aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* | | net.server_epoll: Use method to update peername on connectKim Alvefur2018-10-291-3/+1
| | |
* | | net.server_epoll: Reschedule delayed timers relative to current timeKim Alvefur2018-10-281-3/+7
| | | | | | | | | | | | | | | This should normally never happen, but can be reproduced by suspending the process a while.
* | | Merge 0.11->trunkKim Alvefur2018-10-283-418/+4
|\| |
| * | mod_storage_sql1: Remove pre-0.10 SQL moduleKim Alvefur2018-10-281-414/+0
| | |
| * | util.sslconfig: Recognise TLS 1.3 as a protocol versionKim Alvefur2018-10-281-1/+1
| | | | | | | | | | | | This enables it to understand protocol = "tlsv1_3+"
| * | CHANGES: Update with new branchKim Alvefur2018-10-281-3/+3
| | |
| * | Diverge 0.11Kim Alvefur2018-10-280-0/+0
| | |
* | | net.server: Add an API for holding writes of outgoing dataKim Alvefur2018-10-253-14/+50
| | |
* | | net.server_select: Still allow buffering outgoing data on write-locked ↵Kim Alvefur2018-10-281-4/+1
| | | | | | | | | | | | connections
* | | net.server_select: Replace use of deprecated :lock_read in server.linkKim Alvefur2018-10-281-2/+2
| | |
* | | net.server_select: Deprecate :lock methodKim Alvefur2018-10-281-0/+1
| | | | | | | | | | | | Exists only in server_select and I found nothing using it
* | | server_select: Fix :lock methodKim Alvefur2018-10-281-1/+1
| | | | | | | | | | | | | | | | | | This always unlocks reading. I don't believe this is used anywhere. server_event does not implement this.
* | | net.server_select: Move code from :lock_read into :pause and :resumeKim Alvefur2018-10-281-13/+15
| | |
* | | net.server_event: Deprecate :lock_read here tooKim Alvefur2018-10-281-0/+1
| | |
* | | net.server_select: Deprecate connection:lock_read() methodKim Alvefur2018-10-281-1/+1
|/ /
* | net.server_epoll: Shuffle variables earlier to improve readabilityKim Alvefur2018-10-271-1/+1
| |
* | mod_ping, mod_uptime: Use module:providesKim Alvefur2018-10-272-2/+2
| |
* | mod_adhoc: Add compat marker for older handling of adhoc itemsKim Alvefur2018-10-271-1/+1
| | | | | | | | | | | | | | Most modules use the module:provides() API which add the '-provides' suffix. Previously module:add_item() was used. See 58714123f600
* | util.serialization: Encode non-fatal error in way that can be restoredKim Alvefur2018-10-271-1/+1
| | | | | | | | It could previously encode to eg `{ [nil] = ... }` which doesn't get decoded
* | util.serialization: Rename non-fatal fallback handler for clarityKim Alvefur2018-10-271-3/+3
| |
* | util.serialization: Disable use of unquoted table keys by defaultKim Alvefur2018-10-271-1/+3
| | | | | | | | For safety against future new keywords
* | util.serialization: Add option for allowing multiple references to the same ↵Kim Alvefur2018-10-272-1/+25
| | | | | | | | table (but not cycles)
* | util.serialization: Test rejection of multiple references to same tableKim Alvefur2018-10-271-0/+8
| |
* | util.serialization: Separate errors for multiple table references and max depthKim Alvefur2018-10-271-2/+5
| |
* | util.serialization: Make maximum table depth configurableKim Alvefur2018-10-271-1/+2
| |
* | modulemanager: Fix issues introduced in previous commit acf74ad0b795 [thanks ↵Matthew Wild2018-10-261-7/+4
| | | | | | | | luacheck, scansion]
* | Many things: switch from hacky multi-arg xpcall implementations to a ↵Matthew Wild2018-10-267-31/+23
| | | | | | | | standard util.xpcall
* | util.xpcall, util.compat: Add non-hacky multi-argument xpcall() for Lua 5.1Matthew Wild2018-10-263-1/+40
| |
* | util.dependencies: Remove ztact compatability warningMatthew Wild2018-10-261-13/+0
| |
* | util.promise: Switch from pcall to xpcall to get tracebacks on exceptionsMatthew Wild2018-10-261-1/+1
| |
* | util.promise: Remove the non-standard ability to pass a promise to reject()Matthew Wild2018-10-261-3/+1
| |
* | util.promise: Add tests ensuring returning a promise resolves the current ↵Matthew Wild2018-10-261-0/+49
| | | | | | | | promise with that promise
* | moduleapi: Prevent loading disabled module as dependency of enabled oneKim Alvefur2018-10-251-3/+3
| | | | | | | | Explicitly disabled module should stay disabled.
* | mod_mam: Remove embedded fallback archive driver (fixes #972)Kim Alvefur2018-10-252-92/+0
| | | | | | | | It was equivalent to mod_storage_memory, which is included as a separate module now.
* | mod_mam: Upgrade case of invalid archive store driver to hard errorKim Alvefur2018-10-251-5/+2
| |
* | mod_mam: Ignore case of null storage driverKim Alvefur2018-10-251-7/+3
| |
* | mod_muc_mam: Fix missing form container elementKim Alvefur2018-10-251-1/+1
| |
* | util.promise tests: Fix declared but unused variablesMatthew Wild2018-10-251-2/+2
| |
* | util.promise: Fix missing parametersMatthew Wild2018-10-252-4/+30
| |
* | util.promise: Ensure chained promises always receive a value/rejection even ↵Matthew Wild2018-10-252-22/+60
| | | | | | | | if an intermediate promise has no handlers
* | util.promise: Export is_promise()Matthew Wild2018-10-251-0/+1
| |
* | MUC: Remove debug logMatthew Wild2018-10-251-1/+0
| |
* | util.promise: Remove debug print and assertMatthew Wild2018-10-251-1/+0
| |
* | Merge 0.10->trunkKim Alvefur2018-10-241-2/+6
|\|
| * prosodyctl: Change ownership of certs to same as the cert base dir when ↵Kim Alvefur2018-10-241-1/+4
| | | | | | | | running from a source checkout
| * prosodyctl: Allow cert dir to not be owned by root (fixes #1075)Kim Alvefur2018-10-241-1/+2
| |
* | mod_pep: Add automatic subscriptions on node creation (fixes #1222)Kim Alvefur2018-10-201-1/+21
| | | | | | | | | | Replaces automatic node creation on subscription, since this casues problems when trying to keep nodes private or other specific settings.
* | mod_pep: Add PEP service owners username in configuration to ease identificationKim Alvefur2018-10-201-0/+1
| |
* | mod_pep: Remove use of recipients table for signaling pending disco#itemsKim Alvefur2018-10-231-16/+7
| | | | | | | | | | | | | | | | | | | | It stored the caps hash as a string in the recipients table while waiting for a disco#info response, then replaces it with the set of nodes that had +notify. This mixing of types is bad, so here it gets removed. This can be improved once an IQ tracking API (#714) is added.
* | util.pubsub: Clarify comment about return valueKim Alvefur2018-10-211-1/+1
| |