This is used to provide default implementations for createConfigWriteStream
and createConfigReadStream, which can be used by most backends.
Note that backends can still override createConfigRead/WriteStream;
this could be useful if settings on some port are not stored in a
regular file (think 'Windows registry', for a hypothetical example).
This fixes linker problems with the unit tests on Mac OS X. it is also
"more proper" to use the high-level FSNode API. StdioStream is a
(relatively) low-level wrapper intended for use inside of backends only.
This should assure vsnprintf and snprintf will now also always null
terminate the result even for MSVC.
Currently the functions are placed in scummsys.h, but that causes us
to include two standard C library headers there (for MSVC at least).
This is not particulary nice, so we should think of a better solution
here.
This includes a FIXME, since _snprintf behaves differently to snprintf.
Not only in the return value (which is a minor difference, since we usually
do not use it), but also since it does not always include a terminating null.
Since we only want to disable a MSVC specific warning with it and other
compilers might have different warnings numbers it is safer to only target
MSVC here.
Instead of defining a hacked up Common::Error code, split the return
value of processSettings into two parts: An error code, and a value
which indicates whether the specified command was completely handled
by processSettings or not.
This also removes the dependency of engines on the event recorder header
and API, and will make it easier to RandomSources that are not properly
registered.
We also change how alloc failures are handled: Instead of using
assert(), which is usually disabled in release builds on various
platforms, we now *always* catch this situation and invoke error() if
necessary.
In addition, we use them if in GCC >= 3.3 if unaligned access is
possible. The GCC variants of these macros also contain protection
against overzealous compilers' static aliasing optimizations.
SafeSubReadStream is basically a SeekableSubReadStream that
re-seek()s the parent stream before each read(). That way, more than
one SafeSubReadStream of the same parent stream can be used safely
at the same time, at the cost of increasing IO.
Right now, a few places in the frontend code still use printf and
consorts. We mark the affected files with a FIXME for now, and
add a dedicated exception for each. To be fixed!
Also tweak FORBIDDEN_SYMBOL_REPLACEMENT to hopefully really always
enforce a compiler error
- Add check in listMembers and skip files that can't be enumerated.
- Add checks for all function calls in createReadStreamForMember (and no longer return a stream from an uninitialized buffer when the file cannot be read).
Instead I moved the call to ReadStream::eos into SubReadStream::eos.
fuzzie pointed out that in certain cases this resulted in a lot of eos calls
before. Those were rather needlessly, since checking the eos bit of the
decorated stream in read does not give any advantage over checking it when
the client code really wants to test the eos flag.
MKID_BE relied on unspecified behavior of the C++ compiler,
and as such was always a bit unsafe. The new MKTAG macro
is slightly less elegant, but does no longer depend on the
behavior of the compiler.
Inspired by FFmpeg, which has an almost identical macro.
I moved this TODO to the Wiki TODO page and extended it there. Yet
while I originally wrote it, I now think that we are probably best off
keeping things as they are.
Since various apps can modify the type/creator of the files, we can't rely on the header of the resource fork to take the md5. I've therefore also recalculated all of the detector entries (all 5 of them) that use the Mac resource fork code.
svn-id: r55764
Currently there is no iterator returned from this method, to have some
similarity to associative containers of the STL.
I also "added" one unit test for this method, which is basically just
a copy of the HashMap::erase(const Key &) test with the required adaptions.
svn-id: r55661
The way we determine if a domain is a game domain or not when loading from the config file is by checking for a 'gameId' key. This required changing the method of loading somewhat. MiscDomains (ie. non-game domains) are saved to the config file before game domains but after the main domains.
This change shouldn't mess anything up but it could probably use a look-through just in case.
svn-id: r55060
- Broaden the range of language tags that could be accepted as fallback in
auto-detection. For example, for a system with locale 'pt_PT', all of
these would be accepted as valid fallbacks: 'pt_PT_*', 'pt', 'pt_*'
- Make language codes matching case-insensitive as specified in RFC 4647.
- Add support language code blocks longer than 2 characters.
svn-id: r54715
* Get rid of obsolete comment about external documentation
(link was broken for at least one year)
* Remove unnecessary Common:: prefix
* Push #include stream.h from .h to .cpp
svn-id: r54354
This is the only we need right now, and it saves a few bytes per
instance. The template approach I used before has the drawback that it
increases the binary size, which negates the benefit. Thanks to LordHoto
for pointing this out.
svn-id: r54344
This commits a slightly modified patch from my patch tracker item #3104630
"OSYSTEM: Add logging API as proposed by Max on -devel".
I was not able to test compilation on Android and SamsungTV, since there is no
toolchain for those on buildbot (or I was too blind to find them).
svn-id: r54339
The assert() macro may be compiled to be empty. In that case, its
arguments are *NOT* evaluated. Hence, things like
assert(doSomething())
must not be used whenever doSomething() has important side effects.
Also document BufferedWriteStream::flushBuffer() and explain why it
exists parallel to BufferedWriteStream::flush().
svn-id: r54322
This makes it possible to write
DECLARE_SINGLETON(foo);
instead of
DECLARE_SINGLETON(foo)
without causing a warning about an extra semicolon.
The extra semicolon helps some editors at parsing the C++ code.
svn-id: r54258
One-plugin-at-a-time can have fragmentation caused by the ConfigManager if a game changes any configuration value. By reallocating and copying over the ConfigManager, we avoid this problem.
svn-id: r54243
Revision 54206 broke translations support because it uncovered a bug in
the TranslationManager when reading the translations.dat file. In that file
all the stored string lengths include the terminating 0 but Common::String
expect a length without the terminating 0. Therefore all String objects
created from reading the translations.dat file had an incorrect size. This
caused the font file names to be wrong after adding the charset.
svn-id: r54235
This commit contains the AudioCDManager changes from the gsoc2010-opengl
branch. The other changes in that branch are restricted to the backends
directory only (plus configure).
The Nintendo DS and Dreamcast ports still need to be ported over to
the new Audio CD system, but that should be fairly easy to do.
svn-id: r54147
* names now comply to our naming conventions
* the function computeStreamMD5AsString which computes the MD5
as a hex string now returns it as a Common::String
* add doxygen comments
svn-id: r54121
The deletePointer() method approach cannot work, as it is called
by the destructor of the base class.
A possible correct solution would be to enhance ScopedPtr with a
"deleter" object like SharedPtr. But this seems overkill as long as we
need it in only one place.
svn-id: r54057
This merge was extremely difficult to carry out. It wasn't entirely SVN's fault -- there were several merges to the branch that were done by hand. Please check for any issues and regressions. Also note that the DS makefile was not copied over since the "one at a time" plugin mode currently has too much fragmentation ie. it doesn't work.
svn-id: r54051
Rational: hexdump() is used for debug output. An even better alternative
might be to change it to return a string, instead of printing anything.
This way, it could be used inside e.g. GUI debug consoles.
This is left as an exercise to the interested developer :).
svn-id: r54010
This is a first step towards getting rid of all uses of regular printf,
fprintf, vprintf, vfprintf, puts, fputs, etc. in our codebase.
The name format() reflects the purpose of the function, and parallels
String.format() in Java, boost::format, and others.
svn-id: r54004
A new header file common/forbidden.h is included by scummsys.h and it
re-#defines numerous symbols like fopen(), fread(), system(), etc. with
garbage, in order to provoke compile errors in any code using them.
If a .cpp file really *must* use any of these (e.g. because it is a
backend file), then these redefinitions can be disabled by #defining
FORBIDDEN_SYMBOL_ALLOW_ALL as the first thing in the .cpp file. Whenever
this is done, an explanatory comment should be added.
Note that this system cannot catch all "bad" usages (notably the Lua
code in the sword25 engine), as it can only work if scummsys.h is
included.
svn-id: r53961
Almost all places where we used XMLParser::parseIntegerKey were using
it like this:
XMLParser::parseIntegerKey(str.c_str(), ...)
Since this makes the code harder to read, I overloaded the method to
also accept Commmon::String directly.
Also removed all .c_str() invocations where necessary.
svn-id: r53479
This includes an rather hacky attempt to merge all the recent gp2x backend
changes into the branch. I suppose the gp2x backend and probably all new
backends, i.e. gph, dingux etc., might not compile anymore.
Since I have no way of testing those it would be nice if porters could look
into getting those up to speed in this branch.
svn-id: r53399
The return value of vsnprintf when the provided buffer is not large
enough to hold the formatted string is implementation-dependent:
C99: The size the formatted string would take up.
MSVC: -1, with no indication of how large the buffer should be.
IRIX: The number of characters actually written, which is at most
the size of the buffer minus one, as the string is truncated
to fit. This means the only way to be sure the entire string
is written is if the return value is less than the capacity - 1.
This change means that whenever we try to format a string where the size
is 1 below the built-in capacity, that the capacity will be needlessly increased.
If this turns out to be problematic we could make this behaviour __sgi conditional.
svn-id: r53143
In SCI01 and up, each typed word may be interpreted as multiple
class,group pairs. This patch adds support to the vocabulary and
parser. It uses the matcher support added in r52985.
This fixes parser issues in German LSL3, but needs testing.
svn-id: r52989
When switching to a theme that do not have the fonts needed to properly
display the current language, it now revert to the previously used theme
and display an error message telling the user to change the language
first if he wants to use the theme he selected.
svn-id: r52969
The advantage is that we get to do inlining and even use lwl and lwr instructions where appropriate. We have to do it ourselves because the PSP doesn't tolerate built-in instructions, but also we have a more efficient memcpy than the lib's.
svn-id: r52817
Resource names can be the same across types (ie. 'PICT' "Foo" and 'snd ' "Foo"), so this ensures that the right type is chosen in those cases.
svn-id: r52756
This is a better solution for the gcc aliasing problem that happens when aliasing a struct onto something else. What happens is that the compiler assumes no aliasing can happen when -O2 and -O3 are activated, and a call to READ_UINT32() followed by WRITE_UINT32() and another READ_UINT32() will be optimized to return the original read value instead of re-reading.
svn-id: r52480
Currently static strings are marked with the _s() macro so that xgettext knows
the string needs to be translated. But there was no way to tell xgettext that
the string could get translated in different contexts. This is the role of this new
macro.
svn-id: r52460
If the translations.dat file is not found using the SearchMan, it now looks
for that file into the themepath directory and sub-directories (if the path is
defined).
svn-id: r52428
I need the write buffering for the new version of the PSP streams and thought the simplest way to implement it would be along the lines of BufferedReadStream. Sadly, I found a nasty little bug in BRS but that's taken care of.
Also, I adapted these streams for target-specific memory allocation by using virtual functions for allocation/deallocation.
svn-id: r52325
This change means there can now be different translations for the same
english string depending on the context. It is based on gettext msgctxt feature.
There is a new macro _c(msg, ctxt) that should be used instead of _(msg) in
the source code when we want to add a context to the message. For the
moment I have added contexts to only one message ("None") so that I could
test the changes. Context could be added also to get shorter translations when
GUI is in 1x mode.
I have also added back the fuzzy option to msmerge since it is useful when
adding contexts to populate the translations for the new contexts.
svn-id: r52308
Previous commit of translation.cpp introduced a defined but not
implemented operator when the translation support was disabled.
This was probably fine since it was not used, but not very clean.
svn-id: r52241
Based on patch #3048388: "Hash map for ZipArchive class", though
I made the hashing by default and removed the old code.
It is a really good thing and will speed up any archive operations.
svn-id: r52204
Prior to this change lastPathComponent would not create a correct result,
when the input of lastPathComponent did not contain a single separator.
I also added a test case for this in our unit tests.
svn-id: r52123
It's now a static function since it doesn't require any local variables and is really separate from the resource fork code. Also, actually assigning the keycolor (which was just being ignored).
svn-id: r52075
This fix a regression introduced in revision 50448 ("Add support for
locale-dependant font'). When disabling localization a function was
declared, called but not implemented.
svn-id: r52048
