Hi,

On Thu, 2 May 2019, SZEDER Gábor wrote:

> On Wed, May 01, 2019 at 06:01:08AM -0400, Denton Liu wrote:
> > > Is it not possible to exclude certain directories for certain semantic
> > > patches?
> > >
> > > I guess we could also simply declare that *all* Coccinelle patches should
> > > leave `compat/` alone, on the basis that those files are likely coming
> > > from some sort of upstream. But then, `compat/mingw.c` and `compat/win32/`
> > > seem not to fall into that category...
> > >
> > > Ciao,
> > > Dscho
> >
> > Deciding whether this is a good idea is above my paygrade ;)

:-)

As a software developer, you surely have an opinion, though :-D

Thank you for the patch, it is a good conversation starter in the least,
and I hope that some variation of it will even make it to `master`.

> [...] what about 'compat/win32mmap.c' and 'compat/winansi.c'?  They look
> like "ours" as well.

Indeed, this is probably a good indicator that we'd want this to be an
opt-out, rather than an opt-in list.

For example, we know that we want to exclude compat/regex/ and
compat/poll/ from the `coccicheck` target.

> FWIW, out of curiosity I've run 'make coccicheck' on Linux with
> 'compat/mingw.c' and its friends explicitly added to C_SOURCES, and it
> seems to work...  it even found two places in 'mingw.c' where
> COPY_ARRAY could replace memcpy() :)

TBH I had not even known that those files were excluded from coccicheck by
default. I had assumed that all of Git's sources (and not just the
Linux-specific ones) were included in the target.

Since you *could* include it, I now assume that Coccinelle does not need
to follow the `#include`s (otherwise, it would have complained about not
finding the `windows.h` header in your setup).

If this new assumption is true, I wonder why we cannot make my former
assumption true as well: why not include *all* of Git's `.c` files in
`coccicheck`?

Ciao,
Dscho