- ChangeLog: https://mogan.app/guide/ChangeLog_v1.2.5.html
- Install now: https://mogan.app/guide/Install.html
Mogan v1.2.5.4 LTS: May the force be with you
Your efforts for Mogan v1.2.5.4 LTS is really appreciable. However, for R users you had promised to have a plugin center with all languages, the likes of R. I started with Texmacs earlier, as being a statistician I needed such a WYSIWYG editor which also merges statistical contents alongside results of R codes. But it’s R plugin was not working. The developer Jeroen found a workaround by providing a jupyter plugin approach, but that didn’t provide the colours to the codes. Said he would be working on it. I am still waiting for it. Meanwhile, I am doing with other such systems.
I wish to know if there is any such developments made by you and your team. Can R users may expect any such things in Mogan?
Many thanks for keep this going on.
I am pleased to check your last fixes for the reason the follow.
Also with previous versions of TexMacs I always experienced an unpredictable (until now) issue relative to the cpu usage. That is, a spike, from a point on, that can’t be reversed until closing the app. Although the spike push only a single cpu core to max usage, it is enough to speed up the fan to a noisy level.
I don’t know if other have experienced this or if it is specific to my configuration … but … I see your last fix: “copy/cut of the whole graphics area when it was selected via mouse” and now I verify that the cpu spike begins from the selection of a word with the mouse! … The issue still remain but I think now it is greatly narrowed the cause (provided that it is not a specific behavior on my system).
It is often caused by fonts related routines. I’ve improved it a lot in Mogan.
I don’t know if other have experienced this or if it is specific to my configuration … but … I see your last fix: “copy/cut of the whole graphics area when it was selected via mouse” and now I verify that the cpu spike begins from the selection of a word with the mouse! … The issue still remain but I think now it is greatly narrowed the cause (provided that it is not a specific behavior on my system).
There are many reasons why TeXmacs is slow, I’m trying to improve the speed one by one. Some of the improvements are actually wrong. I’m not a super hero or talented software engineer. I’m trying my best to make it better. And let users’ feedbacks tell me that I did make it better. When it is ready, the improvements in Mogan will be ported to GNU TeXmacs.
A great clue. I have other clues on why in beamer style TeXmacs/Mogan v1.2.5.x LTS is slow. Hopefully, the performance issue will be solved in the near future.
I’m not sure what causes the spiking. But a possibility to investigate is that some scheme “delayed message” is triggered too often, or alternatively that something is making the update function in qt_gui
trigger too often. Probably it is a bug which when occurs break some assumptions in the code of these mechanisms and they start to behave erratically. Is there a way to reproduce the bug consistently?
I currently use Mogan on a physical machine with Kubuntu 22.04 and there is some possibility I could have done some messy with qt installation.
I will configure a fresh VM and try to confine this behavior.
I commit to give you some precise indications.
After searching if someone else have had a similar issue i found the right one : https://savannah.gnu.org/bugs/?62156.
I can confirm the issue also on a new VM with Debian and also that if the autosave option is enabled the general cpu usage is lower.
Clearly I have no question on enable autosave now! but i wonder if - maybe - this could bring light on other performance issue.
I’ve improved the performance of beamer style now: https://github.com/XmacsLabs/mogan/commit/27a335d8cf3ff91ca712212d86851b0004c1d525
I reproduced the bug. I will try to fix it today.
The fix is now shipped in Mogan Research v1.2.6-beta2: https://github.com/XmacsLabs/mogan/releases/tag/v1.2.6-beta2
You can help us test it. And Mogan Research v1.2.6 will be released next week.