I totally and vehemently agree with Hermes here: We should NOT ditch VS2010.

Arguments that were raised by others against VS2010 and my reply:
-"ditching it brings in new developers to ros magically" <- I do ask then, 
where are they? I don't see any.
-"we should not be limited to strict C89" <- No one did request for that. All 
we want is to remain compatible to VS2010. A subset.
-"syncing BTRFS is a bit harder" <- So what? Then let others do the job. 
Actually BTRFS is not even a mandatory feature of ros. Having it in the tree is 
just luxury.
-"libc++" <- I see no urgent need and nothing it would give in return that 
would outweight what we would sacrifice.

My arguments again for keeping VS2010:
-VS2010 creates the smallest binaries of all compilers we do support
-VS2010 CAN be installed in ros, when ros is installed as Server during 2nd 
stage, (yes this was not the case for a very short moment, unfortunately 
exactly when we discussed in https://github.com/reactos/reactos/pull/2658 but 
now it works again, even in 0.4.14-RC51)
-VS2010 can now even open the VS2010 cmd prompt see 
-no other VS>2010 can even complete its setup in ros, and that will remain like 
that for many years to come
-VS2010 is the last version that runs on XPSP3 (which is important for some of 
our devs including myself)
-VS2010 is the last version that runs on 2k3SP2 which is our current target
-VS2010 is reliable with industry-proven stability, and itself no moving target 
(unlike VS2019 which breaks our builds every few weeks when MS upgrades it)
-VS2010 DOES provide std::unique_ptr. Stating anything else is just a lie. It 
covers > 95% of CPP2011-standard. It is absolutely possible and not complicated 
to write leak-free-code with it.

Ros-dev mailing list

Reply via email to