Am Donnerstag, 20. Oktober 2016 um 16:23:05, schrieb Jean-Marc Lasgouttes <lasgout...@lyx.org> > Le 20/10/2016 à 16:03, Tommaso Cucinotta a écrit : > > On 20/10/2016 15:48, Jean-Marc Lasgouttes wrote: > >> Le 18/10/2016 à 18:42, Tommaso Cucinotta a écrit : > >>> So, the problem with Advanced F&R might be: how do we package on Mac > >>> OS-X ? > > > > I didn't see an answer to this, yet. Do we package with std::regex on > > Mac ? What problem in packaging using boost instead ? > > These days, we use std::regex by default when it is available. We may > have to reconsider this policy until AF&R is fixed. > > I do not think that it would be a problem for Stephan to build with > boost::regex explicitly. > > >> Tommaso, since regexes are mostly write-only, you are probably the > >> person who knows best what kind of Perl extensions you like to use. > > > > so the issue would be to double-check whether all regex-es in > > findadv.cpp can be converted to std::regex ones. But, is there a common > > syntax among boost::regex and std::regex that we can use so that > > libraries become interchangeable? Otherwise, if we convert all regexes > > to std:: syntax, we break everything with boost. > > As I understand it, Boost::regex uses Perl syntax, which is a extension > of normal-people-syntax. And I do not know what is the status of > ECMAscript syntax (used by std::regex). > > But suddenly I am not sure anymore of what the problem is. The original > message in this thread complains about "[0-9]+" not working. In theory > this should work both in Perl regexes and ECMAscript.
Actually, in this case, if you use '-dbg find', you can see that the used expression is lyxfind.cpp (871): Replaced text (to be used as regex): [0-9]+ lyxfind.cpp (874): Setting regexp to : '\`[0-9]+' lyxfind.cpp (879): Setting regexp2 to: '\`.*.*[0-9]+' > I am confused now. > > >> Otherwise, we would need some Advanced F&R automated tests to checks > >> how well our regexps behave. > > > > some are already there, see development/autotests/findadv-re-*-in.txt. > > Good. I am not sure any more that the problem we have is with internal > regexes... The person we need on this is Georg. > > JMarc Kornel
signature.asc
Description: This is a digitally signed message part.