I am hoping OpenWRT developers here can help me out.

In a normal situation, if I find some bugs on a package, it is as simple as using quilt to create patches, recompiling the package, and installing on the target to test. In a rare situation, however, one may encounter a package that dynamically downloads additional source packages to add to the package during the compilation process. Of course, creating patches for a package like this is more involving, nevertheless is doable. However, applying the patches has become trickier. For instance, take a look at the OpenWRT Makefile for the net/freeswitch package, where there are some packages, i.e. mod-celt, json-c (for mod-curl), mod-flite, etc., whose source packages are dynamically downloaded when the net/freeswitch package is compiled. This is done through line# 274-276 on the OpenWRT Makefile for the net/freeswitch package during the Build/Prepare stage as shown on line# 358 of the same OpenWRT Makefile. However, this only downloads the flite source package and the process to extract it won't be done way after the process of applying the patches. As you can see, unless there is a way to instruct the process to extract the downloaded file right after it is downloaded and/or before the process of applying patches is called, this has become a nightmare because the compilation will crash during the process of applying patches (flite package source hasn't been extracted).