staticlibs

> sometimes libreoffice crashes and you can't figure out why > why libreoffice stopped publishing artefacts to mvn repo

I think both questions would be a perfect fit for the paid support bugtracker of LibreOffice maintainers. Hopefully paid by some hospital funds that are not spent on MS Office licenses.

Fltkhs [1] was the best ever experience I've had while doing small/auxiliary GUI utilities. Now using NWG [2] for that (UI forms examples [3]) and it is so much less elegant and so much more verbose than Fltkhs. Resulting lean EXEs without any bundled runtime are nice though.

[1] https://github.com/deech/fltkhs

[2] https://github.com/gabdube/native-windows-gui

[3] https://github.com/wiltondb/wiltondb/wiki/WiltonDB-Data-Tran...

Chasing memory leaks in BCP [1] data load implementation in Babelfish [2]. Spotted unusually high RAM usage, decided to look deeper and got into a rabbit hole. Haven't had this fun of overnight debugging (of Postgres guts) for a long time. As a result found 4 different leaks and one (unlikely to be triggered) crash. And now have BCP data load impl with 100% constant-bounded RAM usage in a DB server process.

[1] https://learn.microsoft.com/en-us/sql/tools/bcp-utility?view... [2] https://babelfishpg.org/

Transact-SQL support (both queries and stored procedures) for PostgreSQL is implemented in Babelfish extensions [1]. It is implemented from the ground up as a "pl/tsql" language for Postgres [2], without using SQL-to-SQL transpilation like JOOQ.

[1] https://babelfishpg.org/ [2] https://github.com/babelfish-for-postgresql/babelfish_extens...

What is cool about Postgres extensions ecosystem, is that new functionality can be built upon existing extensions. For example, Babelfish (SQL Server compatibility layer for Postgres) uses pg_hint_plan to implement Transact-SQL hints support [1]. This is really useful, because existing queries, that were optimized for MSSQL, may require tuning when moved to Postgres/Babelfish, and it is great to have additional knobs for tuning.

[1] https://github.com/babelfish-for-postgresql/babelfish-for-po...

I meant that Jet Brains do care about GUI support in OpenJDK (mainstream Java), their flagship products depend on this, they employ a number of ex-Oracle GUI devs and contribute fixes upstream. Just there are no new major projects in GUI area in OpenJDK for years, because none of big contributors is interested in it. Compare this to improvements in non-GUI areas like Java 21 Threading and also the whole Graal thing.

None of major Java/OpenJDK contributors (Oracle, Red Hat, SAP etc) care about desktop GUI Java libs. Jet Brains do care, but they are not major. All Java progress is concentrated on backend cloud services for 10-15 years already. This can explain why Swing is so underdeveloped and JavaFX was thrown away. Basically much more effort is required to make Java GUI look and behave nicely, comparing to Delphi/Lazarus or .NET GUI libs or Qt.

DBeaver uses SWT toolkit, its widgets are as platform-native as Java can do. Some of them can be much faster with long text editing than default Java Swing widgets.

DBeaver works surprisingly nicely with less popular DBs. I work with Babelfish for PostgreSQL [1], it supports connections with SQL Server client libs. Most GUI client tools (like SSMS) expect "real" SQL Server on the other end of the wire - depend on various system views for DB introspection, so only partially work with Babelfish. Even if client tool is based on JDBC (like SQuirell SQL), it doesn't guarantee that this tool won't use additional SQL Server-specific queries for introspection. DBeaver is much better at this, I guess it is using JDBC API or DB-neutral INFORMATION_SCHEMA views for introspection.

[1] https://babelfishpg.org/

Feature set of Postgres on AWS in the article omits Babelfish feature. It is definitely worth mentioning, at least for some unfortunate people who are stuck with SQL Server, but would like to use Postgres as all others do. When you can use BCP or OPENQUERY to move your data to Postgres, instead of COPY - this can make a big difference for some of us.

> WARNING: Files not signed! I don't have a Google Inc. private key at the moment.

Its readme is worth reading.

> moving some data from SQL Server to Postgres

I don't have any first-hand experience with Postgres replication to share, just, when moving DB from MSSQL, Babelfish extensions for Postgres (https://babelfishpg.org/) may be of interest.

Project author here, just noticed that someone submitted this on HN. Thanks for your comment, I didn't thought about this. I wonder what would be a proper way to cover "who is behind"? Perhaps will add some kind of note to the bottom of the web page, something like: "WiltonDB Software founder previously worked on OpenJDK maintenance and packaging at Red Hat".

To clarify, this is a single-person effort, it took about a year to bring the project to current state. Though with current packaging setup (for Windows and Linux) future updates are expected to be smooth. Most of the efforts now are around the tooling like backup/restore.

Looks like this is just one of possible "non-default" extension allocators, that can be selected during libstdc++ compile time: https://gcc.gnu.org/onlinedocs/libstdc++/manual/memory.html#...