[fpc-devel] [RFC] Modernising the FPC Release Process -- Proposal for Review

Nikolay Nikolov nickysn at gmail.com
Tue Apr 21 14:40:35 CEST 2026


On 4/19/26 9:47 PM, Graeme Geldenhuys via fpc-devel wrote:
> On Sunday, 19 April 2026 18:46:46 BST Martin Frb via fpc-devel wrote:
>> area goes up, then that is a risk factor for conflicts.
> I million lines of code in FPC (I didn't count - so don't hold me to it) and
> what... 15 developers with commit access. Chances are super slim.

Changes to the compiler in trunk almost always lead to merge conflicts. 
Adding a target, refactoring the compiler to add such things as LLVM 
support (by adding the high level code generator), the current 
refactoring of the front end to eliminate global variables (and allow 
multithreaded compilation in the future) touch a lot of files! So, your 
statistic is too simplistic and wrong.

See e.g. the (unfinished) changes I'm currently working on in the 
eliminate_global_state branch:

https://gitlab.com/freepascal.org/fpc/source/-/compare/main...eliminate_global_state?from_project_id=28644964

These changes will probably be merged to main after the branching of the 
next stable branch for FPC 3.4.0, so merging of fixes between 3.6-main 
and 3.4-fixes will very often lead to conflicts!

Nikolay



More information about the fpc-devel mailing list