above... I have some control over.

denyer · 2020-01-23 14:18:14

geekland wrote:

With the CSD the global menu is useless... because all the CSD programs tested doesn't have implemented menus to appear in the xfce4 panel.

Interesting use case I hadn't considered, and it's doubtful has been considered in the course of Xfce's CSD initiative. That sucks, since the WM was built to be standards-compliant with that kind of flexibility in mind.

There doesn't seem to be any sense that key devs use functionally built into the window manager beyond the basic min/max/close stuff, and maybe aero snap style tiling, so they aren't bothered about (or listening to any users about) abandoning it. Which is honestly surprising; on a normal 24" monitor snapping a window to half of the screen is in no way a substitute for maximising eg a terminal window vertically. And it isn't a "power user" type function if Windows does it.

Aravisian wrote:

Correct me if my assumption is wrong, but doesn't using gtk-nocsd require a window manager to be present on the system? If it is removed...
Besides, I suspect the gtk-nocsd hack will get targeted before long to break it.

gtk-nocsd (and the gtk-classic mod which goes way beyond it) does need a window manager that handles decorations, yeah. As I understand it the WM could be developed to pass stuff it's doing at the moment to GTK, and given that two of the developers pushing CSD very hard are active with https://git.xfce.org/xfce/xfwm4/ that may happen at some point.

I doubt Xfwm4 will get theming removing and stop being able to provide SSD decorations before 2022, and there is the possibility of using another window manager.

Aravisian wrote:

it feels like we are being forced into a box

It's the Gnome way. I only found Xfce about three years ago, when Windows became too awkward to be using outside of work, and knowing that it's survived since 1996 without major changes was part of the appeal -- it's a shame to see existing functionality cut and vandalised.

It wouldn't be surprising if more desktop-type functionality comes back into Gnome, though, the project's recovered in other ways since 3 was launched. Someone will probably pick up this eventually. And the fad for tablet interfaces isn't as strong as it once was.

Personally I like retro but am not too bothered about theming, so if that goes completely from the WM and Xfce retains enough other functionality I'd consider sticking with it. My coding level's anything useful for IT audit or other tasks at hand, more scripting, database CLRs and web interfaces than desktop software (point and click dev like VSTO doesn't really count), but it's hard to see Xfce being successfully forked when it struggles for maintenance. Even that's not a given since the chances are that current devs who don't like the way things are going will just move on.

Aravisian wrote:

an alternative to XFCE is necessary. However, I will ask the obvious question:
What alternative?
Mate?
Cinnamon?

I'm planning to look at those again, and both work, but having spent a few hours with Kubuntu in a VM that's looking like the favourite. Can't really get comfortable with Dolphin or Caja, but Nemo works well in all of those desktop environments.

KDE has taken a fundamentally different approach to Wayland than Gnome has, which doesn't focus on CSD, so I reckon it's going to be the best DE to follow in order to retain functionality (at least the functionality I use; I've not tried things like panel-located application menus).

alcornoqui · 2020-01-23 17:44:52

Thanks to all for the discussion and links (denyer, interesting blog!)

Reaccion Raul · 2020-01-25 17:30:33

I guess xfwm is kind of a mess and embracing GTK3 solves this issue. We have this visual borders since forever and It seems that this 1px zone to grab the borders is almost unfixable. I hope GTK3 solves this issue since it seems is not viable through xfwm.

Aravisian · 2020-01-25 19:42:26

Reaccion Raul wrote:

I guess xfwm is kind of a mess and embracing GTK3 solves this issue. We have this visual borders since forever and It seems that this 1px zone to grab the borders is almost unfixable. I hope GTK3 solves this issue since it seems is not viable through xfwm.

It is fixable by simply using a wider bordered theme. If the theme you are using has a 1px border, you can easily change to a theme that has a larger border (mine is 9px wide.)
Or, you can use alt key plus scroll wheel to easily (and finely) resize.
You don't need CSD to fix that. With CSD borders, you cannot theme the border past 1px as you can with XFWM. They will always be 1px wide, period. (which is also ugly and removes functionality in the good ol' Gnome style.)
You can, however, if you get into the .css, change the margins to achieve that task, but the margins are not visible. I would hardly call that a fix.
You may as well say, "We can fix this by switching to KDE."
I also can increase the border size- Easily- in XFWM while maintaining only a visible 1px wide border if that is what you want. It's a very minor tweak to any existing theme.
We don't need to remove functionality, customize-ability and move to a known faulty system to achieve what We Already Have.
What you just said is like taking a new car and removing all the components that pull electrical power and downgrading the components instead of putting in a more powerful alternator. It makes No Sense.

Last edited by Aravisian (2020-01-25 19:50:19)

sixsixfive · 2020-01-26 01:27:56

Just let the developers do what they want - If you dont want CSDs use

* MATE
* UKUI
* LXQt
* KDE

you can also use plenty of window managers with Qt or non-CSD GTK apps

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

I personally switched to MATE a year ago when all this was beginning with catfish or the taskmanager.

And i must really say MATE is a lot better than I remember GNOME2(which I never liked - damn i miss KDE3), I still use thunar or pcmanfm for file management since caja is terrible slow at image previews.

But I really like that MATE disabled the overlay scrolling and the dialog headerbars by deault. Not to mention that it has a polkit client, archive manager and atril. All MATE apps have a traditional style: menubar, toolbar and normal menus with icons (not those ugly osx-like popovers).

PS: I have used gtk3-nocsd in the past but this is a dead end since GTK4 won't support modules anymore.

Last edited by sixsixfive (2020-01-26 02:06:40)

denyer · 2020-01-26 03:12:01

Yeah, assume the task manager will go full headerbar soon if it hasn't already.

Reaccion Raul wrote:

I guess xfwm is kind of a mess and embracing GTK3 solves this issue. We have this visual borders since forever and It seems that this 1px zone to grab the borders is almost unfixable. I hope GTK3 solves this issue since it seems is not viable through xfwm.

It might be viable to either borrow code from another window manager or to outright replace Xfwm4. Marco (MATE) is an example of a lightweight WM with resize areas rather than struggle-to-hit borders. And since MATE plans for Wayland are in progress, sharing might cover Xfce's longer-term aspirations in that respect as well. If it was an outright replacement it'd have similar implications for Xfwm4 theming as CSD, of course.

And it's a moot point if devs are sold on CSD/headerbar visuals and that way of picking up resize areas regardless of impact on other functionality.

I think once GtkHeaderBars are in place, the temptation to add non-window controls into those areas will be as strong as it has been with the Settings interface and the main slowing factor will be reluctance by devs of specific applications.

Aravisian wrote:

You may as well say, "We can fix this by switching to KDE."

Or Qt, anyway. Whilst at the moment most GTK-based software may not be using headerbars, the wider adoption is the less incentive there is to fully support traditional desktops in future versions of GTK. GTK apps could end up looking and behaving more consistently under KDE; https://www.omgubuntu.co.uk/2019/12/kde … integrated

Very impressed with resource usage in Plasma so far; from boot the desktop is using about 400MB of a 2GB virtual machine. Fairly straightforward to turn off anything flashy.

Aravisian · 2020-01-26 07:26:02

sixsixfive wrote:

I personally switched to MATE a year ago when all this was beginning with catfish or the taskmanager.
And i must really say MATE is a lot better than I remember GNOME2(which I never liked - damn i miss KDE3), I still use thunar or pcmanfm for file management since caja is terrible slow at image previews.
But I really like that MATE disabled the overlay scrolling and the dialog headerbars by deault. Not to mention that it has a polkit client, archive manager and atril. All MATE apps have a traditional style: menubar, toolbar and normal menus with icons (not those ugly osx-like popovers).
PS: I have used gtk3-nocsd in the past but this is a dead end since GTK4 won't support modules anymore.

You make some strong points about Mate. Personally, I have a more critical view. But I would be willing to re-examine my views.
Missing KDE3? You are not alone in that. In fact, the Trinity desktop is based on that very premise- it is a "remake" of KDE3 desktop environment. I tried it out... But again... I find KDE very difficult to work with. It is more controlled and more scattered at the same time. Rather than neat organization and structure, the components of KDE desktop (Especially Plasma) are just all over the place. Modifying KDE seems to rely heavily on the user relying heavily on the desktop to do it.

denyer wrote:

It might be viable to either borrow code from another window manager or to outright replace Xfwm4. Marco (MATE) is an example of a lightweight WM with resize areas rather than struggle-to-hit borders. And since MATE plans for Wayland are in progress, sharing might cover Xfce's longer-term aspirations in that respect as well. If it was an outright replacement it'd have similar implications for Xfwm4 theming as CSD, of course.
And it's a moot point if devs are sold on CSD/headerbar visuals and that way of picking up resize areas regardless of impact on other functionality.
I think once GtkHeaderBars are in place, the temptation to add non-window controls into those areas will be as strong as it has been with the Settings interface and the main slowing factor will be reluctance by devs of specific applications.
Aravisian wrote:
You may as well say, "We can fix this by switching to KDE."
Or Qt, anyway. Whilst at the moment most GTK-based software may not be using headerbars, the wider adoption is the less incentive there is to fully support traditional desktops in future versions of GTK. GTK apps could end up looking and behaving more consistently under KDE; https://www.omgubuntu.co.uk/2019/12/kde … integrated
Very impressed with resource usage in Plasma so far; from boot the desktop is using about 400MB of a 2GB virtual machine. Fairly straightforward to turn off anything flashy.

I downloaded the source code for XFWM4 4.14. I found within:

    {
        XResizeWindow (display_info->dpy, screen_info->shape_win, frameWidth (c), frameHeight (c));
    }

Interesting. So, XFWM4 relies on Xwindow, the same way that gtk3 or Gnome 3 does.
It is the same mechanism being used. What makes gtk3 different is the availability of margins which was not available in gtk until version 3.20.
So prior to gtk 3.20, gtk2, gtk3 and XFWM were all using the same method of resizing. HMMMMMM...
After version 3.20, margins allowed the user to easily readjust the area with which to grab by simply increasing the frame margins in the .css file of the theme in use (to say, 10px). Which, anyone can verify with a simple internet search.
What this clearly shows, is that the same is true for gtk2 and gtk3 <3.20.
And for XFWM. The fix was and still is the same for all of these window management systems: Define the border width.
Which I had already pointed out above.

But what interests me the most is hearing the repeated claim implying that XFWM is somehow special- somehow DIFFERENT in a resizing problem that just cannot be solved. Apparently, at least according to the source code- This is Not Possible. Because all of the above reply on the same system. What troubled one would HAVE to trouble the other... UNLESS certain XFCE4 users were consistently using a Thin Border theme on XFCE that they did NOT use on other desktops, giving the illusion of a special problem.

With the horrible Wayland being crammed down our throats... I wonder what the result of that will be. Personally, I find Wayland to have some good points and some pros- but overall to be a dead end in development. Yet Gnome seems very (VERY) insistent that everyone conform to it. I wonder why.

denyer · 2020-01-26 12:51:17

Oh, it doesn't seem to be a special problem at all. It's been a design decision;

https://bugzilla.xfce.org/show_bug.cgi?id=16375
https://bugzilla.xfce.org/show_bug.cgi?id=11808
https://bugzilla.xfce.org/show_bug.cgi?id=9490

Olivier doesn't want to make providing an increased resize area contingent on the compositor.

In MATE, Marco takes this approach as a WM; without its adaptive compositor windows have small theme-based resize areas, and with it there are large convenient resize areas.

In Xfce, rather than make it so so that users have the option of using a compositor and having a workable resize area independent of theme, nobody gets one unless they pick a suitable theme. Or unless they opt to lose other window manager functionality and use CSD apps. Or replace the window manager.

This is particularly a problem with small development projects; decisions can end up being taken for ideological or vanity reasons rather than with input from across the user base and with regard to the impact on accessibility.

I wasn't just being bullish by describing the approach as vandalism. It's been deliberate and avoidable.

KDE development can be a bit chaotic from the outside and embracing of rapid change, but it's hard to see similar design decisions being taken there -- the active community is much larger and more vocal, as is the development resource available. The project gets a lot of scrutiny from stakeholders.

Apart from being a bit of a PITA to configure, if using (Ubuntu) MATE I'd be more concerned at the encroachment of headerbar apps and the GTK3/GTK4 transition -- that's with the caveat of not having tried a stock MATE install, though. The development path of GTK is still of concern for any software using it, but I think if anything can shield users reasonably effectively it's probably KDE.

Aravisian wrote:

Rather than neat organization and structure, the components of KDE desktop (Especially Plasma) are just all over the place. Modifying KDE seems to rely heavily on the user relying heavily on the desktop to do it.

It's certainly easier if you can make a comfortable environment without too much tweaking, yeah, and doing so isn't always intuitive with detailed settings being a bit hidden away. Fortunately I like a fair amount of the default stuff, just with bling (animations, shadows, etc) dialled back. It is noticeable that even with a fairly developed icon set such as Elementary there'll be a few missing icons in parts of the Plasma system settings, as an example.

Last edited by denyer (2020-01-26 13:00:01)

Aravisian · 2020-01-26 20:30:49

denyer wrote:

In Xfce, rather than make it so so that users have the option of using a compositor and having a workable resize area independent of theme, nobody gets one unless they pick a suitable theme.

My point is that the solution to this is very simple.
As they all rely on the same method, defining the border size is all that is needed to solve the problem- Which I am sure we both just agreed on.
So, the problem may well be that themers, (Perhaps unaware) took a simplistic approach when making an XFWM4 border theme and when creating a 1px wide side border, inserted the logical 1px wide image.
When I first started using XFCE4 desktop, I noticed that problem right away. So, on my themes, I went into the XFWM4 folder and adjusted those images accordingly, giving them more space (Usually 9px) while leaving the Visible portion at 1px. Problem solved. Which is exactly the same thing as what CSD using a margin does. Similar solutions to a simple problem. And this is a problem easily solved (or prevented!) by the theme maker, as long as they know to do it. And knowing to do it is par for the course.
I recently taught myself how to make Metacity themes (for Mate and Cin) and that is simply part of theme-making. Knowing that there are things you must do to make it work. Even if it is a bit of trial and error.
In such a case, yes you could say it was a poor design decision on the maker of the Window Manager. But given such a very simple solution- I cannot support the idea that the decision is really all that poor. Where trouble hit was when many people did not know how it worked and did not know how to make the solution.
If, long ago, it was more readily known- then we would not be discussing it now... But this is also why I try hard to encourage others to do a little digging and make their own themes.

It is interesting... In making Metacity Themes, in making XFWM4 themes- how much control you have and how very easy it is to make a custom theme exactly to your liking. But in CSD, there is much less of that functioanlity. You must force fit, only the headerbar can be themed- the rest of the window is not, the borders cannot be aligned (the alignment changes per application) and you cannot create a consistent border all the way around. Images in CSD are lesser quality- where in the XFWM and Marco or Metacity, crisp and clear and stretch properly- in CSD, this is not the case. The images are fuzzier and stretch poorly.

denyer wrote:

This is particularly a problem with small development projects; decisions can end up being taken for ideological or vanity reasons rather than with input from across the user base and with regard to the impact on accessibility.

True. You have a knack for concisely making a strong point. I thoroughly enjoyed reading the Blog link posted earlier.

denyer wrote:

It's certainly easier if you can make a comfortable environment without too much tweaking, yeah, and doing so isn't always intuitive with detailed settings being a bit hidden away. Fortunately I like a fair amount of the default stuff, just with bling (animations, shadows, etc) dialled back. It is noticeable that even with a fairly developed icon set such as Elementary there'll be a few missing icons in parts of the Plasma system settings, as an example.

I have had a Painful time trying to learn QT/KDE.
Not just a bit lacking in intuitiveness... It's bass ackwards. I would prefer to learn it and be good at it but frankly, getting started on it usually results in me wiping all traces of it off my machine to my own great relief.
I've not done too horribly on my own but for QT, I may need a tutor that can reign in my temper.

Last edited by Aravisian (2020-01-26 20:37:04)

Misko_2083 · 2020-01-26 22:11:42

I've heard that in mate they use invisible resize window borders.
https://github.com/mate-desktop/marco/issues/248
https://github.com/mate-desktop/marco/pull/503
They backported something from Metacity.

Xfce Forum

Sub domains

#1 2020-01-17 07:30:12

Do you want to use CSD? No please!!

#2 2020-01-17 09:39:49

Re: Do you want to use CSD? No please!!

#3 2020-01-17 10:01:44

Re: Do you want to use CSD? No please!!

#4 2020-01-17 12:03:04

Re: Do you want to use CSD? No please!!

#5 2020-01-17 15:59:40

Re: Do you want to use CSD? No please!!

#6 2020-01-17 19:49:25

Re: Do you want to use CSD? No please!!

#7 2020-01-17 21:19:58

Re: Do you want to use CSD? No please!!

#8 2020-01-17 22:33:30

Re: Do you want to use CSD? No please!!

#9 2020-01-17 23:19:54

Re: Do you want to use CSD? No please!!

#10 2020-01-18 00:37:36

Re: Do you want to use CSD? No please!!

#11 2020-01-18 15:04:43

Re: Do you want to use CSD? No please!!

#12 2020-01-18 20:34:19

Re: Do you want to use CSD? No please!!

#13 2020-01-22 23:07:44

Re: Do you want to use CSD? No please!!

#14 2020-01-23 07:06:07

Re: Do you want to use CSD? No please!!

#15 2020-01-23 07:18:22

Re: Do you want to use CSD? No please!!

#16 2020-01-23 14:18:14

Re: Do you want to use CSD? No please!!

#17 2020-01-23 17:44:52

Re: Do you want to use CSD? No please!!

#18 2020-01-25 17:30:33

Re: Do you want to use CSD? No please!!

#19 2020-01-25 19:42:26

Re: Do you want to use CSD? No please!!

#20 2020-01-26 01:27:56

Re: Do you want to use CSD? No please!!

#21 2020-01-26 03:12:01

Re: Do you want to use CSD? No please!!

#22 2020-01-26 07:26:02

Re: Do you want to use CSD? No please!!

#23 2020-01-26 12:51:17

Re: Do you want to use CSD? No please!!

#24 2020-01-26 20:30:49

Re: Do you want to use CSD? No please!!

#25 2020-01-26 22:11:42

Re: Do you want to use CSD? No please!!

Board footer