Hello, On 3/1/13 3:40 PM, Jasmin Schnatterbeck wrote: the cache is for PV names, not for PV values. What is printed as value of fscript anyhow? It might get 0 which will result in same logic with IF. Cheers, Daniel

Hi Daniel, that issue seems to be fixed in 3.3... may the corresponding commit is related to app_lua or pvapi. Cheers Jasmin Am 02.03.2013 04:06, schrieb Jasmin Schnatterbeck: > Hi Daniel, > > when printing fscript only, there is no output line in the log. So I > added the following code: > > if(fscript == nil) then > sr.log("warn","Filterscript is nil") > elseif(fscript == 0) then > sr.log("warn","Filterscript is 0 n") > elseif(fscript == "0") then > sr.log("warn","Filterscript is 0 s") > else > sr.log("info","fscript found") > end > > And the output is: > > WARNING: app_lua [app_lua_sr.c:99]: Filterscript is nil > > So nothing is returned at all. But the htable entry is definitely set, > because it works in some cases (without changing anything) and in some > cases not. According to sht_dump, everything seems to be ok. > > When using pv index like sr.pv.get("$(hdr(X-Test)["..i.."])") , a > similar problem occurs (sometimes the value is returned, sometimes not). > When omitting the pv index at all, it works everytime... > So may the dynamic name of the variable is the problem? But (according > to the debug message) the pv cache check function gets the correct name... > > Cheers > Jasmin > > Am 01.03.2013 17:49, schrieb Daniel-Constantin Mierla: >> >> Hello, >> >> On 3/1/13 3:40 PM, Jasmin Schnatterbeck wrote: >>> >>> Hi, >>> >>> when using sr.pv.get() in app_lua the pv value is not always returned. >>> >>> Lua code is: >>> >>> local fscript = sr.pv.get("$(sht(myscripts=>" .. args["filterid"] >>> .."))") >>> >>> if(fscript) then >>> sr.log("info","Filterscript is " .. fscript) >>> else >>> sr.log("warn","fscript not found!") >>> end >>> >>> >>> And log shows: >>> >>> app_lua [app_lua_sr.c:888]: pv get: (sht(myscripts=>1)) >>> <core> [pvapi.c:318]: pvar [$(sht(myscripts=>1))] found in cache >>> WARNING: app_lua [app_lua_sr.c:99]: fscript not found! >>> >>> So the pv is being discovered in the pv cache, but sr.pv.get() returns >>> no value. >>> >>> I think the problem occurs especially in the case, when using >>> sht_reload mi command. So I have had a short look into htable.c and I >>> did not find any code, that clears the pv cache... is that >>> necessary/may the reason for the problem, that pv is found in pv cache >>> but nothing returned? (kamailio 3.2.3) >>> >>> Thanks for any ideas... >> >> the cache is for PV names, not for PV values. What is printed as value >> of fscript anyhow? It might get 0 which will result in same logic with >> IF. >> >> Cheers, >> Daniel