This is used to differentiate LOG_ERROR (which is a *non* fatal error, but a error none the less) and fatal errors which result in program termination with no *visible* output (on windows/macs) on our end because of the crappy GUI functions so we usually dump it to stderr.txt and there is no clue that the program had to exit for whatever reason.
Hopefully, this will be helpful to windows people so they can differentiate between *our* exit/abort routines, and crashes that occur in drivers which look exactly the same since they get dumped back to the desktop with no visible clue why.
LOG_FATAL is *always* on in both debug & release builds.
On LOG_FATAL debug lines, on windows, we now throw up a modal dialog box with the current error message. At this time, this only happens when we have a error, and we use abort() right after the error, which makes it fatal.
That explains why this touches ~60 files :)
git-svn-id: https://warzone2100.svn.sourceforge.net/svnroot/warzone2100/trunk@8320 4a71c877-e1ca-e34f-864e-861f7616d084
* At the only place where these ID numbers where used (scrValDefSave and scrValDefLoad), use the key-string instead (e.g. "ZNULLBODY" instead of 17)
* Rename function strresGetStringByID to strresGetString
* Increment event save file version to 4 (version of savegame.es files)
* Reject loading of savegames that need to load non-NULL ST_TEXTSTRING variables from version < 4
NOTE: This may break current savegames of games that use scripts with variables of type TEXTSTRING in them (AFAIK currently only ''some'' campaign and tutorial savegames have these).
This closes ticket:19
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5685 4a71c877-e1ca-e34f-864e-861f7616d084
* Don't maintain a secondary linked list of strings in strres.c
* Add the string that's tracked and it's ID number to the TREAP_NODE structure
* Let the treap code manage its own memory
* Have the treap code perform reverse lookups using either the resource string (i.e. not the "key" string) or associated ID number
NOTE: Searching a treap for anything else than the key is O(n), but that's no worse than a linked list, which is O(n) as well. Walking a treap has O(log n) as memory requirement, whereas walking a linked list has O(1) as memory requirement (stack size). So the only disadvantage is that searching the treap has a (highly remote!) possibility of stack overflows.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5666 4a71c877-e1ca-e34f-864e-861f7616d084
* Replace use of the forwarding macro TREAP_ADD with a direct call to treapAdd
* Remove treap debugging code (actually that's what the above changes effectively accomplished)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5627 4a71c877-e1ca-e34f-864e-861f7616d084
* Replace use of the forwarding macro TREAP_CREATE with a direct call to treapCreate
* Return the TREAP* pointer directly from treapCreate and use NULL as an indicator for failure
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5626 4a71c877-e1ca-e34f-864e-861f7616d084
Because the treap management code is only used by the code for string resources, make it specific for these string resources and don't use casting back & forth from void* all over the place. This simplifies the interaction of string resource code with treap code.
This will allow an easier cleanup and eventually a removal of string resource code. As the string resource code can eventually be replaced with gettext code instead.
This closes#18
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5586 4a71c877-e1ca-e34f-864e-861f7616d084
* Don't strdup strings retrieved from __FILE__ directives; just store const char* the pointer instead (prevents a memory leak as well)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5549 4a71c877-e1ca-e34f-864e-861f7616d084
* Declare these functions with static linkage instead.
* Update autotools buildsystem, Code::Blocks project and MSVC project, raw win32 Makefiles don't need an update
NOTE: This may require updating of buildsystems not mentioned above
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@5544 4a71c877-e1ca-e34f-864e-861f7616d084
* Turn some signed integers ("int" or "SDWORD") into unsigned integers where it made more sence (and signed integers caused warnings about comparing signed with unsigned integers)
* At some places cast to an unsigned integer (where the type of the expression is a signed integer but the result of it logically _always_ is an absolute value)
* Remove unused struct-type EVENT_INIT
* Convert use of malloc->memcpy->free to one single realloc call (and now make sure to actually check realloc's return value, which wasn't done with the malloc call)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@3037 4a71c877-e1ca-e34f-864e-861f7616d084
* Refactor NETlogEntry() to use the array of strings (packetname[]) instead of a large switch-statement with a lot of copy & pasted code
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@2617 4a71c877-e1ca-e34f-864e-861f7616d084
* Use decent assertion expressions for some asserts (e.g. !"string", so that debuggers actually display something more interesting than 'FALSE')
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1815 4a71c877-e1ca-e34f-864e-861f7616d084
This patch will create crashes where the code expects FREE to set the pointer to NULL! (Those problems should be fixed anyway.)
The only occassion where I saw this is when quiting a game, so it is not too much of an issue.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1459 4a71c877-e1ca-e34f-864e-861f7616d084
- Simplify definition of ASSERT(). Now is the same for GCC and MSVC.
This may not work on versions older than MSVC 2005. If you experience this, please report it!
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@613 4a71c877-e1ca-e34f-864e-861f7616d084