Re: [Bug-apl] SVN 624 Warnings

2015-05-08 Thread Juergen Sauermann
Hi, I have rewritten the macro causing this warning, so the warning should disappear. SVN 626. Elias, emacs mode was affected by this change, so please sync back. Peter, I would report this as an error to the authors of your compil

Re: [Bug-apl] SVN 624 Warnings

2015-05-07 Thread Juergen Sauermann
Hi Jay, true, but: #define rw_sv_def(q, _txt) { const char * qu = #q;    \    if (!strncmp(qu, "Quad_", 5)) { UCS_string ustr(qu + 5);   \     if (ustr.starts_iwith(qxx)) matches.push_back(ustr); } }

Re: [Bug-apl] SVN 624 Warnings

2015-05-07 Thread Jay Foad
Hi Jürgen, It's warning about line 51: rw_sv_def(CHI, "{ ... } axis argument") In this case you get "CHI" + 5, which is indexing way off the end of the string, which is why it's warning. Jay. On 7 May 2015 at 11:27, Juergen Sauermann wrote: > Hi, > > I believe by string they mean a simpl

Re: [Bug-apl] SVN 624 Warnings

2015-05-07 Thread Elias Mårtenson
How about rewriting it as &"somestring"[n] instead? That should result in the same code while silencing the compiler warnings. Regards, Elias On 7 May 2015 at 18:27, Juergen Sauermann wrote: > Hi, > > I believe by *string* they mean a simple string literal. What I am doing > is. for example >

Re: [Bug-apl] SVN 624 Warnings

2015-05-07 Thread Juergen Sauermann
Hi, I believe by string they mean a simple string literal. What I am doing is. for example "Quad_CT" + 5 (and so on for many ⎕xx functions and variables). The "Quad_CT" string literal, is, of course, not spelled directly (because t

Re: [Bug-apl] SVN 624 Warnings

2015-05-06 Thread Elias Mårtenson
I think the problem here is that std::string doesn't have an overloaded operator that takes an int. Thus, what happens is that the string gets coerced into a char *, which is then offset using the integer. Thus, if that's what you want you should make it clear using (someString.c_str() + offset).

Re: [Bug-apl] SVN 624 Warnings

2015-05-06 Thread Juergen Sauermann
Hi Peter, I believe I managed to fix the sem_xxx warnings in SVN 625. The Postgres warning should remain because it points at a harmless but somewhat incomplete SQL installation. The final warning seems to make no sense at all. And yes,

[Bug-apl] SVN 624 Warnings

2015-05-05 Thread Peter Teeson
Hi Jürgen: /Volumes/Data/Development/MyProjects/GNUAPLWS/apl-svn/src/apl-sqlite.cc:71:3: "PostgreSQL unavailable since ./configure could not detect it." /Volumes/Data/Development/MyProjects/GNUAPLWS/apl-svn/src/Parallel.cc /Volumes/Data/Development/MyProjects/GNUAPLWS/apl-svn/src/Parallel.cc:62:4