Replies: 2 comments 1 reply
-
I'm in favor of this, especially now that we've moved to C++17 starting with 1.13. Probably not everywhere, but in places where it really makes sense. |
Beta Was this translation helpful? Give feedback.
0 replies
-
#4332 @teksturi you can work with @matejk on this for 1.14.0 |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello.
We use Poco and are quite happy with it. However, we have already converted our entire internal codebase to use std::string_views. In many places, we now need to convert string_views to strings, as Poco only accepts the latter. These are the same places where we previously needed to create new strings, so it's not a catastrophic issue.
I'm still interested in whether we could support string_views in Poco and would like to know if this is something you would consider. I'm happy to start working on this, but first, we need a plan to determine if this will move forward. I understand that the minimum version Poco supports is C++17. However, there are still some places in the CMake files where C++14 is checked, but I wonder if that's just an oversight?
My proposal for proceeding with this is to start by converting to std::string_views in internal classes and source files. This should not affect external users at all. After that, we should decide whether to make a breaking change at some point and switch to using only string_views. Alternatively, we could take an intermediate step where we make this a compiler option. I've created a concept on Godbolt 1. We could get away of lot of things which currently use string iterators. These could in many places made just with string_views.
This is just an initial post on this subject. I have also done some prototyping with real code. If necessary, I can make a PR for a small portion, and maybe it would be easier to discuss it there.
Beta Was this translation helpful? Give feedback.
All reactions