I’m in favor of updated the CMS to a more current C++, given the sustained interest in C++ and ActiveMQ.
Generally, forking isn’t great for anyone in the long term. Perhaps its best suited for a v4.x release and baseline against newer C++ libs? -Matt Pavlovich > On Apr 14, 2021, at 11:28 AM, Arjun Ray <ara...@gmail.com> wrote: > > On Tue, 20 Oct 2020 00:11:24 -0400, I wrote: > > | Compiling the ActiveMQ-CPP distribution generates hundreds of warnings > | on using std::auto_ptr, which has been deprecated since C++11 nearly a > | decade ago, and in fact has been removed altogether in C++17. > | > | Is there a show-stopping issue with replacing std::auto_ptr with > | std::unique_ptr? I couldn't find any discussion of this. My resources > | are limited to the Linux platform (specifically Ubuntu), where I find > | that there are very few pain points in the replacement. > > I've since found that a 'discussion' of sorts took place several years > ago. Someone raised a JIRA ticket for precisely this issue, and > posted a proposed patch, whereupon the ticket was summarily closed as > "Won't fix", without any explanation. > > https://issues.apache.org/jira/browse/AMQCPP-628 > > | With these changes, the library compiles and all tests pass. I don't > | have the resources to verify this with other compilers or on other > | platforms, which is why I haven't yet investigated how to submit a > | pull request. > > It seems that a PR would meet the same fate. Forking may be the > better option for those looking to improve the codebase.