(readded cc: cscope-devel) 2011/2/16 Hans-Bernhard Bröker <hbbroe...@t-online.de>: > On 17.02.2011 01:56, Ken Smith wrote: >> >> 2011/2/16 Hans-Bernhard Bröker<hbbroe...@t-online.de>: >> >>> Actually, you would _always_ have to do cscope -R -b -q. The ability to >>> concatenate single-letter options that don't use arguments is a GNU >>> feature >>> offered by getopt_long(). No _GNU_SOURCE, no -Rbq. > >> Ok. I'm compiling on Linux with the GNU compiler which doesn't define >> it by default. > > As of when? It's actually predefined by GNU libc, not the compiler, if > memory serves.
I don't know when this changed if it was previously not the case but let me demonstrate what I observe. % gcc --version gcc (Ubuntu/Linaro 4.4.4-14ubuntu5) 4.4.5 Copyright (C) 2010 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. % cat macro-check.c #if defined(_GNU_SOURCE) #error _GNU_SOURCE is defined #endif % gcc -c macro-check.c macro-check.c:4: error: #error _GNU_SOURCE is not defined Interestingly, % g++ -c macro-check.c macro-check.c:2: error: #error _GNU_SOURCE is defined and this difference is complained about elsewhere on the internet. When I do a fresh checkout of cscope, ./configure, and build it, _GNU_SOURCE is not defined. I verified this by putting a check similar to the above in src/main.c. Perhaps it would be worth your while to check with someone at one of the distros to see if they have to apply any patches to cscope to get it to build with _GNU_SOURCE enabled. >> I'm happy to learn that cscope already does what I want but so far, I >> don't think -P does. I'm happy to be proven wrong. > > You don't want to pass $(pwd) to -P. It's meant to be used like this: > > cscope -d -f /some/where/cscope.out -P/some/where This does work but I find it more straightforward to build the database with absolute paths. Is there any harm in that? The vim/cscope integration mirrors the above by allowing an argument to :cs add. cs add /path/to/cscope.out /path/to But isn't it more convenient to simply build the database with absolute paths and then you don't have to use -P to load each one of them every time. Indeed, the invocation is somewhat redundant don't you think? Perhaps the default behavior of -f should be to automatically prepend the supplied path when that path is not the same as $PWD. The default behavior of using paths relative to the current directory at database creation time is in harmony with that. At any rate, the patch is yours to do with as you please. I personally find it useful and would love to see it integrated into the mainline but I'm also content with patching my own copy. The patch probably needs to be modified to avoid using unistd.h for Windows and such but I don't have a Windows development box handy to work out the details. ks ------------------------------------------------------------------------------ The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: Pinpoint memory and threading errors before they happen. Find and fix more than 250 security defects in the development cycle. Locate bottlenecks in serial and parallel code that limit performance. http://p.sf.net/sfu/intel-dev2devfeb _______________________________________________ Cscope-devel mailing list Cscope-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cscope-devel
