QGIS bugs and features

121 views
Skip to first unread message

Emma Hain

unread,
Jun 3, 2026, 7:02:30 PMJun 3
to QGIS Australia User Group
Hi All

As users of QGIS, if you encounter a bug, and issue or have an idea for a new feature, you can add it here:

It is set up to guide you through how to do it. 
image.png
If it is your first time and you come into any issues, let me know and I will try and help you. You will need a github account to do it and this is free.

If you are adding a new feature, can you please ping me with a link to the feature you have added. It is also good to search through the current list to see if someone also has requested that feature. If someone has, add in your request and requirements as well to that entry - the more the merrier! 

Cheers
Em

Emma Hain — Product Manager/Senior GIS Analyst
em...@north-road.com
https://north-road.com North Road
Cartography • Development • Spatial Analysis


north-road.com

FOSS4G 2026
August 30 – September 5, Hiroshima, Japan
QGIS User Conference
5-6 October, LAAX, Switzerland

Leonard Gouzin

unread,
Jun 3, 2026, 9:01:03 PMJun 3
to australian-qg...@googlegroups.com
Hi Emma,

I have added one recently about a bug loading EPT files from S3 onto QGIS 4. 


I thought it was the SRS of the input point cloud causing the issue (proj_cs vs compd_cs) but it's not. So I closed the bug later (https://github.com/qgis/QGIS/issues/66232)

I did more testing and now I think it is the EPT version 1.1.0 causing issue with point clouds that have RGB (PDRF 7). 
EPT 1.0.0 is ok with RGB in QGIS 4. 
EPT 1.1.0 with RGB doesn't load in QGIS 4. It reads the extent, but doesn't load points. 

PDRF 6 cause no issue loaded as EPT 1.1.0

Very hard to be certain that's the issue though. I'll loop you in if I add another bug.

Also, those are ept from a private bucket. QGIS3 had issue handling AWS credentials with EPT on private buckets. QGIS 4 fixed that. 

Support with bug reporting is much appreciated! 


Leo

--
You received this message because you are subscribed to the Google Groups "QGIS Australia User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to australian-qgis-use...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/australian-qgis-user-group/CACGKw0aYjp5hPD6AmGww%3DZVJTKQ3xgin8VXT_zMSo6u_DLQ76g%40mail.gmail.com.

Emma Hain

unread,
Jun 3, 2026, 11:01:01 PMJun 3
to australian-qg...@googlegroups.com
Thanks Leonard!
Did you add in another bug or was your intention to re open the issue?

Kind regards
Em


Leonard Gouzin

unread,
Jun 3, 2026, 11:15:43 PMJun 3
to australian-qg...@googlegroups.com
I want to add one once I know a bit more and have followed your guide to do it well :)

Rae Allen

unread,
Jun 4, 2026, 5:46:33 PMJun 4
to australian-qg...@googlegroups.com
Thanks Emma,

I do admit to always thinking it's my user error 🙂

Cheers

Rae

On Thu, 4 June 2026, 09:02 Emma Hain, <em...@north-road.com> wrote:

Emma Hain

unread,
Jun 7, 2026, 8:30:53 PMJun 7
to australian-qg...@googlegroups.com
Shows how smart and humble you are 🧙!
Kind regards
Em


Adam Steer

unread,
Jun 8, 2026, 4:54:06 AMJun 8
to australian-qg...@googlegroups.com
heya Emma - great idea, is there a non-github alternative for people who have fled from github?

[currently wrestling with QGIS 4.0 not rendering a new 3D map layout when one has already been made in the same project - no idea how to formally write that up though]

[also realising I need to find a new place to put a few small orphan things I still have on github. ugh]

Regards,

Adam



Brad Hards

unread,
Jun 8, 2026, 6:23:26 AMJun 8
to australian-qg...@googlegroups.com, Adam Steer

On Monday, 8 June 2026 6:53:52 PM AEST Adam Steer wrote:

> [also realising I need to find a new place to put a few small orphan things

> I still have on github. ugh]


I've had limited experience with it so far, but Codeberg

(https://codeberg.org/)  seems a valid alternative if you're looking for somewhere that looks different to github / gitlab and you don't want to self-host.


Brad

Emma Hain

unread,
Jun 8, 2026, 9:17:01 AMJun 8
to australian-qg...@googlegroups.com
Unfortunately no - times are changing but I think the project needs to use what they have. But interesting point Adam. 

Leonard Gouzin

unread,
Oct 11, 2026, 7:54:55 PMOct 11
to australian-qg...@googlegroups.com
Hi Emma,

I have done a bit more work on this bug and found the root cause! 
It's a problem due to how LAS 1.2 vs LAS 1.4 differ in how they store the point count.

I entered a bug


As LLMs do open a few doors, I experimented with doing a custom build over the weekend, and after many hours of head scratching and waiting, I did my first QGIS custom build and tested my one line fix in qgslazdecoder.cpp. 
It worked! 

I also got a bit too enthusiastic and submitted a PR, but also closed it because I realised afterwards that LLM usage was not allowed for PR, which is fair enough. 

Not sure what to do next. I thought I'd message here and let you know so the potential fix doesn't get lost. 

What would you recommend for the next steps?

Thanks in advance,
Leo



On Thu, 4 June 2026, 13:01 Emma Hain, <em...@north-road.com> wrote:

Adam Steer

unread,
Oct 11, 2026, 9:55:45 PMOct 11
to australian-qg...@googlegroups.com
hey Leonard, all

[digression into las standards]

...well, the fix for las/z 1.2 has been well known for a while: store las/z data in the OGC community standard las/z 1.4 ;)

...and there are a lot of people already who can handle this, no LLM needed.

(yes,I know the arguments around formats and provenance etc and they are all invalid for a lossless translation)

Is it really a bug for QGIS to expect a modern, accepted format? The long term fix in this case would be to reuqest that the vendor update their data store ;). At least, thats what I'd do...

Cheers,

Adam
--
Dr Adam Steer

Leonard Gouzin

unread,
Oct 11, 2026, 10:46:35 PMOct 11
to australian-qg...@googlegroups.com
Hey Adam,

I agree with you, a good LAS produced in 2026 is a LAS/Z 1.4 :)
The problem here is the other way around, only EPT 1.2 works, it's when they are 'EPT 1.4' that QGIS still prevent the display of points.

So the bug is that QGIS doesn't read EPT 1.4. the one line fix in my bug gets the point count correctly and therefore enable the display of points.

Leo
 

--
You received this message because you are subscribed to the Google Groups "QGIS Australia User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to australian-qgis-use...@googlegroups.com.

Adam Steer

unread,
Oct 12, 2026, 7:25:11 AMOct 12
to australian-qg...@googlegroups.com
hey Leonard - arguably a good las/z produced in 2016 was las/z 1.4 ;)

All my EPT resources are encoded as laz 1.4, some as EPT with version 1.1.0 listed in ept.json, others 1.1.0 - all with las/z 1.4 headers (unsigned long - 64 bit integer - point counts)

...although none are point format 7.

Do the data parse correctly with other methods (eg PDAL info? Read with LASpy? Look at a single EPT tile in CloudCompare?)  Is it just one problem dataset / data source? or many? From your report it looks like just one dataset?

We learned in a project a decade ago that las/z datasets even from the same vendors are not very standard, regardless of the commercial or scientific reputation of the supplier. The lesson is that data source issues are ore common than imagined. More so than read/write software which uses standards as reference on how to do things. And yes, I am well aware of the challenges in getting an institution to adjust how their data are stored.

Although, well, bugs creep in. However, given that the issue seems inconsistent across users, maybe it is worth looking at the data again?

Regards, and OK, maybe back to QGIS bug reporting discussion. Feel free to contact me directly about lidar things, however come with a purchase order or hiring contract ready ;)

Adam

--
Dr. Adam Steer

Leonard Gouzin

unread,
Oct 15, 2026, 2:45:33 AMOct 15
to australian-qg...@googlegroups.com
Hi Adam, 

Yes, LAZ 1.4 EPT 1.1.0, tiled with Entwine 3.2.1, pdrf 6.

It is not affecting just one dataset, far from it. It's been an issue since having the EPT with LAZ 1.4, when it is LAZ 1.2 inside, QGIS 4 has no problems with it. 

Interestingly, the LAZ 1.4 EPT tiles do not load in QGIS 3 LTR (Extra byte specification exceeds point length beyond base format length), but do load in QGIS 4.2.3 and CloudCompare 2.14beta. 

The LAZ themselves are fine, it's QGIS who doesn't seem to read 'Extended number of point records' as it should for a LAZ 1.4, and instead reads 'number of point records', which is 0 in a LAZ 1.4

I've done the work, found the fix, made a QGIS 4 build and tested it. It works! The repo doesn't allow LLM contributions, which is fair enough, so I've withdrawn my work... 

What's good practise now? 


Thanks for your help,
Leo


--
You received this message because you are subscribed to the Google Groups "QGIS Australia User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to australian-qgis-use...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages