aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* util.poll: Lua 5.1 compatKim Alvefur2018-09-151-0/+6
|
* util.poll: Remove 'inline' from function declarationKim Alvefur2018-09-151-1/+1
| | | | Why does this fix building?
* util.poll: Test that it loads after being compiledKim Alvefur2018-09-151-0/+6
|
* MUC: Don't kick user of a reserved nick if it's theirs (thanks pep.)Matthew Wild2018-09-142-1/+66
|
* net.server_epoll: Use util.pollKim Alvefur2018-05-161-69/+74
|
* util.poll: Import poll library with epoll and select supportKim Alvefur2018-05-162-1/+425
|
* MUC: Include 'nick' attribute in affiliation lists (thanks jc)Matthew Wild2018-09-142-1/+18
|
* net.server_epoll: Make read_size a server-wide optionKim Alvefur2018-09-131-1/+2
| | | | Can still be overriden per connection
* net.server_epoll: Rename pattern to read_sizeKim Alvefur2018-09-131-10/+10
|
* net.server_epoll: Refactor Direct TLS assumptions outwardsKim Alvefur2018-09-131-8/+17
| | | | | | | | | | The assumption that connections are "Direct TLS" when a TLS context is supplided should be broken. The goal is to make it easy to add a new API that can be given a TLS context at creation even if it should do STARTTLS. With this commit, only the exposed server_select-compatible API assumes Direct TLS when a TLS context is included.
* util.stanza tests: Remove unused variable #luacheckMatthew Wild2018-09-131-2/+2
|
* util.stanza: Don't automatically generate ids for iq stanzasMatthew Wild2018-09-132-12/+18
| | | | | | | | | | Users of this API should provide their own id attribute that meets their uniqueness requirements. The current implementation leaks information (i.e. how many iq stanzas have been sent by the server to other JIDs). Providing any strong guarantees of randomness here would need to pull in additional dependencies that we don't want in this simple library.
* net.server_epoll: Factor Direct TLS handling out of connection initializationKim Alvefur2018-09-131-6/+6
| | | | | Eventually it should be possible to create a socket that has a TLS context but that does not automatically start TLS on connect.
* spec/scansion/basic_roster.scs: Add tests for actually modifying the rosterMatthew Wild2018-09-131-8/+40
|
* .luacheckrc: Remove aspirational goals from strict check, keep it simple for nowMatthew Wild2018-09-131-3/+0
|
* rostermanager: Use util.id to generate ids for roster pushedMatthew Wild2018-09-131-1/+3
|
* util.stanza tests: Fix test name (copy/paste error?)Matthew Wild2018-09-131-1/+1
|
* net.server_epoll: Remove coercion of connections to stringsKim Alvefur2018-09-131-8/+8
| | | | No longer needed since util.logger does this now
* net.server_epoll: Rename all variables holding TLS contexts for consistencyKim Alvefur2018-09-131-13/+13
|
* MUC: Fix to correctly return 'node' in disco#info responses (thanks jc)Matthew Wild2018-09-132-3/+17
|
* MUC: If a user with an existing affiliation registers, preserve that ↵Matthew Wild2018-09-132-1/+66
| | | | affiliation (thanks jc)
* spec/scansion/prosody.cfg.lua: daemonize by defaultMatthew Wild2018-09-121-0/+1
|
* spec/scansion/prosody.cfg.lua: Add pidfile option so Prosody can be daemonizedMatthew Wild2018-09-121-0/+2
|
* doc/net.server.lua: Not an actual source file, instruct luacheck to ignore itMatthew Wild2018-09-121-0/+2
|
* spec/scansion/prosody.cfg.lua: Config file to be used for integration testsMatthew Wild2018-09-121-0/+76
|
* mod_storage_memory: Import from prosody-modules 4c3230c22c18Matthew Wild2018-09-121-0/+178
|
* mod_auth_insecure: Fix module provider nameMatthew Wild2018-09-121-1/+1
|
* spec/scansion: Ensure all scripts have sensible titlesMatthew Wild2018-09-1112-1/+22
|
* spec/scansion: Multiple fixes to make all tests pass in latest ↵Matthew Wild2018-09-106-155/+9
| | | | (c35386e4b456) scansion
* GNUmakefile: Tell Mercurial to behave consistenlyKim Alvefur2018-09-091-1/+1
| | | | See `hg help environment.HGPLAIN`
* GNUmakefile: Allow overriding path to bustedKim Alvefur2018-09-091-1/+2
|
* GNUmakefile: Allow overriding path to luacheckKim Alvefur2018-09-091-1/+3
|
* mod_vcard_legacy: Add some notesKim Alvefur2018-09-091-0/+2
|
* mod_vcard_legacy: Fix EXTVAL conversion (thanks pep.)Kim Alvefur2018-09-091-1/+1
| | | | Good thing uri and url doesn't look similar at all!
* mod_vcard_legacy: Fix uppercase tag name (thanks pep.)Kim Alvefur2018-09-091-1/+1
| | | | Must have slipped on my shift key
* mod_vcard4: Advertise feature on account instead of hostKim Alvefur2018-09-081-1/+3
|
* util.pposix: Allow _DARWIN_C_SOURCE to be overridenKim Alvefur2018-09-081-0/+2
| | | | Like the others there
* Merge 0.10->trunkKim Alvefur2018-09-081-0/+3
|\
| * util.pposix: Fix or silence a warning on OS X (#1202)Kim Alvefur2018-09-081-0/+3
| |
* | Merge 0.10->trunkKim Alvefur2018-09-081-1/+1
|\|
| * util.pposix: Fix building on OS X (#1202)Kim Alvefur2018-08-181-1/+1
| |
* | util.pposix: Restrict use of malloc_info to glibcKim Alvefur2018-09-071-1/+1
| | | | | | | | Fixes build on musl based distros like Alpine linux
* | util.pposix: Do not attempt to assign to stdio variables [pedantic]Kim Alvefur2018-09-071-3/+3
| | | | | | | | | | | | Fixes build on gcc 6.4.0 See also https://stackoverflow.com/a/586416
* | mod_auth_insecure: Accept any username/password (import of mod_auth_any from ↵Matthew Wild2018-09-071-0/+53
| | | | | | | | prosody-modules)
* | mod_pubsub: Exclude value of FORM_TYPE from stored dataKim Alvefur2018-09-071-3/+3
| |
* | mod_vcard_legacy: Include avatar data even if metadata can't be loadedKim Alvefur2018-09-071-3/+4
| | | | | | | | | | | | | | | | | | Normally both nodes should have the same configuration and matching items, but we can't depend on it without having some code that enforces it, which does not exist at the time of this commit. Including the avatar itself should be prioritised. The image format can be derived from magic bytes.
* | mod_vcard_legacy: Return old vcard if neither vcard4 or avatar availableKim Alvefur2018-09-071-0/+4
| |
* | mod_vcard_legacy: Handle avatar without vcard4Kim Alvefur2018-09-071-4/+0
| | | | | | | | | | | | | | Since vcards are just avatar containers in many modern clients, aborting in case of no vcard4 data is not optimal. The upgrade mechanism needs further tweaks.
* | mod_vcard_legacy: Rename variable for consistencyKim Alvefur2018-09-071-2/+2
| | | | | | | | | | The variable holding the PEP service was `pep_service`, except in the XEP-0153 section.
* | mod_vcard_legacy: Return error in case saving avatar failsKim Alvefur2018-09-061-2/+7
| |