Monday 23 June 2014

    Finally this might be the last post about my project and after this my posts will be about working on incremental goals. I'm glad that till now I'm on my schedule and accomplished the things that I proposed to be done by me by this time. The project implementation is over and now it is time to get reviewed some gtk+ patches and style issues and land all the patches on master branch.

    I'll give one last overview of whatever I've done so far. The ultimate goal of my project was to expand the range of cities available in libgweather location entry suggestions. So that for any user input if the place exists and even if it is not in the Locations.xml database then also it is shown in the suggestions. That is done by forward geocoding. 
For example,  for "abc" input, there is no location in Locations.xml file but still it shows suggestions for places named "abc".

Invalid input

    The next step was to create a detached location using geocode place and make it addable. The locations that are shown in the suggestions should be selectable and should be addable to the application. For this, I added detached location support to location-entry.c and made it addable.
Here, in the next screenshot "Abc,Canada" is shown added to the application.

"Abc, Canada" location is added
    After this point, it seemed working fine and I made the first patch of my work. I faced few SegFaults in between my work and with the help of my mentor I solved them. After my mentor reviewed and tested my patch he suggested me few changes that I could not think that they also may cause SegFault but after that I merged those changes and landed up with new final patch. 

    During final testing I discovered one more thing that I was not able to delete one detached location that I added this way. After digging up and discussing with my mentor I came to know the loophole and I found that serialization and deserialization of detached locations was not consistent in weather-location.c file. So I fixed that bug and made a patch. So there are 2 patches that I made for libgweather which are ready to land and the only one hurdle left to be crossed is the issue with gtk+. Hope to see my changes in master branch as soon as possible.

    Finally thanks to my mentor to keep supporting me throughout the journey. Now it's time for midterm evaluation. Hope to keep doing well during the next half of the journey !

Monday 2 June 2014

    Hi, I'm again back with few updates. Finally, I've made some good progress and "The mission" is about to get over. I'm done with the implementation of getting the list of geocode places in the suggestions list. In this I'm facing a problem of showing the "Loading...." string, while the asynchronous task is processing, as the string seems to be hiding somewhere behind the input-text box. I've got nothing after digging up too much so now I'm thinking for another option or may be Giovanni(My mentor) can suggest something. Otherwise it is working fine.
   The second part that I was supposed to implement was to make those geocode places addable and for that I was supposed to find the nearest city from the  database using coordinates of the fetched geocode places. Then using that location I'm supposed to make a detached location and add it as selected location. For now I'm adding the nearest city to test it because creating detached location is not a big deal. In this the problem that I'm facing is if I click on the add button then it works perfectly but if I press enter key to add the location then it gives segmentation fault and I cant understand this strange behaviour till now.
     So this is it. And after this, I'll be working on few other things related to libgweather and Gnome-weather.