History log of /1.7.1/memcachetest/ (Results 1 - 25 of 79)
Revision (<<< Hide revision tags) (Show revision tags >>>)Date Author Comments
(<<< Hide modified files)
(Show modified files >>>)
Revision tags: 0.8.3
88ae3b3f27-Apr-2011 Matt Ingenthron <ingenthr@cep.net>

Remove pointless assertion check.

Checking an assertion on an enum generates a warning. I don't think
the compiler should be warning here since it's an assertion, but
it does and se

Remove pointless assertion check.

Checking an assertion on an enum generates a warning. I don't think
the compiler should be warning here since it's an assertion, but
it does and seeing the warning is annoying.

show more ...

4fcfa25927-Apr-2011 Matt Ingenthron <ingenthr@cep.net>

Change variable name so as to not shadow a global from time.h.

7a862ca427-Apr-2011 Matt Ingenthron <ingenthr@cep.net>

Formatting should be correct for size_t.

Revision tags: 0.8.2
d810452505-Jan-2011 Matt Ingenthron <ingenthr@cep.net>

Re-add static to thread entry, only used from main.

e32e6c1f05-Jan-2011 Matt Ingenthron <ingenthr@cep.net>

Use size_t consistently for total aggregation.

011f385c05-Jan-2011 Matt Ingenthron <ingenthr@cep.net>

Remove old error reporting method.

ac9c2a2f06-Dec-2010 Aliaksey Kandratsenka <alk@tut.by>

fixed off-by-one in textual_seterrmsg

d269bf1806-Dec-2010 Aliaksey Kandratsenka <alk@tut.by>

increased connection buffer to 1M + 256

This is necesssary so that ascii gets of maximal size work

711139bb06-Dec-2010 Aliaksey Kandratsenka <alk@tut.by>

Fixed gcc warning.

The original patch had converted chunk to int, but after
checking about, it seems better to keep that a size_t and simply
remove the typedef.

This is more

Fixed gcc warning.

The original patch had converted chunk to int, but after
checking about, it seems better to keep that a size_t and simply
remove the typedef.

This is more c99, and also requires the addition of stddef.h.
http://stackoverflow.com/questions/3597743/where-is-ptrdiff-t-defined-in-c

show more ...

e7bc7eb906-Dec-2010 Aliaksey Kandratsenka <alk@tut.by>

fixed mistake causing compile warnings to be disabled

1416b3b107-Dec-2010 Matt Ingenthron <ingenthr@cep.net>

Conditionally include binary error text strings.

The whole world isn't binary, and we support that elsewhere so
we should here.

Revision tags: 0.8.1
d778856603-Dec-2010 Matt Ingenthron <ingenthr@cep.net>

Give more info about the failure when setting.

d8a6d24703-Dec-2010 Matt Ingenthron <ingenthr@cep.net>

Be verbose on sets too, when required.

1441b92130-Nov-2010 Matt Ingenthron <ingenthr@cep.net>

If the server disconnects us, let someone know.

3865f32030-Nov-2010 Matt Ingenthron <ingenthr@cep.net>

Correct and portable error string handling.

2ffa4ada30-Nov-2010 Matt Ingenthron <ingenthr@cep.net>

Support for continuing when SERVER_ERROR on get.

Revision tags: 0.8.0
1a01b7be19-Nov-2010 John David Duncan <jdd@sam-by-land.lenoxway.net>

Portability fixes for FreeBSD and Mac OS X.

FreeBSD requires include of <netinet/in.h>
Mac gettimeofday() has problems; use a non-standard monotonic native
clock instead.

5b5edbe715-Nov-2010 Trond Norbye <trond.norbye@gmail.com>

Add support for key prefix

Change-Id: I2af7cd09f89bfd919721118ecfc6a94763bf9407

6883df7b16-Nov-2010 Aliaksey Kandratsenka <alk@tut.by>

return copy of error message from libmemc_get_error

560f1c7210-Oct-2010 Trond Norbye <trond.norbye@gmail.com>

Refactor: change the get wrapper to return true/false

792b4af810-Oct-2010 Trond Norbye <trond.norbye@gmail.com>

Increase error information if set fails

4551b65312-Oct-2010 Matt Ingenthron <ingenthr@cep.net>

Data isn't missing if it's size should be 0.

Revision tags: 0.7.0
3d87e1f826-Sep-2010 Matt Ingenthron <ingenthr@cep.net>

Added 99th percentile latency.

8dcf54b230-Aug-2010 Matt Ingenthron <ingenthr@cep.net>

Support the membase backoff response value.

When membase is under memory pressure and has not yet migrated
enough things to disk to accept new incoming ones, it will send
either a bi

Support the membase backoff response value.

When membase is under memory pressure and has not yet migrated
enough things to disk to accept new incoming ones, it will send
either a binary or an ascii response telling the client that
there is a temporary failure to accept the store. At this time
it's appropriate to backoff for a period of time.

This implements a very simple backoff. A better one would have
some randomness for the interval. This gets the job done.

show more ...

4bb89df907-Sep-2010 Matt Ingenthron <ingenthr@cep.net>

Fixes for make dist, need header file.

1234