<div>It would be also great if there was an official set of interfaces (not necessarily in interface type form but API description) describing all these generic containers and algorithms.</div><div>Thus, all implementations which follow these interfaces are interchangeable.</div><div>I believe this is even more important than 4-th or 5-th implementation.</div><div> </div><div>---</div><div>Best regards, George</div><div> </div><div>06.06.2019, 17:58, "Ben Grasset" <operator97@gmail.com>:</div><blockquote><div><div>On Wed, Jun 5, 2019 at 11:11 AM Sven Barth via fpc-devel <<a href="mailto:fpc-devel@lists.freepascal.org">fpc-devel@lists.freepascal.org</a>> wrote:</div><div><blockquote style="border-left-color:rgb( 204 , 204 , 204 );border-left-style:solid;border-left-width:1px;margin-bottom:0px;margin-left:0.8ex;margin-right:0px;margin-top:0px;padding-left:1ex"><div><div>You could try to adjust the FGL unit to use the pluggable sorting system introduced here: <a href="https://svn.freepascal.org/cgi-bin/viewvc.cgi?view=revision&revision=41167">https://svn.freepascal.org/cgi-bin/viewvc.cgi?view=revision&revision=41167</a></div></div></blockquote><div> </div><div>Wouldn't that system itself need to be generic and not just use void pointers first? Or is that part of what you meant?</div><div> </div><div>IMO something FGL would benefit from also is the removal of the probably-not-the-best-design inheritance from non-generic TFPSList to everything else, which prevents inlining in a large number of places (the compiler notes indicating such are visible basically any time you use anything from FGL.)</div></div></div>,<p>_______________________________________________<br />fpc-devel maillist - <a href="mailto:fpc-devel@lists.freepascal.org">fpc-devel@lists.freepascal.org</a><br /><a href="http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-devel">http://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-devel</a></p></blockquote>