Re: [PULL 00/21] Vga 20200528 patches

[Peter Maydell]

On Fri, 29 May 2020 at 17:15, Philippe Mathieu-Daudé <f4bug@amsat.org> wrote:
> On 5/29/20 12:29 PM, Peter Maydell wrote:
> > Could somebody send a followup patch to fix the indentation
> > error checkpatch notices, please?
>
> If this is part of your scripts, this is a nice feature :)

No, I just noticed the patchew email.

> >
> > 5/21 Checking commit 97f369f2479d (hw/display/cirrus_vga: Use
> > qemu_log_mask(ERROR) instead of debug printf)
> > ERROR: suspect code indent for conditional statements (16, 12)
> > #34: FILE: hw/display/cirrus_vga.c:1038:
> >                 if (s->cirrus_blt_pixelwidth > 2) {
> > +            qemu_log_mask(LOG_GUEST_ERROR,
>
> I explained on the patches:
>
>   False positive.

The code is
            if (s->cirrus_blt_mode & CIRRUS_BLTMODE_TRANSPARENTCOMP) {
                if (s->cirrus_blt_pixelwidth > 2) {
            qemu_log_mask(LOG_GUEST_ERROR,
                          "cirrus: src transparent without colorexpand "
                          "must be 8bpp or 16bpp\n");
                    goto bitblt_ignore;
                }

checkpatch seems correct; the qemu_log_mask line is misindented,
and looking at the commit this is a misindent introduced in
commit 2b55f4d3504a9f34 "hw/display/cirrus_vga: Use
qemu_log_mask(ERROR) instead of debug printf". The old
fprintf() line was using indent of tab+tab+4 spaces, but
the new qemu_log_mask line is indented by 12 spaces, not 20.
(Tabs are always 8 spaces equivalent.)

Some days I wonder whether we should just do a bulk detabify
of the QEMU sources.

thanks
-- PMM