Age | Commit message (Collapse) | Author | Files | Lines |
|
For consistency with `pc->addfame()`, the argument to pc->famerank() is
now an enum fame_list_type.
The function was renamed to `pc->fame_rank()` to avoid silently
compiling old non-compliant code.
Signed-off-by: Haru <haru@dotalux.com>
|
|
The function now takes the rank type as argument, rather than guessing
it from the character's class. If the wrong fame point type for the
current character is requested, the request is ignored.
This fixes some (unofficial) edge cases where a Taekwon or an Alchemist
refined a signed item, they could obtain rank points.
Signed-off-by: Haru <haru@dotalux.com>
|
|
This commit ensures that `sd->job` and related variables only contain
MapIDs (i.e. the serverside optimized values, where High Novice is
MAPID_NOVICE | JOBL_UPPER) and are never checked against the
client-based Job Class values.
As a rule of thumb, from now on, when a variable is named `job` or
`jobid`, it is intended to contain a MapID and not a Job Class ID. The
type of such variable shall be an unsigned `uint16` or `uint32`.
To ensure that related third party code is also verified when this
commit is merged, the variable `struct map_session_data::class_ (i.e.
`sd->class_`) is renamed to `job`.
Some issues in related lines are also fixed, including:
- The atcommand `@mount` would not check properly that the requesting
character is a 2nd class Swordsman type.
- `pc->addfame()` would silently accept invalid MapIDs, sending unknown
values to the client (more on this in a subsequent commit).
- `pc->famerank()` would not use the passed job as a bitmask, causing
the caller to have to mask it beforehand (more on this in a
subsequent commit).
- The Soul Linker check in TK_JUMPKICK wasn't future-proof (no harm
caused currently).
- Gunslingers would be able to be targeted by Spiritual Bestowment
(`MO_KITRANSLATION`) and Spiritual Sphere Absorption
(`MO_ABSORBSPIRITS`) due to a faulty check introduced when the
Rebellion class was implemented (causing unintended interaction with
Coins).
Signed-off-by: Haru <haru@dotalux.com>
|
|
This commit ensures that `sd->status.class` and related variables only
contain Job Classes (i.e. the client-compatible values, where High
Novice 4001) and are never checked against the bitmask-based MapID
values.
As a rule of thumb, from now on, when a variable is named `class`, it
is intended to contain a Job Class ID and not a MapID. The type of such
variable shall be a signed `int16` or `int`.
To ensure that related third party code is also verified when this
commit is merged, the variable `struct mmo_charstatus::class_ (i.e.
`sd->status.class_`) is renamed to `class`.
Some issues in related lines are also fixed, including:
- A wrong check in the char server would prevent the correct detection
of babies in code related to the family exp sharing.
- Baby Arch Bishops would not be affected by Eucharistica.
- A wrong check would cause the `questinfo()` script command not to
display its information for most classes (except 1-1 classes).
- Map IDs and Job Classes were mixed up in `itemdb_jobid2mapid()` and
`itemdb_jobmask2mapid()` for 1-1 classes (causing currently no harm,
since they just happen to coincide).
- The Baby Sura class would not cause parties to be marked as
containing a monk type character (for SLS Team Up purposes).
- Baby Geneticists would bypass the cart check when trying to equip
ammunitions.
- Baby Mechanics would bypass the Mado Gear check when trying to equip
ammunitions.
- Transcendent Shadow Chasers would lose the Stalkers' ability to clone
transcendent skills.
Signed-off-by: Haru <haru@dotalux.com>
|
|
The enum values are now written in a more explicit way, showing how
they're composed (upper mask + base job).
A future-proof static assertion is added, to make sure the base jobs
won't overlap the upper masks.
Signed-off-by: Haru <haru@dotalux.com>
|
|
This introduces the macro `STATIC_ASSERT(ex, msg)`, that works like its
C11 counterpart `_Static_assert(ex, msg)`, on compilers that support it.
Support is provided, depending on the compiler:
- When in C11 mode, use the native `_Static_assert`
- If the compiler advertises having the `c_static_assert` feature, use
`_Static_assert` (according to the clang documentation).
- If the compiler is GCC >= 4.7 (tested to be compatible), use
`_Static_assert` (nonstandard extension, C11 feature available in C99).
- If the compiler is MSVC (all the currently supported versions have
been tested to be compatible), then use `static_assert` (nonstandard
extension, C++ feature)
- Otherwise, define it as a no-op macro.
Signed-off-by: Haru <haru@dotalux.com>
|
|
Signed-off-by: Haru <haru@dotalux.com>
|
|
Add *makeitem2 script command.
|
|
Fixed areawarp cell check
|
|
Update party_check_state
|
|
The script command areawarp has been checking the wrong target map for ages. It has been checking the mapindex rather than the mapid and therefore returned a wrong result.
Original commit: rathena/rathena@ac2ba09
|
|
The apostrophe (`'`) is no longer an allowed symbol in libconfig key
names. It is no longer necessary since it is no longer allowed in
constant names.
Signed-off-by: Haru <haru@dotalux.com>
|
|
Supporting apostrophes (`'`) inside constants or variables is no longer
necessary thanks to the more strict AegisName ruleset we're using now.
Signed-off-by: Haru <haru@dotalux.com>
|
|
Signed-off-by: Haru <haru@dotalux.com>
|
|
Some missing items were copied to the pre-renewal database. In general,
when an item ID is hardcoded in source, it's preferable that the item
exists in both databases.
Signed-off-by: Haru <haru@dotalux.com>
|
|
The config_t objects need to be initialized through libconfig->init()
before use (except when loaded through libconfig->load_file(), which
initializes the object).
Signed-off-by: Haru <haru@dotalux.com>
|
|
This also fixed a typo (misplaced parentheses around LIBCONFIG_OBJ) in
the *_DEPENDS variable of makefiles
Signed-off-by: Haru <haru@dotalux.com>
|
|
The library common.a depends on common_sql.a.
Signed-off-by: Haru <haru@dotalux.com>
|
|
|
|
As we use int64 (instead of long long int) through the code, it's
preferable if our libconfig interface does the same, avoiding us to
create temporary interchange variables for that purpose.
Signed-off-by: Haru <haru@dotalux.com>
|
|
|
|
The issue was caused by an evident patch mishap in commit
5e04f472e1ef618d9201fd22e382dca829992b3b. The misplaced piece of code
has now been moved to the most likely place it was supposed to be in.
This commit also removes an extra 'break' introduced by accident in
commit f4e8189545642ee0bf87f3da546bddc918682935 (which caused no harm).
Related: http://herc.ws/board/topic/13672-heaven-driver/
Signed-off-by: Haru <haru@dotalux.com>
|
|
|
|
|
|
|
|
into packets.h
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fixed copyright year for 2016
|
|
|
|
|
|
- missing classes checking.
|
|
- Fixes #1483
Signed-off-by: Haru <haru@dotalux.com>
|
|
b0006317e5cc4a48cd5b8a501b94813cd3b93670
- Their original value was `17` (0x11), which is equivalent to
`SCFLAG_NOAVOID|SCFLAG_NOICON`, rather than `SCFLAG_ALL` (0x1f)
Signed-off-by: Haru <haru@dotalux.com>
|