Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: "Rémi Denis-Courmont" <remi@remlab.net>
To: ffmpeg-devel@ffmpeg.org
Subject: Re: [FFmpeg-devel] [PATCH 2/8] lavu: new AVWriter API
Date: Tue, 02 May 2023 18:53:09 +0300
Message-ID: <3106500.VeKRRRbgAb@basile.remlab.net> (raw)
In-Reply-To: <20230428095508.221826-2-george@nsup.org>

Le perjantaina 28. huhtikuuta 2023, 12.55.02 EEST Nicolas George a écrit :
> +/**************************************************************************
> * + * Generic API
> +
> ***************************************************************************
> / +
> +#define FIELDOK(st, f) ((char *)(&(st)->f + 1) <= (char *)(st) +
> (st)->self_size) +
> +#define methods_assert_abi(methods) av_assert0(FIELDOK(methods, flush))
> +
> +static void printf_unchecked(AVWriter wr, const char *fmt, ...)
> +{
> +    va_list va;
> +
> +    va_start(va, fmt);
> +    wr.methods->vprintf(wr, fmt, va);
> +    va_end(va);
> +}

Indirecting the printf function seems pretty pointless. The last thing you 
want are different implementations of printf() with different limitations. And 
it makes writing different backends needlessly complex, while you could just 
use vasprintf().

Typically, with this kind of abstraction, only the raw bytes writing is 
abstracted away. Examples include funopen() and fopencookie().

As for hypothetical use cases whence vasprintf() wouldb e "too slow", then 
should not use printf()-style or function pointers to begin with. Besides if 
you _really_ want to avoid the heap allocation, you can also use fopencookie() 
on systems that provide it or equivalent.

> +void av_writer_write(AVWriter wr, const char *data, size_t size)
> +{
> +    methods_assert_abi(wr.methods);
> +    if (wr.methods->write) {
> +        wr.methods->write(wr, data, size);
> +    } else if (wr.methods->get_buffer) {
> +        size_t buf_size;
> +        char *buf = wr.methods->get_buffer(wr, size, &buf_size);
> +        if (buf_size > 0)
> +            memcpy(buf, data, FFMIN(size, buf_size));
> +        write_or_discard(wr, buf_size, size);

That sounds like it belongs in whichever backend actually wants to heap-
allocate the output buffer, not the frontend.

> +    } else if (wr.methods->vprintf) {
> +        size_t i;
> +        for (i = 0; i < size; i++)
> +            printf_unchecked(wr, "%c", data[i]);

This is an abstraction inversion (and also highly inefficient) + what I noted 
above.

> +    } else {
> +        av_writer_impossible(wr, "av_writer_write()");
> +    }
> +}
> +
> +void av_writer_print(AVWriter wr, const char *str)
> +{

This is an analogue of puts/fputs, not "print".

> +    av_writer_write(wr, str, strlen(str));
> +}
> +
> +void av_writer_printf(AVWriter wr, const char *fmt, ...)
> +{
> +    va_list va;
> +
> +    va_start(va, fmt);
> +    av_writer_vprintf(wr, fmt, va);
> +    va_end(va);
> +}
> +
> +void av_writer_vprintf(AVWriter wr, const char *fmt, va_list va)
> +{
> +    methods_assert_abi(wr.methods);
> +    if (wr.methods->vprintf) {
> +        wr.methods->vprintf(wr, fmt, va);
> +    } else if (wr.methods->get_buffer) {
> +        size_t buf_size;
> +        char *buf = wr.methods->get_buffer(wr, 128, &buf_size);
> +        va_list va2;
> +        int ret;
> +        va_copy(va2, va);
> +        ret = vsnprintf(buf, buf_size, fmt, va2);
> +        va_end(va2);
> +        if (ret < 0)
> +            return;
> +        if ((size_t)ret + 1 > buf_size) {
> +            buf = wr.methods->get_buffer(wr, ret + 1, &buf_size);
> +            ret = vsnprintf(buf, buf_size, fmt, va);
> +            if (ret < 0)
> +                return;
> +        }
> +        write_or_discard(wr, buf_size, ret);
> +    } else if (wr.methods->write) {
> +        AVBPrint bp;
> +        av_bprint_init(&bp, 0, AV_BPRINT_SIZE_UNLIMITED);
> +        av_vbprintf(&bp, fmt, va);
> +        if (av_bprint_is_complete(&bp)) {
> +            wr.methods->write(wr, bp.str, bp.len);
> +        } else {
> +            wr.methods->write(wr, bp.str, bp.size - 1);
> +            if (wr.methods->notify_discard)
> +                wr.methods->notify_discard(wr, bp.len - bp.size + 1);
> +        }
> +        av_bprint_finalize(&bp, NULL);
> +    } else {
> +        av_writer_impossible(wr, "av_writer_vprintf()");
> +    }
> +}

Same problems and overengineering as above.

> +
> +void av_writer_add_chars(AVWriter wr, char c, size_t n)
> +{
> +    char buf[512];
> +    size_t m;
> +
> +    m = FFMIN(n, sizeof(buf) - 1);
> +    memset(buf, c, m);
> +    while (n) {
> +        av_writer_write(wr, buf, m);
> +        n -= m;
> +        m = FFMIN(n, sizeof(buf) - 1);
> +    }
> +}
> +
> +void av_writer_flush(AVWriter wr)
> +{
> +    methods_assert_abi(wr.methods);
> +    if (wr.methods->flush)
> +        wr.methods->flush(wr);
> +}
> +
> +int av_writer_get_error(AVWriter wr, int self_only)
> +{
> +    methods_assert_abi(wr.methods);
> +    if (wr.methods->get_error)
> +        return wr.methods->get_error(wr, self_only);
> +    return 0;
> +}
> +

-- 
レミ・デニ-クールモン
http://www.remlab.net/



_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

  parent reply	other threads:[~2023-05-02 15:53 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-28  9:55 [FFmpeg-devel] [PATCH 1/8] lavu: add macros to help making future-proof structures Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 2/8] lavu: new AVWriter API Nicolas George
2023-04-28 10:37   ` Rodney Baker
2023-04-28 11:20     ` Nicolas George
2023-05-02 15:53   ` Rémi Denis-Courmont [this message]
2023-05-02 16:53     ` Nicolas George
2023-05-02 18:29       ` Rémi Denis-Courmont
2023-05-02 18:36         ` Nicolas George
2023-05-02 18:46           ` Rémi Denis-Courmont
2023-05-02 18:47             ` Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 3/8] lavu/writer: add test Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 4/8] lavf/dump: use a writer Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 5/8] lavu: add a JSON writer API (WIP) Nicolas George
2023-04-29  9:11   ` Nicolas George
2023-04-29  9:41     ` Anton Khirnov
2023-04-29 14:06       ` James Almer
2023-04-29 17:17         ` Nicolas George
2023-04-29 15:06       ` Derek Buitenhuis
2023-04-30  0:29         ` Kieran Kunhya
2023-05-01  6:20           ` Vittorio Giovara
2023-04-29 17:11       ` Nicolas George
2023-04-29 18:27         ` Anton Khirnov
2023-04-29 18:33           ` Nicolas George
2023-05-01  6:57             ` Leo Izen
2023-05-01  9:51               ` Nicolas George
2023-05-01 10:18               ` Jean-Baptiste Kempf
2023-04-30 15:06           ` Michael Niedermayer
2023-04-30 21:51             ` Kieran Kunhya
2023-05-01  9:46             ` Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 6/8] lavu: add JSON writer test (WIP) Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 7/8] lavf/options: add av_disposition_write() Nicolas George
2023-04-28  9:55 ` [FFmpeg-devel] [PATCH 8/8] lavf/dump: use av_disposition_write() Nicolas George
2023-04-29  8:17 ` [FFmpeg-devel] [PATCH 1/8] lavu: add macros to help making future-proof structures Anton Khirnov
2023-04-29 15:11   ` Derek Buitenhuis
2023-05-02 15:36 ` Rémi Denis-Courmont
2023-05-02 16:42   ` Nicolas George
2023-05-02 18:31     ` Rémi Denis-Courmont
2023-05-02 18:38       ` Nicolas George

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3106500.VeKRRRbgAb@basile.remlab.net \
    --to=remi@remlab.net \
    --cc=ffmpeg-devel@ffmpeg.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git