Hi Jaroslaw,

I was sureprised about some lost features.
1. No mapbrowser available
2. issues while storing a changed query in connection with wrong critera, which were formerly o.k.

screenshots in https://www.magentacloud.de/share/8di6z38pdo
Albrecht

awill wrote:
I was sureprised about some lost features.
1. No mapbrowser available
2. issues while storing a changed query in connection with wrong critera, which were formerly o.k.

screenshots in https://www.magentacloud.de/share/8di6z38pdo

Hi Albrecht
I am glad to see this report about Kexi 3 from you.

Kexi 3.0.0 Alpha 1 has been published on September 14 with the note "Ported all modules known from version 2.9.11 except for: map form widget, experimental scripts".

https://community.kde.org/Kexi/Releases

It's always a good idea to read these announcements. We're spending time to write them and test many scenarios and keep users informed.

Moreover, technically Kexi 3 can be used alongside of Kexi 2, and even Kexi 3.1 will be co-installable with Kexi 3.0. It took a lot of effort to make it possible. If this is not available for the used distribution, its support can be contacted with a request. Knowing you opinion about deployment of the app itself is also a valuable input.

Regarding regression in smaller feature set such as the mentioned cases for queries, could you please report them to bugs.kde.org, attach test databases with minimal data? For best results please use commands logs and screen shots in English. To force that instead of e.g. German, instead running 'kexi' from the command line you can use the 'LANG=C kexi' command.

All I can see without test databases from you, without knowing database schema is that Kexi 3 has extra careful checks for validity of SQL commands so maybe issues missed in version 2 are now catch. Maybe. If you prefer to discuss issues on this forum please start a new thread, one per issue.

In general and to be honest, we've been waiting for these reports after Alpha 3.0 and Beta 3.0. Let's fix more before 3.0.1. If we had to wait more you would probably not be able to find Kexi 3.0 packages and would not be able to test anyway. That gives a full catch-22 situation typical for the software development cycles with somewhat limited budget.

Best regards
Jarek

Hi Jaroslaw,

sure you're right with you note to 3.0. My problem is, that this new package came from my stable repos of Manjaro. So I didn't see the hint.
Now I looked for a possibility to regress. All I found in Archlinux-mirrors was kexi 3. At first I will search and install another Linux in a VM with Kexi 2, because I need just now my queries.
If I get any newer version during a rolling release-update I will test it and write anyhow.
Albrecht

Yes that's a problem, maybe not your but general problem of distribution. Requiring software projects to keep quality of N deployment methods does not lead to best results. Also co-installability of version 2 and other is a feature that distributors may or may not agree with.

As mentioned not once, ultimate solution will be to distribute universal installers. I hope users will support us with this effort in any possible way.