Re: [PATCH v6 05/13] pkt-line: add packet_write_gently()

From: Lars Schneider <larsxschneider@gmail.com>
Date: Aug 26, 2016, 09:40 UTC
Message-ID: <0BD3A97F-A9FD-45A6-AF82-E01CC6B9BA89@gmail.com>
In-Reply-To: <xmqq4m68ldrg.fsf@gitster.mtv.corp.google.com>

Show 13 quoted lines

> On 25 Aug 2016, at 23:50, Junio C Hamano <gitster@pobox.com> wrote:
> 
> larsxschneider@gmail.com writes:
> 
>> From: Lars Schneider <larsxschneider@gmail.com>
>> 
>> packet_write_fmt() has two shortcomings. First, it uses format_packet()
>> which lets the caller only send string data via "%s". That means it
>> cannot be used for arbitrary data that may contain NULs. Second, it will
>> always die on error.
> 
> As you introduced _gently in 3/13, the latter is no longer a valid
> excuse to add this function.  Just remove the sentence "Second, ...".

Agreed!

Show 39 quoted lines

>> Add packet_write_gently() which writes arbitrary data and returns `0`
>> for success and `-1` for an error. This function is used by other
>> pkt-line functions in a subsequent patch.
>> 
>> Signed-off-by: Lars Schneider <larsxschneider@gmail.com>
>> ---
>> pkt-line.c | 12 ++++++++++++
>> 1 file changed, 12 insertions(+)
>> 
>> diff --git a/pkt-line.c b/pkt-line.c
>> index cad26df..7e8a803 100644
>> --- a/pkt-line.c
>> +++ b/pkt-line.c
>> @@ -3,6 +3,7 @@
>> #include "run-command.h"
>> 
>> char packet_buffer[LARGE_PACKET_MAX];
>> +static char packet_write_buffer[LARGE_PACKET_MAX];
>> static const char *packet_trace_prefix = "git";
>> static struct trace_key trace_packet = TRACE_KEY_INIT(PACKET);
>> static struct trace_key trace_pack = TRACE_KEY_INIT(PACKFILE);
>> @@ -155,6 +156,17 @@ int packet_write_fmt_gently(int fd, const char *fmt, ...)
>> 	return (write_in_full(fd, buf.buf, buf.len) == buf.len ? 0 : -1);
>> }
>> 
>> +int packet_write_gently(const int fd_out, const char *buf, size_t size)
>> +{
>> +	if (size > sizeof(packet_write_buffer) - 4)
>> +		return -1;
>> +	packet_trace(buf, size, 1);
>> +	memmove(packet_write_buffer + 4, buf, size);
>> +	size += 4;
>> +	set_packet_header(packet_write_buffer, size);
> 
> It may not matter all that much, but from code-reader's point of
> view, when you know packet_write_buffer[] will contain things A and
> B in this order, and when you have enough information to compute A
> before stasrting to fill packet_write_buffer[], I would prefer to
> see you actually fill the buffer in that natural order.

I did that because when packet_write_stream_with_flush_from_fd() calls packet_write_gently() then buf[] is actually packet_write_buffer[]:

https://github.com/larsxschneider/git/blob/d474e6a4c2523b87624a07111eb7a4f2dcd12426/pkt-line.c#L185-L192

Therefore I would override the first 4 bytes. However, the code evolved for some reason in that way but looking at it now I think that is an obscure, likely meaningless optimization. I'll use a separate buffer in packet_write_stream_with_flush_from_fd() and then fix the order here following your advice.

> Do you anticipate future need of non-gently variant of this
> function?  If so, perhaps a helper that takes a boolean "am I
> working for the gently variant?" may help share more code.

With helper you mean "an additional boolean parameter"? I don't see a need for a non-gently variant right now but I will add this parameter if you think it is a good idea. How would the signature look like?

int packet_write_gently(const int fd_out, const char *buf, size_t size, int gentle)

This would follow type_from_string_gently() in object.h.

Thanks, Lars