Based on semver, we should not remove any code that has not previously
been deprecated.
We have `@Removal` tags on much of our deprecated code so we observe
these - that is we shouldn't delete any deprecated methods with
removal set for POI 7.
I have no problem with deprecating code that no longer seems needed -
but I would object to removing it without deprecating it first.

There are multiple jars that we use that don't have proper module-info
setups. It isn't just SparseBitSet.

On Fri, 7 Nov 2025 at 03:17, axh <[email protected]> wrote:
>
> Hi,
>
> once 5.5.0 is released, I think we should decide what our targets are for 
> 6.0. I have a list with things I'd like to see (and would like to implement):
>
> - raise the baseline JDK to 17
> - remove obsolete code and utility classes not needed anymore and replace 
> with standard JDK 17 fearures
> - modernize the code base
> - solve the last remaining JPMS compatibility problems, namely SparseBitSet 
> not fully supporting the module system (uses an automatic module name) by 
> either removing the dependency on com.zaxxer.sparsebitset or getting the 
> author to add a real module-info.java to his project (honestly, I have very 
> little hope for that to happen - the maintainer seems quite oblivious to the 
> fact that automatic module names make life hard for people using jlkink). We 
> could use another sparse BitSet implementation, fork the project, copy the 
> source or change the POI code to not use a bitset at all.
>
> What are your thoughts?
>
> Axel
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to