Timo Sachsenberg

QSql definitely looks nicer compared to our sql wrapper https://github.com/OpenMS/OpenMS/blob/develop/src/openms/include/OpenMS/FORMAT/SqliteConnector.h so the question would be if we should replace our SqliteConnector (and usage of QSql) by https://github.com/SRombauts/SQLiteCpp or one of...

[sqdbcpp](http://code.google.com/p/sqdbcpp/): RAII design, simple, no dependencies, UTF-8/UTF-16, new BSD license [sqlite3cc](http://ed.am/dev/sqlite3cc): uses boost, modern design, LPGPL [sqlite3pp](https://github.com/iwongu/sqlite3pp): modern design inspired by boost, MIT License [SQLite++](http://sqlitepp.berlios.de/): uses boost build system, Boost...

I think it is clear that for the library we ideally would not like to depend on qt (or e.g. boost) at all. For QString it is not as bad...

I guess it is not a qt version thing? I know that at least in Qt6 they tried to reduce dependency on QCoreApplication a bit

oms related tests still fail on ubuntu 20.04 

``` checking void store(const String& filename, const IdentificationData& id_data) ... Progress of 'converting protein identification runs': Progress of 'converting protein hits (run 1/2)': -- done [took 0.00 s (CPU), 0.00...

on gitpod: ``` -- done [took 0.00 s (CPU), 0.00 s (Wall)] -- -- done [took 0.00 s (CPU), 0.00 s (Wall)] -- 202: Progress of 'converting peptide identifications': --...

and seems it also exists on linux. - what now?

hack also doesn't seem work (constructs QCoreApplication if none is running): ``` void OMSFileStore::store(const IdentificationData& id_data) { // ugly hack to force loading of qsqlite plugin static int argc =...

- check our github action how we upload to a pypi server - check if we can do it with out server