Hi,

7051 noticed today that in Delight's server log there are many occurences of

Code: Select all

^7FS_Open("data/maps/.mapcfg", "rb", false): nasty filename rejected
^7FS_Open("maps/.mapcfg", "rb", false): nasty filename rejected


I think this happens cos the condition in fs.c in the engine code, line 1559, applies:

Code: Select all

// all: forbid leading dot on any filename for any reason
       if (strstr(path, "/."))
               return 2; // attempt to go outside the game directory

It looks like these "/.mapcfg" names are coming from GotoNextMap im qcsrc/server/g_world.qc, line 1580ff:

Code: Select all

// cvar "nextmap" always gets priority
if(TryFile(strcat("maps/", cvar_string("nextmap"), ".mapcfg")))
{
   localcmd(strcat("exec \"maps/", cvar_string("nextmap"), ".mapcfg\"\n"));
   return;
}

(Source is from 2007-01-23)

Maybe "nextmap" should be tested for content before trying to exec the mapconfig?

The "filename rejected" message doesn't occur when I change map with chmap or when I set "nextmap" to anything before time limit or frag limit is reached.

Or are the "nast filename rejected" messages hints on some other problem?

why the fuck is it nasty o_O

esteel ur a nerd u figure it out

Nil wrote:Or are the "nast filename rejected" messages hints on some other problem?

No, the messages are obviously caused by engine changes between 2.2.2 and 2.2.3...
Anyway, i added a check if nextmap is empty, which seemed redundant before and the error mesages are gone...

cyan wrote:why the fuck is it nasty o_O

The error message is there to prevent the use of ".." within filenames, which would allow to read/write files outside of the gamedir. It's a bit too strict, since it also forbids names like ".mapcfg", but it's fast and reliable.

KadaverJack wrote:No, the messages are obviously caused by engine changes between 2.2.2 and 2.2.3...
Anyway, i added a check if nextmap is empty, which seemed redundant before and the error mesages are gone...

Thanks!

ok, and for me: what does this mean? is it fixed in >=2.2.4 ?

ZyanKLee wrote:ok, and for me: what does this mean? is it fixed in >=2.2.4 ?

yes

thx

I just wanted to add that one may safely ignore the messages in 2.2.3. The added check just prevents the warning from appearing - it never was a critical problem, that is, it never caused wrong or missing map changes.

yeah thats right ... but it causes annoying messages in the IRC-chan