Looks like goahead is not a good choice to replace LuCI.
It uses too much space(850KB), maybe memory cost is low, but flash cost is not acceptable.
Also it request librt.so and libpthread.so, that is about 100KB addition flash cost.
I think maybe write a real light weight one is the only choice. We do not need that complex cgi function, we just need a way to transfer data between browser and the server.
Currently vocore.io/store, the trace page is based on a light weight C language web application(I am not good at php, so have to write one in C), it will not be very hard if we just have a simple cgi and do not use https.
>>It uses too much space(850KB)
It is a proper space compared 8M FLASH. My experience: 90% simple implement will be replace later because different requirements are required later. Finally, the beginning solution will be a good solution.
I think vonger.io could run goAhead without any problem. If it could run, it is proper to make goAhead as a standard solution for vonger.io until clear requirement defined.
It might be a noobish question of mine but
How about using the SD card to store the os or a partition on the SD and have the flash act like a boot loader.
Or is it not possible to make it work like that?
It is possible, but not easy 🙂
What about mini-httpd? According to http://wiki.openwrt.org/doc/howto/http.overview it looks like 21k for the binary w/o SSL and the libc and libgcc might already be included. Looks like another 70k-630k if you want to add SSL.
I went looking to compare lighttpd with your numbers above and it seems like if you want real web server performance, the extra ~110k (if you can compile it w/o SSL) is probably worth it.
Looking forward to my package arriving so I can play too!
🙂
Oops, that would be 200k for lighttpd versus 21k for mini-httpd (each w/o SSL).
Depending on the use-case, might be worth the additional storage.