> On Apr 21, 2021, at 01:23, Hans Verkuil <hverkuil@xxxxxxxxx> wrote:
>
>> On 21/04/2021 09:20, Rosen Penev wrote:
>> The array has a nullptr and 0 member for some reason. Remove and convert
>> loop to a for range one.
>>
>> Signed-off-by: Rosen Penev <rosenp@xxxxxxxxx>
>> ---
>> utils/common/v4l2-info.cpp | 33 +++++++++++++++------------------
>> 1 file changed, 15 insertions(+), 18 deletions(-)
>>
>> diff --git a/utils/common/v4l2-info.cpp b/utils/common/v4l2-info.cpp
>> index cb3cb91f7..0359cf137 100644
>> --- a/utils/common/v4l2-info.cpp
>> +++ b/utils/common/v4l2-info.cpp
>> @@ -3,6 +3,8 @@
>> * Copyright 2018 Cisco Systems, Inc. and/or its affiliates. All rights reserved.
>> */
>>
>> +#include <array>
>> +
>> #include <v4l2-info.h>
>>
>> static std::string num2s(unsigned num, bool is_hex = true)
>> @@ -411,21 +413,20 @@ std::string mbus2s(unsigned flags, bool is_hsv)
>> return flags2s(flags, mbus_ycbcr_def);
>> }
>>
>> -static const flag_def selection_targets_def[] = {
>> - { V4L2_SEL_TGT_CROP_ACTIVE, "crop" },
>> - { V4L2_SEL_TGT_CROP_DEFAULT, "crop_default" },
>> - { V4L2_SEL_TGT_CROP_BOUNDS, "crop_bounds" },
>> - { V4L2_SEL_TGT_COMPOSE_ACTIVE, "compose" },
>> - { V4L2_SEL_TGT_COMPOSE_DEFAULT, "compose_default" },
>> - { V4L2_SEL_TGT_COMPOSE_BOUNDS, "compose_bounds" },
>> - { V4L2_SEL_TGT_COMPOSE_PADDED, "compose_padded" },
>> - { V4L2_SEL_TGT_NATIVE_SIZE, "native_size" },
>> - { 0, nullptr }
>
> The idea of having this sentinel is that this makes it easy to add new
> entries without having to update the array size.
Not following this. I assume it’s some C feature.
>
>> +static constexpr std::array<flag_def, 8> selection_targets_def{
>
> Something you need to do here, adding a new flag means updating the size.
I assume this is a small issue. It’s an immediate compile error anyway.
>
> New flags are added regularly, so keeping that robust is a good idea IMHO.
>
> If it were possible to write:
>
> static constexpr std::array<flag_def> selection_targets_def{
>
> i.e. without an explicit size, then this would make sense, but C++
> doesn't allow this. And std::vector allocates the data on the heap,
> which is less efficient as well.
But this is possible. With C++17 :). That would necessitate a minimum of GCC6 though.
>
> Let's just keep using normal arrays in this case, they do the job
> just fine. Just because you have a hammer, it doesn't mean everything
> is now a nail :-)
>
> Regards,
>
> Hans
>
>> + flag_def{ V4L2_SEL_TGT_CROP_ACTIVE, "crop" },
>> + flag_def{ V4L2_SEL_TGT_CROP_DEFAULT, "crop_default" },
>> + flag_def{ V4L2_SEL_TGT_CROP_BOUNDS, "crop_bounds" },
>> + flag_def{ V4L2_SEL_TGT_COMPOSE_ACTIVE, "compose" },
>> + flag_def{ V4L2_SEL_TGT_COMPOSE_DEFAULT, "compose_default" },
>> + flag_def{ V4L2_SEL_TGT_COMPOSE_BOUNDS, "compose_bounds" },
>> + flag_def{ V4L2_SEL_TGT_COMPOSE_PADDED, "compose_padded" },
>> + flag_def{ V4L2_SEL_TGT_NATIVE_SIZE, "native_size" },
>> };
>>
>> bool valid_seltarget_at_idx(unsigned i)
>> {
>> - return i < sizeof(selection_targets_def) / sizeof(selection_targets_def[0]) - 1;
>> + return i < selection_targets_def.size();
>> }
>>
>> unsigned seltarget_at_idx(unsigned i)
>> @@ -437,13 +438,9 @@ unsigned seltarget_at_idx(unsigned i)
>>
>> std::string seltarget2s(__u32 target)
>> {
>> - int i = 0;
>> -
>> - while (selection_targets_def[i].str != nullptr) {
>> - if (selection_targets_def[i].flag == target)
>> - return selection_targets_def[i].str;
>> - i++;
>> - }
>> + for (const auto &def : selection_targets_def)
>> + if (def.flag == target)
>> + return def.str;
>> return std::string("Unknown (") + num2s(target) + ")";
>> }
>>
>>
>
