[FFmpeg-devel] [PATCH v14 4/5] libavformat: Remove MAX_PATH limit and use UTF-8 version of getenv()

[Soft Works]

> -----Original Message-----
> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of Nil
> Admirari
> Sent: Monday, June 13, 2022 6:26 PM
> To: ffmpeg-devel at ffmpeg.org
> Subject: [FFmpeg-devel] [PATCH v14 4/5] libavformat: Remove MAX_PATH
> limit and use UTF-8 version of getenv()
> 
> 1. getenv() is replaced with getenv_utf8() across libavformat.
> 2. New versions of AviSynth+ are now called with UTF-8 filenames.
> 3. Old versions of AviSynth are still using ANSI strings,
>    but MAX_PATH limit on filename is removed.
> ---
>  libavformat/avisynth.c    | 39 +++++++++++++++++++++++++++----------
> --
>  libavformat/http.c        | 20 +++++++++++++-------
>  libavformat/ipfsgateway.c | 35 +++++++++++++++++++++++------------
>  libavformat/tls.c         | 11 +++++++++--
>  4 files changed, 72 insertions(+), 33 deletions(-)
> 
> diff --git a/libavformat/avisynth.c b/libavformat/avisynth.c
> index 8ba2bdead2..a97d12b6b6 100644
> --- a/libavformat/avisynth.c
> +++ b/libavformat/avisynth.c
> @@ -34,6 +34,7 @@
>  /* Platform-specific directives. */
>  #ifdef _WIN32
>    #include "compat/w32dlfcn.h"
> +  #include "libavutil/wchar_filename.h"
>    #undef EXTERN_C
>    #define AVISYNTH_LIB "avisynth"
>  #else
> @@ -56,6 +57,7 @@ typedef struct AviSynthLibrary {
>  #define AVSC_DECLARE_FUNC(name) name ## _func name
>      AVSC_DECLARE_FUNC(avs_bit_blt);
>      AVSC_DECLARE_FUNC(avs_clip_get_error);
> +    AVSC_DECLARE_FUNC(avs_check_version);
>      AVSC_DECLARE_FUNC(avs_create_script_environment);
>      AVSC_DECLARE_FUNC(avs_delete_script_environment);
>      AVSC_DECLARE_FUNC(avs_get_audio);
> @@ -137,6 +139,7 @@ static av_cold int avisynth_load_library(void)
> 
>      LOAD_AVS_FUNC(avs_bit_blt, 0);
>      LOAD_AVS_FUNC(avs_clip_get_error, 0);
> +    LOAD_AVS_FUNC(avs_check_version, 0);
>      LOAD_AVS_FUNC(avs_create_script_environment, 0);
>      LOAD_AVS_FUNC(avs_delete_script_environment, 0);
>      LOAD_AVS_FUNC(avs_get_audio, 0);
> @@ -807,26 +810,38 @@ static int
> avisynth_create_stream(AVFormatContext *s)
>  static int avisynth_open_file(AVFormatContext *s)
>  {
>      AviSynthContext *avs = s->priv_data;
> -    AVS_Value arg, val;
> +    AVS_Value val;
>      int ret;
> -#ifdef _WIN32
> -    char filename_ansi[MAX_PATH * 4];
> -    wchar_t filename_wc[MAX_PATH * 4];
> -#endif
> 
>      if (ret = avisynth_context_create(s))
>          return ret;
> 
> +    if (!avs_library.avs_check_version(avs->env, 7)) {

I like the version check. I don't know about all the derivatives
of AviSynth, but I assume you have checked that it's valid for
the common ones (or at least the original non-Plus variant)?

> +        AVS_Value args[] = {
> +            avs_new_value_string(s->url),
> +            avs_new_value_bool(1) // filename is in UTF-8
> +        };
> +        val = avs_library.avs_invoke(avs->env, "Import",
> +                                     avs_new_value_array(args, 2),
> 0);
> +    } else {
> +        AVS_Value arg;
>  #ifdef _WIN32
> -    /* Convert UTF-8 to ANSI code page */
> -    MultiByteToWideChar(CP_UTF8, 0, s->url, -1, filename_wc,
> MAX_PATH * 4);
> -    WideCharToMultiByte(CP_THREAD_ACP, 0, filename_wc, -1,
> filename_ansi,
> -                        MAX_PATH * 4, NULL, NULL);
> -    arg = avs_new_value_string(filename_ansi);
> +        char *filename_ansi;
> +        /* Convert UTF-8 to ANSI code page */
> +        if (utf8toansi(s->url, &filename_ansi)) {

Two ideas came to my mind how this could be done better.
What's actually needed here is not a string conversion, we need
a valid and usable filename, and the function could be more
something like "get_ansi_filename()".

The first thing that this function could do is to convert the
the filename to ANSI and right back to UTF-8, then compare the
UTF-8 result with the original UTF-8 string. When both are equal,
we know that the conversion is safe, otherwise we know that it
won't work.

Then, we can use the win32 API GetShortFileName(). Which returns
file and directory names in 8.3 notation which (IIRC) contains
only letters which are valid in the ANSI code page.

8.3 file names do not always exist (depending on system config), 
but it's always worth trying.

Should both of these procedures fail, we could at least output
a useful message, explaining why it doesn't work.

Let me know what you think.

sw