Re: [PATCH V6 1/3] gdb: support for eBPF

[Andrew Burgess <andrew.burgess@embecosm.com>]

* Jose E. Marchesi <jose.marchesi@oracle.com> [2020-08-04 16:57:51 +0200]:

> 
> >> Hello.
> >> 
> >> > On Monday, August 3, 2020 4:03 PM, Jose E. Marchesi wrote:
> >> >> This patch adds basic support for the eBPF target: tdep and build
> >> >> machinery.  The accompanying simulator is introduced in subsequent
> >> >> patches.
> >> >>
> >> >> +/* Internal debugging facilities.  */
> >> >> +
> >> >> +/* When this is set to non-zero debugging information will be
> >> >> +   printed.  */
> >> >> +
> >> >> +static unsigned int bpf_debug_flag = 0;
> >> >> +
> >> >> +/* The show callback for 'show debug bpf'.  */
> >> >> +
> >> >> +static void
> >> >> +show_bpf_debug (struct ui_file *file, int from_tty,
> >> >> +	        struct cmd_list_element *c, const char *value)
> >> >
> >> > Just my two cents.  The 'struct' keyword can be removed in numerous places
> >> > throughout the file.  Also, NULL can be replaced with nullptr.
> >> 
> >> I prefer to write (and maintain) C.
> >
> > GDB is written in C++.  Though it's C history shows in many places,
> > the code base is ever moving towards C++, IMHO new code should be
> > quality C++.
> 
> I don't think anything in the proposed patch is invalid C++.  I was just
> expressing a personal preference in style, call it Cish C++ if not
> C.

I was thinking about this last night and I realised that "quality" was
absolutely the wrong word for me to use, so I apologise for that.

The issue here is not what is good (or quality) vs bad, which is what
I implied.  As you said, your code is perfectly valid C++.

What I should have said is that your preferences don't take precedent
over the projects agreed coding standard.  There are many
non-functional alternatives to the current coding standard, but
people are not free to just go with their personal preferences.  We
all stick to some agreed rules, and hopefully, over time we end up
with a consistent looking code base.

That said, as a more junior maintainer, I certainly defer to Simon on
this.  If he's happy with this patch being merged then I'm not going
to block it.

Once again, apologise for miss-speaking in my previous email.

Thanks,
Andrew


> 
> Of course, if the global maintainers decide that "quality" C++ requires
> avoiding `struct' keywords and using nullptr instead of NULL, and that
> it is important for new code to stick to it, then sure I will just
> change it without further discussion :)