Le 06/06/2016 à 11:16, Joseph Van Riper a écrit :
On Thu, Jun 2, 2016 at 6:09 PM Vicente J. Botet Escriba < vicente.botet@wanadoo.fr> wrote:
Hi,
Some of us have to work yet with C++98 compiler at work (due to well know reasons) and this for a long time. This is a reality.
I was wondering about forking a public domain C++11/14/17 standard library and adapt it to C++98 compilers. This library could be seen as another TR1, an extension of the C++03 standard library.
Does this make any sens? Would this project be of interest for the Boost/C++ community?
Yes. At least for what we do where I work.
The cyber-security field works with very old machines, either for forensics (people forget they have some old machine until it gets hacked and some old process that always "just worked" stops working), teaching (teaching old, outmoded techniques helps to illustrate cyber security problems without giving people tools for modern exploitation), or competitions.
I kind of hate having to work with clunky, antediluvian C++ compilers that magnify the level of woe required to deal with old machinery. Especially while the rest of the C++ world moves on to the spiffy-shiny features found in the later compilers. Having at least a library that helps us use some of those nifty features can help us use our skills on more modern projects when we move on to another job, while easing the efforts required to do our work. A library wouldn't change your compiler ;-), but it could change the way you program :).
Anyway, if you find that this will be useful to you at your work we will
need your participation. Would you be interested in working on it?
Hmmm... am I even knowledgeable enough?
While I may find myself learning quite a lot about C++ by studying the boost code people write, I am also humbled by what I see there. I fear I am only beginning to dip into some of the intermediate meta-programming techniques that seem to confer tremendous power to this language.
as a Boost community project?
outside Boost, as an independent project?
Hmmm...
The boost 'brand' confers a certain gravitas. If you wish to maintain this gravitas in the new library, I'd say keep the boost brand, but require the same rigor. We have a License issue, so it couldn't be a Boost library :(
Vicente P.S. Please, contact me privately if you are interested.