I ended up removing the offending paragraph completely – it really didn’t add anything to the post and gave people the wrong impression about the differences between the C skeleton and C++ skeleton. If you are playing with tools like Flex and Bison you probably know enough to decide for yourself whether you should use C or C++, so the pros and cons of this argument are left for the reader to decide.
-- Robert Hollencamp rhollenc...@gmail.com On Mon, Mar 8, 2010 at 3:33 AM, Istvan Sandor <sand...@rht.bme.hu> wrote: > > Hi, > > I definitely wouldn't say this: > > "Edit: March 7 2010 [...] In C mode, the generated > code uses global variables to store information, making them > non-thread-safe. It is also a pain to create multiple different parsers > / scanners and include them in the same program. Using C++ mode, all of > the data needed is encapsulated by the object, making all of the > generated code reentrant. You can have multiple different parsers / > scanners in the same program, and instantiate and use multiple instances > of each at the same time." > > This simply isn't true, you can generate reentrant parsers with bison > and you can also safely have multiple different parsers by renaming > yyparse() and the like. > > Just my 2 cents. > > Istvan > > > On Sat, Mar 06, 2010 at 09:05:30PM -0500, Robert Hollencamp wrote: > > I made an example of using Flex and Bison together in C++ and was > wondering > > if you guys could provide some feedback. Despite the C++ interface being > > 'experimental' I found it works out pretty good after you figure it out. > > > > http://www.thewaffleshop.net/2010/03/06/flex-bison-cpp-example/ > > > > -- > > Robert Hollencamp > > rhollenc...@gmail.com > > _______________________________________________ > > help-bison@gnu.org http://lists.gnu.org/mailman/listinfo/help-bison _______________________________________________ help-bison@gnu.org http://lists.gnu.org/mailman/listinfo/help-bison
