Hi, I am debugging the backend using valgrind to investigate any possible memory leaks. This was indicated to me as a debugging necessity by jerome.
Edward On 12/10/2015, aitor_czr <[email protected]> wrote: > No problem :) > > On 12/10/15 22:24, Edward Bartolo wrote: >> Hi, >> >> This is embarrassing to me. I discovered a bug in the main auto >> connect loop. This loop uses two lists: >> a) a sorted list of wifis in power descending order >> b) a list of installed connections (essids) >> >> The algorithm was originally intended to take the first item from the >> power sorted list (a), searches in list (b) for its presence. If it >> finds the item in list b, it attempts to connect. If not it continues >> iterating until it uses all the items in the power sorted list. >> >> My error was to swap these two lists! >> >> I discovered this bug because netman started to connect to a >> neighbour's wifi notwithstanding that my wifi signal was far stronger. >> This set me thinking there was something fishy in my algorithm. >> >> I uploaded the code changes to correct this horrible bug. >> >> Edward >> >> >> On 12/10/2015, aitor_czr<[email protected]> wrote: >>> >Ok, thanks. >>> > >>> >On 12/10/15 17:45, Edward Bartolo wrote: >>>> >>Hi Aitor, >>>> >> >>>> >>fsmithred confirmed that adding: >>>> >> >>>> >>iface wlan1 inet dhcp >>>> >> >>>> >>to /etc/network/interfaces was what was causing disconnecting to >>>> >> fail. >>>> >>This means, all system network devices to be used have to be defined >>>> >>in /etc/network/interfaces as showm or anything that allows ifup and >>>> >>ifdown to function. >>>> >> >>>> >>Before actually building the packages for Devuan, try to ONLY add the >>>> >>backend executable to: >>>> >> >>>> >>/usr/lib/netman/bin >>>> >> >>>> >>All the other compiler generated files are unnecessary as we are only >>>> >>invoking the backend. The same applies for any files generated by the >>>> >>compiler while building the frontend by Lazarus. So, to conclude, the >>>> >>only required executables are >>>> >> >>>> >>/usr/bin/netman >>>> >>/usr/lib/netman/bin/backend >>>> >> >>>> >>We also have an /etc/netman.conf text file. >>>> >> >>>> >>Edward >>>> >> >>>> >>On 12/10/2015, Edward Bartolo<[email protected]> wrote: >>>>>> >>> >Hi Aitor, >>>>>> >>> > >>>>>> >>> >It seems fsmithred bug is not a bug after all. I am waiting for >>>>>> >>> > his >>>>>> >>> >reply as to whether configuring /etc/network/interfaces to >>>>>> >>> > contain the >>>>>> >>> >configuration line: >>>>>> >>> > >>>>>> >>> >iface wlan1 inet dhcp >>>>>> >>> > >>>>>> >>> >will solve his problem. I am convinced it should work, but let us >>>>>> >>> > be >>>>>> >>> >prudent and wait. >>>>>> >>> > >>>>>> >>> >Edward >>>>>> >>> > >>>>>> >>> >On 12/10/2015, Edward Bartolo<[email protected]> wrote: >>>>>>>> >>>> >>Hi Aitor, >>>>>>>> >>>> >> >>>>>>>> >>>> >>fsmithred found a serious bug that prevents wlan1 from >>>>>>>> >>>> >> disconnecting. >>>>>>>> >>>> >>I will debug that, and then, upload again the necessary >>>>>>>> >>>> >> changes. >>>>>>>> >>>> >> >>>>>>>> >>>> >>Edward. >>>>>>>> >>>> >> >>>>>>>> >>>> >>On 12/10/2015, aitor_czr<[email protected]> wrote: >>>>>>>>>> >>>>> >>>Good! >>>>>>>>>> >>>>> >>> >>>>>>>>>> >>>>> >>>I will update the packages of netman. I'm still >>>>>>>>>> >>>>> >>> generating the >>>>>>>>>> >>>>> >>>Packages*.gz and the Contents*.gz of my repository. >>>>>>>>>> >>>>> >>> Adding >>>>>>>>>> >>>>> >>>Contents*.gz >>>>>>>>>> >>>>> >>>slows the proccess. It takes so long... >>>>>>>>>> >>>>> >>> >>>>>>>>>> >>>>> >>>...Patience... >>>>>>>>>> >>>>> >>> >>>>>>>>>> >>>>> >>>Aitor. >>>>>>>>>> >>>>> >>> >>>>>>>>>> >>>>> >>>On 12/10/15 14:00, Edward Bartolo<[email protected]> >>>>>>>>>> >>>>> >>> wrote: >>>>>>>>>>>> >>>>>> >>>>Hi All, >>>>>>>>>>>> >>>>>> >>>> >>>>>>>>>>>> >>>>>> >>>>The changes have been uploaded. The backend should >>>>>>>>>>>> >>>>>> >>>> now support >>>>>>>>>>>> >>>>>> >>>>other >>>>>>>>>>>> >>>>>> >>>>devices besides wlan0 and eth0. >>>>>>>>>>>> >>>>>> >>>> >>>>>>>>>>>> >>>>>> >>>>Edward > > _______________________________________________ Dng mailing list [email protected] https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
