Code review comment for lp://qastaging/~mabac/linaro-image-tools/image-support-for-android

On Fri, Jun 10, 2011 at 3:25 PM, Guilherme Salgado
<email address hidden> wrote:
> On Fri, 2011-06-10 at 10:56 +0000, Mattias Backman wrote:
>> 193   + self.sector_size = 512
>> 194   + self.android_image_size = 2 * 1024**3 # rounded up from ~1.13GiB
>
> I'm confused by the 1.13GiB.  I thought 1024**3 wass 1GiB?

Oh, yes. The sfdisk command (stolen from the production code) creates
partitions that would consume at least 1.13 GiB. So I figured I'd
round that up to 2 GiB since that would be more like an actual size sd
card than fractions of GiBs. I'm still not sure if I should clarify
the comment or change the partition sizes so we don't create a 2 GiB
tempfile.

>
>> 195   + self.android_offsets_and_sizes = [
>> 196   + (63 * self.sector_size, 270272 * self.sector_size),
>> 197   + (270336 * self.sector_size, 524288 * self.sector_size),
>> 198   + (794624 * self.sector_size, 524288 * self.sector_size),
>> 199   + (1318912 * self.sector_size, (self.android_image_size -
>> 200   + 1318912 * self.sector_size)),
>> 201   + ((1318912 + 32) * self.sector_size, ((1048576 - 32) *
>> 202   + self.sector_size)),
>> 203   + ((2367488 + 32) * self.sector_size,
>> 204   + self.android_image_size - (2367488 + 32) * self.sector_size)
>> 205   + ]
>>
>> The reason for adding and substracting 32 here is that it seems that
>> the extended partition uses that much space. That space gets stolen
>> from the first partition even though sizes and offsets are explicitly
>> requested by the sfdisk commands.
>
> You must've had some fun to figure that out, eh? :P

Well there are people who consider pain to be fun, but I'm not one of them. ;)

> It'd be nice to have this as a comment, I think.

True, I'll add a comment about this as well as the map/zip decision.