Re: bug#25821: grep-3.0 FAILS on Make phase

2017-02-21 Thread rennes
Hello Bruno, thanks for the fix, now 'grep' built successfully.

Re: bug#25821: grep-3.0 FAILS on Make phase

2017-02-21 Thread Bruno Haible
Hi Rene, > > glthread/lock.c:54:40: error: 'PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP' > > undeclared (first use in this function) > > > > PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP); > > ^ Thanks for the report. Fixed in gnulib as follows: 2017-02-2

Re: manywarnings for C++

2017-02-21 Thread Reuben Thomas
On 21 February 2017 at 15:05, Bruno Haible wrote: > Hi Reuben, > > > ​For example, would it be sensible to have a single spec file, with each > > row as follows: > > > > warning language-list comment > > > > i.e. the first and third columns as at present, and the second giving the > > languages f

Re: manywarnings for C++

2017-02-21 Thread Bruno Haible
Hi Reuben, > ​For example, would it be sensible to have a single spec file, with each > row as follows: > > warning language-list comment > > i.e. the first and third columns as at present, and the second giving the > languages for which the flag is used? In my opinion 1) How to specify the se

Re: manywarnings for C++

2017-02-21 Thread Bruno Haible
Hi Reuben, This looks reasonable: Since a project can use both gcc and g++ and since the desired warning options for gcc and g++ are likely different, it's good to have different macros, that set different variables (WARN_CFLAGS vs. WARN_CXXFLAGS). > I'd be happy to work this up into a patch give

manywarnings for C++

2017-02-21 Thread Reuben Thomas
I needed this recently, so I made it. I have not attempted to refactor the code, as there seem to be multiple sensible-seeming ways to do this; I attach a simple duplication of the manywarnings files, reconfigured for C++. I'd be happy to work this up into a patch given some guidance on what shape

Re: manywarnings for C++

2017-02-21 Thread Reuben Thomas
On 21 February 2017 at 12:13, Reuben Thomas wrote: > I needed this recently, so I made it. I have not attempted to refactor the > code, as there seem to be multiple sensible-seeming ways to do this; I > attach a simple duplication of the manywarnings files, reconfigured for C++. > > I'd be happy