May 2023 - Boost - lists.boost.org

Re: [boost] New Website !
by Glen Fernandes 11 May '23

11 May '23

On Thu, May 11, 2023 at 10:35 AM Vinnie Falco wrote: > Hey, it’s Vinnie here again! I have some very exciting news to share > with you. The renovated website for Boost that we’ve been working on > for far too long is now going into its Public Beta phase! Feel free to > poke around and kick the tires but keep in mind this is a piece of > software and it is still under development. Some parts are missing, > incomplete,or buggy. Without further ado: > > https://boost.revsys.dev Thank you for working on this. The mechanics of this site are already impressive, and the dark mode design looks very nice (the Boost logo, theme, and all) even in its beta form. New documentation for Boost users about building and installing Boost, and for potential Boost authors about contributing to Boost development is also great to see. If anyone else viewing this like me has their system detected by the site as "light themed", then view the site in the dark mode. It has a much better look, and apparently the light mode design is still in progress. When the site is more developed, I can make sure that the release information also gets published to the new site. Glen

1 0

Re: [boost] charconv for C++11
by Robert Ramey 11 May '23

11 May '23

On 5/11/23 7:40 AM, Matt Borland via Boost wrote: >> How is this different from boost lexical cast? > > Robert, > > Charconv has a much smaller scope than lexical cast. We convert character arrays into built-in arithmetic types and vice versa. That’s it. I believe that this is what lexical_cast does. It might be different in that other than using arrays it uses strings. Since it has such a narrow scope we can heavily optimize these conversions, Honestly, I don't see how the scope is that much narrower than lexical_cast. which we have sunk a lot of time and effort into. I presume you're referring to charconv and not lexical_cast here. I don't question anyone's commitment, motives or competence in their efforts to contributing to boost. But this is not in itself an argument for accepting or rejecting a boost library submission. Thinking ahead a little bit, I'd have the following questions which might be brought up were such a component come up for review. a) How is this different from boost lexical cast? https://theboostcpplibraries.com/boost.lexical_cast. b) What will be the difference in user experience? c) if it's very similar, why wouldn't it be better to just enhance/refine/improve/correct lexical_cast itself? d) lexical cast uses an implementation strategy of leveraging C++ standard io library. This is a header only library which presumably, results in code highly optimized at compile time. Also it doesn't add any dependencies other than the standard library itself. Of course this optimization and hence performance will vary from one standard library/compiler implementation to another. But the same question applies to code written from scratch. Robert Ramey > > Matt > > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

1 0

Re: [boost] charconv for C++11
by Matt Borland 11 May '23

11 May '23

On Thu, May 11, 2023 at 17:16, Andrey Semashev via Boost <[boost@lists.boost.org](mailto:On Thu, May 11, 2023 at 17:16, Andrey Semashev via Boost <<a href=)> wrote: > On 5/11/23 17:40, Matt Borland via Boost wrote: >>> How is this different from boost lexical cast? >> >> Robert, >> >> Charconv has a much smaller scope than lexical cast. We convert character arrays into built-in arithmetic types and vice versa. That’s it. Since it has such a narrow scope we can heavily optimize these conversions, which we have sunk a lot of time and effort into. > > I remember there were a lot of optimizations poured into > Boost.Spirit.Qi/Karma. How does your library compare to those > implementations? > > _______________________________________________ > Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost Andrey, Frankly I’ve never looked at Boost.Spirit before. I can add comparison benchmarks to the charconv docs. Matt

1 0

Re: [boost] charconv for C++11
by Andrey Semashev 11 May '23

11 May '23

On 5/11/23 17:21, Vinnie Falco via Boost wrote: > On Thu, May 11, 2023 at 6:50 AM Robert Ramey via Boost > <boost(a)lists.boost.org> wrote: >> How is this different from boost lexical cast? > > Oh wow this has got to be one of the most amusing things posted to the > mailing list this year with the exception of the mathematical > breakthrough post which attached a gif of the author's monitor showing > the source code instead of providing a github link. I see nothing funny about Robert's post. Asking for comparison with existing solutions is a very reasonable request.

1 0

Re: [boost] charconv for C++11
by Andrey Semashev 11 May '23

11 May '23

On 5/11/23 17:40, Matt Borland via Boost wrote: >> How is this different from boost lexical cast? > > Robert, > > Charconv has a much smaller scope than lexical cast. We convert character arrays into built-in arithmetic types and vice versa. That’s it. Since it has such a narrow scope we can heavily optimize these conversions, which we have sunk a lot of time and effort into. I remember there were a lot of optimizations poured into Boost.Spirit.Qi/Karma. How does your library compare to those implementations?

1 0

Re: [boost] charconv for C++11
by Matt Borland 11 May '23

11 May '23

> How is this different from boost lexical cast? Robert, Charconv has a much smaller scope than lexical cast. We convert character arrays into built-in arithmetic types and vice versa. That’s it. Since it has such a narrow scope we can heavily optimize these conversions, which we have sunk a lot of time and effort into. Matt

1 0

New Website !
by Vinnie Falco 11 May '23

11 May '23

Hey, it’s Vinnie here again! I have some very exciting news to share with you. The renovated website for Boost that we’ve been working on for far too long is now going into its Public Beta phase! Feel free to poke around and kick the tires but keep in mind this is a piece of software and it is still under development. Some parts are missing, incomplete,or buggy. Without further ado: https://boost.revsys.dev This public beta will extend for at least 10 weeks as we finish the last remaining features, and put a few more rounds of polish on the artwork, visual styling, and user interface. After that if you like what you see then we will reset the database, move the repositories and the site to their new homes, and then deploy it for real! If you have suggestions or wish to report problems please feel welcome to open an issue here: https://github.com/cppalliance/temp-site/issues This is a Django site written and maintained in Python by a great group of folks from https://www.revsys.com/. The project is coordinated by Frank Wiles (https://www.revsys.com/about/bio/frankwiles) who shares the same deep commitment to open source that we Boost people do. We chose Django because it is well understood and supported, and because Python composes well with ease of maintenance. This technology stack allows us to execute on our future plans with confidence. Some cool things about this website include: * Log in with GitHub or Google * Light and Dark themes with selector * Asciidoc markdown for dynamic content * Antora content sources for site documentation * Data-driven dynamic pages for libraries and releases * Integrated control panel for producing Boost releases * Full-text search across the documentation and library headers * Professionally designed logo to evoke identity and instant recognition The layout of the site is carefully thought out to make sure that information can be found quickly. Fundamental topics each get a dedicated word in the top level navigation, visible on every page including the site and library documentation. Most of the information which is not library-specific is warehoused in “site docs” which is a collection of individual Antora content sources each authored in Asciidoc. Anyone can update the information on the website simply by contributing to the site-docs via pull request: https://github.com/cppalliance/site-docs Antora and Asciidoc are potent technologies that offer Boost plenty of room to grow to make our site documentation and library documentation more effective for users. While the initial release of the website after the beta will only have the basic functionality, we have many future plans. Some of these you have heard about already and some are new: * A web-based forum which interacts seamlessly with the mailing list * Polls and surveys for measuring the sentiment of Boost users and stakeholders * News section for aggregating off-site links, blog posts, and release progress messages * Integrated Review process. Outcomes go in the database and become dynamic content. * User Profiles record and show activity such as review participation or library submission We’re all very excited and proud of the work that went into this thing and hope that it really fulfills the positive impact that we intended. Thanks! Vinnie

1 0

Re: [boost] charconv for C++11
by Vinnie Falco 11 May '23

11 May '23

On Thu, May 11, 2023 at 6:50 AM Robert Ramey via Boost <boost(a)lists.boost.org> wrote: > How is this different from boost lexical cast? Oh wow this has got to be one of the most amusing things posted to the mailing list this year with the exception of the mathematical breakthrough post which attached a gif of the author's monitor showing the source code instead of providing a github link. Thanks

1 0

Re: [boost] charconv for C++11
by Robert Ramey 11 May '23

11 May '23

On 5/10/23 4:08 PM, Gavin Lambert via Boost wrote: > On 11/05/2023 02:01, Matt Borland wrote: >> What does charconv do? It converts a character sequence into a number >> (from_chars) or converts a number into a character sequence >> (to_chars). How is this different from boost lexical cast? Robert Ramey

1 0

Re: [boost] charconv for C++11
by Gavin Lambert 11 May '23

11 May '23

On 11/05/2023 02:01, Matt Borland wrote: > What does charconv do? It converts a character sequence into a number > (from_chars) or converts a number into a character sequence > (to_chars). This is similar to strtod and printf with a few key > differences. First and most important is a roundtrip guarantee. That > is a number converted into a character sequence, and then back into a > number will be identical to the starting value. Second, you can > easily specify an input/output format with a chars_format argument, > and also provide a target precision. "Roundtrip guarantee" and "specify format/precision" appear to be contradictory goals.

1 0