|
|
|
|
|
by schiffern
3484 days ago
|
|
Possibly dumb question: why doesn't the hash verification take place on the uncompressed contents, moving recompression from the updating-critical-path to the nice-to-have-later-to-save-disk-space category? My dream updater would just be a bunch of stream processors, so the downloading / decompression / disassembly (a la courgette), patching / recompression happens all at once, as the data packets come in (thus utilizing both CPU and IO). If done right an update shouldn't take much longer to complete than the download step, no? |
|