I mean, there's a problem with the count variable. It should not redefine => int count. Outside of while, it will always be 0.
------------------------------------------------------------------发件人：Salvador \"Fandiño\" via RT <bug-Net-SSH2@rt.cpan.org>发送时间：2015年9月7日(星期一) 20:55收件人：goldenspider <goldenspider@aliyun.com>主 题：Re: [rt.cpan.org #106928] Net-SSH2-0.54_01 bug for net_ch_write <URL:
https://rt.cpan.org/Ticket/Display.html?id=106928 >
Show quoted text----- Original Message -----
> From: goldenspider via RT <bug-Net-SSH2@rt.cpan.org>
> To:
> Cc:
> Sent: Monday, September 7, 2015 2:08 PM
> Subject: [rt.cpan.org #106928] Net-SSH2-0.54_01 bug for net_ch_write
>
> Mon Sep 07 08:08:08 2015: Request 106928 was acted upon.
> Transaction: Ticket created by goldenspider@aliyun.com
> Queue: Net-SSH2
> Subject: Net-SSH2-0.54_01 bug for net_ch_write
> Broken in: (no value)
> Severity: (no value)
> Owner: Nobody
> Requestors: goldenspider@aliyun.com
> Status: new
> Ticket <URL:
https://rt.cpan.org/Ticket/Display.html?id=106928 >
>
>
> net_ch_write:1. while (offset < len_buffer) {
> int count = libssh2_channel_write_ex(ch->channel,
> XLATEXT, //one
> pv_buffer + offset,
> len_buffer - offset);
> if (count >= 0)
> offset += count;
> else if (!((count == LIBSSH2_ERROR_EAGAIN)
> && //two
> libssh2_session_get_blocking(ch->ss->session)))
> break;
> }//outside count will always equ 0.2.else if (LIBSSH2_ERROR_EAGAIN ==
> count
> && libssh2_session_get_blocking(ch->ss->session)
> == 0)
> is better?
Sorry, I am not understanding what you say.
The current version is not trivial because if has to handle all the following cases:
1. in blocking mode, write all the data
2. in non-blocking mode, write as much data as possible without blocking
3. if some error happens
a. if some data was already written, discard the error and report the number of bytes written.
b. if no data was written, report the error.
