From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by master.gitmailbox.com (Postfix) with ESMTP id 3419F40E56 for ; Fri, 31 Dec 2021 16:53:47 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id A7AF568B029; Fri, 31 Dec 2021 18:53:46 +0200 (EET) Received: from iq.passwd.hu (iq.passwd.hu [217.27.212.140]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 0845268A962 for ; Fri, 31 Dec 2021 18:53:40 +0200 (EET) Received: from localhost (localhost [127.0.0.1]) by iq.passwd.hu (Postfix) with ESMTP id 32B17E64B6 for ; Fri, 31 Dec 2021 17:53:39 +0100 (CET) X-Virus-Scanned: amavisd-new at passwd.hu Received: from iq.passwd.hu ([127.0.0.1]) by localhost (iq.passwd.hu [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id aoZv0-IF3pXO for ; Fri, 31 Dec 2021 17:53:36 +0100 (CET) Received: from iq (iq [217.27.212.140]) by iq.passwd.hu (Postfix) with ESMTPS id 1638DE5CA0 for ; Fri, 31 Dec 2021 17:53:36 +0100 (CET) Date: Fri, 31 Dec 2021 17:53:36 +0100 (CET) From: Marton Balint To: FFmpeg development discussions and patches In-Reply-To: Message-ID: <23148c44-cec7-8d6c-6060-f719bea7c65c@passwd.hu> References: <20211227002613.25069-1-cus@passwd.hu> MIME-Version: 1.0 Subject: Re: [FFmpeg-devel] [PATCH 1/5] avformat/aviobuf: set AVIOContext->error on bprint buffer ENOMEM X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii"; Format="flowed" Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Archived-At: List-Archive: List-Post: On Fri, 31 Dec 2021, Andreas Rheinhardt wrote: > Marton Balint: >> This makes sure the error condition is kept in AVIOContext even if the user >> does not check the return value of avio_read_to_bprint or >> ff_read_line_to_bprint. >> >> Signed-off-by: Marton Balint >> --- >> libavformat/aviobuf.c | 8 ++++++-- >> 1 file changed, 6 insertions(+), 2 deletions(-) >> >> diff --git a/libavformat/aviobuf.c b/libavformat/aviobuf.c >> index 29d4bd7510..6f8a822ee3 100644 >> --- a/libavformat/aviobuf.c >> +++ b/libavformat/aviobuf.c >> @@ -875,8 +875,10 @@ static int64_t read_string_to_bprint_overwrite(AVIOContext *s, AVBPrint *bp, >> if (ret < 0) >> return ret; >> >> - if (!av_bprint_is_complete(bp)) >> + if (!av_bprint_is_complete(bp)) { >> + s->error = AVERROR(ENOMEM); >> return AVERROR(ENOMEM); >> + } >> >> return bp->len; >> } >> @@ -1351,8 +1353,10 @@ int avio_read_to_bprint(AVIOContext *h, AVBPrint *pb, size_t max_size) >> if (ret <= 0) >> return ret; >> av_bprint_append_data(pb, buf, ret); >> - if (!av_bprint_is_complete(pb)) >> + if (!av_bprint_is_complete(pb)) { >> + h->error = AVERROR(ENOMEM); >> return AVERROR(ENOMEM); >> + } >> max_size -= ret; >> } >> return 0; >> > > I don't really see the point of this: It is not a real read error that > should stick to the AVIOContext (which can still be used afterwards > without any issue). > If the user does not check the errors, then the user > has no one to blame but himself for missing errors. AVIO read/write behaviour is to store IO errors in the context so the user does not have to check for them in every call. It is not well documented which calls should be checked always, so the user might be under the impression that errors during read/write may be checked sometime later. Admittedly, ENOMEM is not an IO error, but I think it is better to store that as well in the context to keep the behaviour consistent, because in case of ENOMEM avio_read_to_bprint reads and drops undefined amount of data, so the context will also be in an undefined state. Other possibilities: - make avio_read_to_bprint read all the data regardless of AVBPrint fullness - mark avio_read_to_bprint av_warn_unused_result. - both :) But these also forces the user to check return values... So I kind of like my original approach better, because it maintains avio_read/write call behaviour that it is safe to check errors sometime later. Regards, Marton > > - Andreas > > PS: If the AVBPrint API had a documented way of marking data as used, > one could avoid those stack buffers and use the AVBPrint buffer directly > with av_bprint_get_buffer(). (Marking data as used would be equivalent > to incrementing len and ensuring that the buffer stays zero-terminated.) > If this were done, no already read data would be lost in case of a later > allocation failure. > _______________________________________________ > 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". > _______________________________________________ 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".