Rich text performance issue

[cookie1909]

Has anyone encountered performance issues when using rich text with html tags? I used either QLabel or QTextArea and passed in setText function an html string, with tags like, <table>, <font>. Leaving the program running overnight, the GUI ended taking 80% of the CPU. Removing the tags fixed the issue. While there's a work around, it'd be nice to use rich text.

I'm using QtEmbedded 4.6.1

[wysota]

If you don't do anything to the text or its surroundings that requires constat re-layouting the text, then no CPU power should be used on it regardless of the content of the text.

[cookie1909]

I forgot to specify that the text is constantly updated very second to show the time. I have no performance issue when using plain text. But rich text would take up a lot of CPU. The bigger of the redrawing area is, the faster the CPU usage will increase.

[wysota]

I forgot to specify that the text is constantly updated very second to show the time. I have no performance issue when using plain text. But rich text would take up a lot of CPU. The bigger of the redrawing area is, the faster the CPU usage will increase.

That seems quite normal, doesn't it?

[cookie1909]

No, not quite

For example:
Updating a label holding a red text by using function label.setText("<font color=red>Some red text</font>"); every second takes up a lot of CPU usage.
Drawing a label using palette.setColor(label.foregroundRole(), Qt::red); every second takes up almost nothing

It would make sense to me if performance is similar either way, which is not in this case.

[wysota]

In the second case you just draw text using a red pen. In the first case you have to parse HTML every now and then and do a bunch of other stuff. However I don't think drawing a red "Some red text" label takes that much cpu power. Your real use-case is much more complex.

[cookie1909]

hmm so I thought about this again today and I'm still not convinced... if it always takes the same cpu usage all the time when parsing html then ok. But initially it'd take only 3%-4%, then overnight it'd take 70%-80%. It just seems like there's a memory leak somewhere while parsing html because with a red pen I don't have that problem.

Appreciate your help, wysota

[wysota]

hmm so I thought about this again today and I'm still not convinced... if it always takes the same cpu usage all the time when parsing html then ok. But initially it'd take only 3%-4%, then overnight it'd take 70%-80%. It just seems like there's a memory leak somewhere while parsing html because with a red pen I don't have that problem.

Appreciate your help, wysota

Please post a minimal compilable example reproducing the problem and I'll happily inspect it for you.