Re: [PATCH] Improve QNX support in GIT

From: Matt Kraai <kraai@ftbfs.org>
Date: Feb 26, 2013, 17:25 UTC
Message-ID: <20130226172504.GA2271@ftbfs.org>

Hi Mike,

Mike Gorchak wrote:

Show 31 quoted lines

> diff --git a/config.mak.uname b/config.mak.uname
> index 8743a6d..2d42ffe 100644
> --- a/config.mak.uname
> +++ b/config.mak.uname
> @@ -527,14 +527,21 @@ ifeq ($(uname_S),QNX)
>  	HAVE_STRINGS_H = YesPlease
>  	NEEDS_SOCKET = YesPlease
>  	NO_FNMATCH_CASEFOLD = YesPlease
> -	NO_GETPAGESIZE = YesPlease
>  	NO_ICONV = YesPlease
>  	NO_MEMMEM = YesPlease
> -	NO_MKDTEMP = YesPlease
> -	NO_MKSTEMPS = YesPlease
>  	NO_NSEC = YesPlease
> -	NO_PTHREADS = YesPlease
>  	NO_R_TO_GCC_LINKER = YesPlease
> -	NO_STRCASESTR = YesPlease
>  	NO_STRLCPY = YesPlease
> +	# All QNX 6.x versions have pthread functions in libc
> +	# and getpagesize. Leave mkstemps/mkdtemp/strcasestr for
> +	# autodetection.
> +	ifeq ($(shell expr "$(uname_R)" : '6\.[0-9]\.[0-9]'),5)
> +		PTHREAD_LIBS = ""
> +	else
> +		NO_PTHREADS = YesPlease
> +		NO_GETPAGESIZE = YesPlease
> +		NO_STRCASESTR = YesPlease
> +		NO_MKSTEMPS = YesPlease
> +		NO_MKDTEMP = YesPlease
> +	endif
>  endif

Is there a point to the version checking? I don't know that anyone has tried to build Git on QNX 4, so adding a case for it seems misleading.

I didn't realize that QNX 6.3.2 provided getpagesize. Its header files don't provide a prototype, so when I saw the warning, I assumed it wasn't available. Since NO_GETPAGESIZE is only used by QNX, if it's OK to reintroduce the warning, NO_GETPAGESIZE might as well be removed entirely.

I don't think it's a good idea to just enable thread support. On QNX, once a process creates a thread, fork stops working. This breaks commands that create threads and then try to run other programs, such as "git fetch" with an https remote. If threads are enabled, I think that the uses of fork need to be audited and, if they can be called after a thread is created, fixed.