| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
| |
Co-authored-by: Joan Touzet <wohali@users.noreply.github.com>
|
|
|
|
| |
Port reduce_false.js and reduce_builtin.js to Elixir
|
| |
|
| |
|
| |
|
|\
| |
| | |
3.0.x backports from master
|
| |
| |
| | |
Fix create db options on secondary shard creation
|
| |
| |
| | |
Set cookie domain when DELETE'ing
|
| |
| |
| | |
Send correct seq values for filtered changes
|
|/
|
| |
Move "users_db_security_editable" to the correct location in the ini file
|
|
|
|
| |
Co-authored-by: Robert Newson <rnewson@apache.org>
|
|
|
|
|
|
| |
This copy slowed down the `erlang:spawn_link/3` call considerably.
Measurements in the wild showed the cost of that `spawn_link/3` going
from roughly 8 uS to 800 uS.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
* Port elixir proxyauth tests from js to elixir (#2660)
* Add support for specify a custom config file for CouchDB startup during testing
|
|
|
|
|
|
|
| |
elixir (3.0.x) (#2629)
* Port _design_docs tests, design_options and design_paths from js to elixir (#2596)
* Port changes.js test suite into elixir
|
| |
|
|\
| |
| | |
Port form submit test to elixir in 3.0.x
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously concurrent db creation requests could get past db_exists/1 and try
to create the shard document. The first one would succeed but the others would
fail with a `500 {"error":"error", "reason":"conflict"}` error instead of
`412`.
Handle the race condition by checking if db exists after getting a conflict
error and return a 412 response.
Fixes: #2613
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There's a race between the meck:wait call in setup and killing the
config_event process. Its possible that we could kill and restart the
config_event process after meck:wait returns, but before
gen_event:add_sup_handler is called. More likely, we could end up
killing the config_event gen_event process before its fully handled the
add_sup_handler message and linked the notifier pid.
This avoids the race by waiting for config_event to return that it has
processed the add_sup_handler message instead of relying on meck:wait
for the subscription call.
|
| |
|
| |
|
|
|
|
| |
This reverts commit afef1b2f5823f6269ecb4a845e286e9d0c0f4f35.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
After commit 27bb45043435828915bdcbdc130b685e5533bbd8 local _users is only
created the first time it is used. So in most case it's expected to not exist.
Update local_dbs to return it only if it is actually created.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Allows `configure.ps1` to correctly pull and build `rebar` on Windows
* Removes the static declarations in `rebar.config.script` on
specific, pre-determined paths to various includes/libraries
necessary for NIFs and external binaries (expectation is these are
passed in env vars INCLUDE, LIB and LIBPATH)
* fixes the SM60 `couchjs` build by telling `windows.h` not to
redefine min and max as macros through a `#define`
* fixes the `make eunit` target on Windows
* Adds the missing `EXE_LINK_CXX_TEMPLATE` that our rebar doesn't have,
but `enc` has today, which is also causing a failed `couchjs` (C++)
build on Windows
* Causes `make python-black` to correctly cause failure in `make check`
if it finds problems
* fixes Mango tests on Python 3.8 by bumping the hypothesis dependency
* fixes one Elixir test on Windows (incorrect calculation of `now(:ms)`
due to Erlang clock precision difference)
* a little bit of python black cleanup (mango tests)
|
| |
|
|
|
|
|
|
|
|
|
| |
Observed on FreeBSD Jenkins test runner:
```
function couch_replicator_clustering:setup/0 (src/couch_replicator_clustering.erl, line 257)
**error:{badmatch,{error,{already_started,<0.3165.0>}}} in module 'couch_replicator_clustering'
```
|
|
|
|
|
|
| |
This changes the couchjs OOM test so that it will trigger more reliably
on SpiderMonkey 60. It appears that newer SpiderMonkeys are better at
conserving memory usage which takes this test longer to trigger.
|
| |
|
|
|
|
| |
This fixes builds on Windows and CentOS.
|
|
|
|
|
|
|
|
|
|
| |
Since the backend port is closed by default, remove the cache service
since it was shown to cause monitor leaks, and it just consumes resources
running in the background when almost nobody uses it. The few utility scripts
that might use the backend port can just get the auth docs directly from the
db.
Issue: https://github.com/apache/couchdb/issues/2493
|
| |
|
| |
|
|\
| |
| | |
Fix highlight_post_tag parsing from json request body
|
| | |
|
|/
|
|
| |
Fixes #2520
|
| |
|
|
|
|
|
|
|
| |
The CouchDB API defines the warning field returned by _find to be
a string (and this is what Fauxton expects). 5d55e289 was missing
a string conversion and returned the warning(s) as an array. This
restores the intended behaviour.
|