Core dump with `corrupted double-linked list` on SQL export

[mattias-aims]

Bug description

When using the pgmodeler-cli to export to SQL then it core dumps with corrupted double-linked list. The same model exports without problems using v1.2.2.

How to reproduce

export PGSQL_VERSION=17.0
pgmodeler-cli --export-to-file --input $INPUT_MODEL --output $OUTPUT_SQL --pgsql-ver $PGSQL_VERSION --silent

Expected behavior
No core dump.

Info about your desktop

• OS: Debian Linux
• Version: Linux debian 6.12.57+deb13-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.12.57-1 (2025-11-05) x86_64 GNU/Linux
• Window manager: Gnome
• pgModeler version: pgmodeler_plus-1.2.3-linux64 from tarball
• Qt version: ?

Stacktrace / Debug info

corrupted double-linked list is the only output I get.

[rkhaotix]

Hm... strange. The crash also happens on GUI?

I'll investigate this ASAP.

[rkhaotix]

Maybe it's something with your model that hit a bug. I've tested the export with sample.dbm in both 1.2.3 and 2.0.0-alpha1, and the CLI didn't crash. Could you please send me a copy of your model or even create a sample that can trigger the crash?

[DAAAAAAAAAAAAAAAAN]

I'm running into the exact same issue. Exporting from the GUI works without issue. With the CLI, It seems to do the export successfully, but something crashes during cleanup?

2026/02/16 14:06:49 Generating migration statements
corrupted double-linked list
/home/XXXX/.local/bin/pgmodeler-cli: line 3: 277551 Aborted                 (core dumped) /opt/pgModelerPlus/pgmodeler-cli $@
Error: failed to generate migration statements: failed to export model: failed to run pgmodeler: exit status 134 
pgModeler Plus command line interface.
Version 1.2.3 - 20260207.e7ed75eeb Qt 6.9.3
PostgreSQL Database Modeler Project - pgmodeler.io
Copyright 2006-2026 Raphael Araújo e Silva <raphael@pgmodeler.io>

Starting model export...
Loading input file: /home/XXXX/XXXX.dbm
Export to SQL script file: /home/XXXX/XXXX.sql
Generating SQL code for PostgreSQL `15.0'
[1%] Generating SQL code: `stock' (Schema)
[1%] Generating SQL code: `sales' (Schema)
[1%] Generating SQL code: `shipping' (Schema)
...
[99%] Generating SQL code: `grant_r_e2bf4860ae' (Permission)
[99%] Generating SQL code: `grant_r_be88535797' (Permission)
[99%] Generating SQL code: `grant_r_99b5cb7a90' (Permission)
[100%] SQL file `/home/XXXX/XXXX.sql' successfully written.
Export successfully ended!

Flushing used memory...
Done!

XXXX@XXXX:~/XXXX$ 

With 1.2.2, this core dump happens only sometimes (I guesstimate 1 in 10 runs). I reverted back to 1.2.2 and I am unable to share my model file, nor do I have time to test with other models, unfortunately. My AI made an analysis:

Exit code 134 = 128 + 6, meaning the process was killed by signal 6 (SIGABRT). It's not a return code from pgModeler or Qt application code.

The sequence is:

glibc's malloc detects heap corruption → prints corrupted double-linked list
glibc calls abort()
Kernel delivers SIGABRT (signal 6) to the process
Shell reports exit code 128 + 6 = 134
So the corruption is detected by glibc, not by pgModeler or Qt. The actual bug causing the corruption could be in either pgModeler's C++ code or Qt, but glibc is the one pulling the trigger. This reinforces hypothesis 1 — something in the 1.2.3 update introduced a memory bug that manifests during cleanup.

Hope this helps

[rkhaotix]

Thanks @DAAAAAAAAAAAAAAAAN

Did you try with 2.0.0-alpha1? The issue also happens?
I'm still investigation, but couldn´t reproduce the crash, that's why I'm suspecting that sometihing on certain input models are causing the problem. :(

[DAAAAAAAAAAAAAAAAN]

I just tested it with 2.0.0-alpha1 and the issue is not present there

[mattias-aims]

Issue fixed in 2.0.0-alpha1