bspatch: use sizeof(header) rather than hardcoded size

[emaste]

Summary

Although the header size is fixed, this change provides an explicit indication of what the size is being used for.

[allanjude]

+1 for killing magic numbers

[emaste]

• update with _Static_assert and use %jd / intmax_t as suggested by @kib
• add an offset variable that we increment by the header / bzctrllen / bzdatalen to avoid repeating the same expression

[kib]

I looked at this stuff again, and having the construct char header[32] and static_assert(sizeof(header) == 32) feels weird. Might be, instead use a #define HEADER_SIZE 32 and then char header[HEADER_SIZE]; and use HEADER_SIZE.

Sorry for bike-shedding, but this is weird when seen completed.

[emaste]

Incorporate comments from @kib

[emaste]

Commit summary will be use #define for header size instead of magic number Also note that the update includes a change to keep an increasing offset rather than repeating the same expression in several places.

[allanjude]