Only notify the player initiating /load about the error and don't send
'Loading initiated by ...' and 'Saving initiated by ...' messages resulting
in two messages per /load and /save.
2245: Stop playing map sound when demo player is paused r=heinrich5991 a=def-
Noticed on Shooting Range map
2266: Reset hook & collision in prediction after kill r=heinrich5991 a=def-
Thanks to Pipou for reporting
Co-authored-by: def <dennis@felsin9.de>
2276: Fix segmentation fault when creating sql tables failed r=heinrich5991 a=Zwelf
`GameServer()->OnShutdown(true);` (which would wait on running threads to exit)
wasn't called if an error occured in the server init. Therefore `CServer` was
freed before the init-thread finished executing, causing a use after free from
`CSqlConnector::ms_ppSqlReadServers` pointing to `CServer->m_apSqlReadServers`.
Fixes#2271
Note that the server exits when creating tables failed. Since the tables are created by default, it has to be disabled explicitly (e.g. `add_sqlserver r teeworlds record teeworlds PASSWORD "localhost" 3306 0`) if the servers should start even when the database connection fails.
Co-authored-by: Zwelf <zwelf@strct.cc>
`GameServer()->OnShutdown(true);` (which would wait on running threads to exit)
wasn't called if an error occured in the server init. Therefore `CServer` was
freed before the init-thread finished executing, causing a use after free from
`CSqlConnector::ms_ppSqlReadServers` pointing to `CServer->m_apSqlReadServers`.
2247: Thread safe SQL interaction r=def- a=Zwelf
This PR intends to make the database interaction thread safe and optimizes some SQL queries. This is still a WIP, but since it is a rather big PR I wanted to get feedback early on.
Benefits:
* remove race conditions leading to undefined behavior and potential crashes
* logging game related database results in teehistorian would be possible
Behavior change:
* /top5team prints ranks in reverse order when passing a negative number (like /top5, /top5points)
* Optimize SQL statements for /rank /teamrank /top5 /top5team /points /top5points
* /load without parameters doesn't pass the SQL error to the user (as most other functions)
* Simplify IScore interface
* Add UUID to /save table (update of database schema necessary):
```
ALTER TABLE record_saves ADD SaveID varchar(64);
```
* /save immediately kills team and loads it again if the database insert fails.
still TBD:
* [x] saving (team) score when finishing
* [x] loading team save
* [x] loading initial time and birthday check
* [x] /map and random map votes
* [x] RFC: generate a passphrase (2-3 word) if save-code exists or no save-code is given making /save failures much more rare and save-codes more secure
* [x] clean up code (removing now unused structs, ordering of functions in IScore)
Co-authored-by: Zwelf <zwelf@strct.cc>
2270: Consider time_freq() r=def- a=Learath2
The timer ticks way faster on macOS leading to some insane pings :D
Co-authored-by: Learath <learath2@gmail.com>
2243: Workaround for client crash on predicted explosion r=def- a=trml
I haven't been able to find the exact source of the problem yet, but at least this seems to get rid of the segfault. The crash only happened with with recent versions of gcc and -O3 for me.
Co-authored-by: trml <trml@users.noreply.github.com>
2231: Don't clamp values that have no reasonable values r=Learath2 a=def-
Still affects many values in editor, for example colors in tiles and quads.
Thanks to Pipou for report.
Co-authored-by: def <dennis@felsin9.de>
2226: Fix image/sound selection in editor r=heinrich5991 a=def-
Follow-up to https://github.com/ddnet/ddnet/pull/2186 by @BannZay
Thanks to Cøke for notifying me
Co-authored-by: def <dennis@felsin9.de>
2202: Send DDNet version early in the connection process r=Learath2 a=heinrich5991
This gets rid of the problem that we don't know whether we should send
full snapshots to clients because they haven't told us about them being
DDNet yet.
Co-authored-by: heinrich5991 <heinrich5991@gmail.com>
2212: Use a consistent PRNG on all platforms r=Learath2 a=heinrich5991
I just used glibc's because that is what we use on all the servers. Based on https://www.mscs.dal.ca/~selinger/random/.
Behavior change: The PRNG is actually seeded now.
Co-authored-by: heinrich5991 <heinrich5991@gmail.com>
As discussed on Discord today, can be enabled inside of teams on the fly
during each run on any server. Finishes don't count. I haven't tested
save/load yet, would do that live on the server if this can be merged.
The following member variables:
* m_name to m_aName
* m_String to m_aString
* m_CpCurrent to m_aCpCurrent
* m_Switchers to m_pSwitchers
* SavedTees to m_pSavedTees
This came from a long discussion comparing PCG-* against xoroshiro*. Do
not generate integers without bias because it doesn't affect us very
much and it is easier to reimplement with modulo.
2208: Implement /load without params to show some info r=Learath2 a=def-
about your existing saves. Not the code of course, but only how many saves you have and when you last saved on this map.
Co-authored-by: def <dennis@felsin9.de>
src/game/editor/auto_map.h:47:9: warning: ‘*((void*)& NewConf +16)’ may be used uninitialized in this function [-Wmaybe-uninitialized]
I don't think it's an actual problem
This gets rid of the problem that we don't know whether we should send
full snapshots to clients because they haven't told us about them being
DDNet yet.
2191: write automapper version, was not initialized r=heinrich5991 a=Patiga
write automapper version, was not initialized
Co-authored-by: patiga <patigatus21@gmail.com>
2180: Don't count freshly loaded map as modified (fixes#2178) r=Learath2 a=def-
The reason is that we increased the sensitivity of the modified flag to
include adding layers and groups, and simply loading the map already
adds those, so every loaded map was considered modified.
Co-authored-by: def <dennis@felsin9.de>
The reason is that we increased the sensitivity of the modified flag to
include adding layers and groups, and simply loading the map already
adds those, so every loaded map was considered modified.
2161: Small prediction fixes/update r=def- a=trml
I looked through the latest changes in the server code, and made some updates to the prediction with it. This includes the hammer-out-of-freeze server fix and a small fix for stopper prediction. Also cleaned up some code, and fixed predicted smoke-trails slightly so they don't go through walls in certain cases.
Co-authored-by: trml <trml@users.noreply.github.com>
2164: Fix empty sound (fixes#2163) r=heinrich5991 a=def-
The problem was that the string is empty after going through the trim, thus the main displaying loop doesn't run, but the sound still does. Verified that it works.
Co-authored-by: def <dennis@felsin9.de>
Add libnotify support for linux client
Also unify notification management
Make libnotify mandatory for the client. It is installed on 100% of Arch
Linux systems and on 70% of Debian systems. I'd guess the remaining
Debian systems are servers.
Detect dependent libraries of `libnotify` using `pkg-config`. Remove
library-specific code from the game module. Decrement refcount for
libnotify notification object before leaving the function.
2119: Change last render logic, when inactive r=def- a=Jupeyy
Just a small fix to let the client not try to get high fps, when it "thinks" it didn't render for a long time.
(is still affected by the SDL check "if window has input" at the cl_refresh_rate calculation, so not a 100% fix).
Co-authored-by: Jupeyy <jupjopjap@gmail.com>
2104: Add spectate_closest now works when in Free-View r=def- a=d3fault
sry for the general refactor along with it xD
Co-authored-by: d3fault <d3fault@d3fault.net>
2103: Add spectate_closest r=def- a=d3fault
Spectates the tee closest to the one you are currently spectating.
This commit brought to you by COVID-19 lockdown boredom.
Fix#1793
Co-authored-by: d3fault <d3fault@d3fault.net>
2096: Fix the double-free of the antibot r=def- a=heinrich5991
I could not find an easy way to ensure that `AntibotDestroy` is only
being called once for each `AntibotInit` but still happening after all
the `CPlayer` destructors.
Co-authored-by: heinrich5991 <heinrich5991@gmail.com>
2078: Fix cl_showrecord with dummy (fixes#2077) r=heinrich5991 a=def-
Shows best time of either dummy or main player depending on which is selected
Co-authored-by: def <dennis@felsin9.de>
I could not find an easy way to ensure that `AntibotDestroy` is only
being called once for each `AntibotInit` but still happening after all
the `CPlayer` destructors.
2069: Move global configuration out of client interface r=heinrich5991 a=ChillerDragon
2076: Allow editing multiple tile layer props r=heinrich5991 a=Learath2
Requested by Ravie on discord
2095: Fix `AntibotDestroy` being called twice r=def- a=heinrich5991
The resource deallocation needs to happen in symmetry with the
allocation, otherwise we get unmatched dealloctions leading to crashes.
Allocations happened in `CAntibot::Init`, but it was not guaranteed that
this was called before the destructor.
Co-authored-by: ChillerDragon <ChillerDragon@gmail.com>
Co-authored-by: Learath <learath2@gmail.com>
Co-authored-by: heinrich5991 <heinrich5991@gmail.com>
2070: Fix demo recording with broken maps r=heinrich5991 a=Learath2
Apparently there are maps out there with broken headers advertising the wrong map size. This allows demos to be recorded on those.
Example of such a map is `ctf4_old_d668e9fa_2f472051b26b6bffaa4af8990cf882cafd6364e00e1333b77762cb9911e49464.map`
Hosted on `Allround Network`
Reported by teini94 on discord
Co-authored-by: Learath <learath2@gmail.com>
2071: Add an option to only start a new demo when connect while automatically record demos r=def- a=sirius1242
Origin auto demo recording of ddnet will generate lots of demo files especially when playing in mods which have shorter rounds, such as infclass. I added a new option to open a new demo file only when connect to server (include map changing), this is disabled by default, and can use `cl_auto_demo_on_connect 1` to open. It works only when `cl_auto_demo_record` open.
Co-authored-by: sirius <sirius.ustc@gmail.com>
The resource deallocation needs to happen in symmetry with the
allocation, otherwise we get unmatched dealloctions leading to crashes.
Allocations happened in `CAntibot::Init`, but it was not guaranteed that
this was called before the destructor.