nwnx_resman no longer works

[stubhead]

it appears that resman no longer monitors its directories. it used to work so that placing a new file in one of the appropriate subdirectories under the resman directory [e.g., .ncs files in the 'ncs' subdirectory] would cause resman to take note of it and import the resource. this no longer happens. the plugin sets the correct directory source path and RegisterDirectoryHandlers() appears to be called correctly, but both HandleResourceExistsEvent() and HandleDemandResourceEvent() are NEVER called... this was working perfectly in the 2008 version, but now does absolutely nothing.

[virusman]

I'm using it to override or add scripts on my PW all the time, so it definitely works.. I'll check.

[stubhead]

any results ? on my end, i still can't see how this can work. i've added detailed logging to all functions, but nothing is ever logged anywhere beyond the initial config phase. the very first thing any of the handlers tries to do is log the call, and... nothing.

i'll pm you over at nwnx.org to provide you with more detailed data.

[virusman]

Do you have a ResMan entry in your nwnx2.ini?

[virusman]

Looking at the code, it seems to be the only case when it may unexpectedly fail to initialize the directory handler.

[virusman]

Also, try disabling nwnx_jvm.

[stubhead]

On 3/21/2018 10:49 PM, Victor Babenko wrote: Looking at the code, it seems to be the only case when it may unexpectedly fail to initialize the directory handler.

i've added tracing to all functions, and according to the output there are no errors. possibly errors are simply not being reported... i do see that nwnx2 says it found the plugin at 0xB7503120 and the handlers are found at 0xb74fxxxx -- actually before the area of memory occupied by the plugin. since i am completely unfamiliar with the design of the plugin, this may be a stupid question, but is this difference in addresses cause for concern ? ================================= in nwnx2.txt - RESMAN: Supports the new plugin interface RESMAN plugin registered. Address: B7503120 in nwnx_core.txt - [2018-03-08 20:41:03] HookEvent(ResMan/Exists, 0xb74fd384) [2018-03-08 20:41:03] HookEventOptionally(ResMan/Exists, 0xb74fd384) [2018-03-08 20:41:03] HookEventOptionally() returning 0x70001 [2018-03-08 20:41:03] HookEvent(ResMan/Demand, 0xb74fd4a1) [2018-03-08 20:41:03] HookEventOptionally(ResMan/Demand, 0xb74fd4a1) [2018-03-08 20:41:03] HookEventOptionally() returning 0x60001 in nwnx_resman.txt - o Logging missing resources: 1 o Debugging CRes construction and destruction: 1 o Log level: 6 o RegisterDirectoryHandlers: o - handleResourceExists (ResMan/Exists, 0xb755b3c4) = 0x70001 o - handleDemandResource (ResMan/Demand, 0xb755b506) = 0x60001 o Directory source path set: /home/nwn/nwn/resman

[virusman]

Do you have nwnx_jvm enabled?

[stubhead]

i completely removed nwnx_jvm.so from the directory and verified that it isn't starting [no log file].
problem persists.

can you say which version of nwnx_resman you're using, virusman ?

also, i'm wondering if this might be a problem with environment. i'm running ubuntu 16.1 on virtualbox. should not be a problem, correct ?

[virusman]

I've just created a new VM with Ubuntu 16.04 32-bit, compiled nwnx2-linux from master – resman seems to work. It did crash at first in a logging call (fixed that) but other than that it shows all resource requests properly:

File: ttr01_h01_02.wok, Exists?: 0, mtime: 0
File: ttr01_h01_02.wok, Exists?: 0, mtime: 0
File: plc_w02.pwk, Exists?: 0, mtime: 0
File: plc_w02.pwk, Exists?: 0, mtime: 0
File: plc_f03.pwk, Exists?: 0, mtime: 0
File: plc_f03.pwk, Exists?: 0, mtime: 0

Are you sure you're using nwnx2.so that was built from the same codebase as nwnx_resman.so?
Please try building a clean install of nwserver and nwnx2-linux.