minimal gpt partition size is wrong

[therjak]

in partition/gpt/partition.go are multiple checks for end > start.
(e.g.

go-diskfs/partition/gpt/partiton.go

Line 253 in 8ff8bc6

case start >= 0 && end > start && size == calculatedSize:

)

Both start and end represent sectors with start the first and end the last sector of the partition. so the minimal allowed sector size of a partition becomes 2 while it should be 1.
this is somewhat correctly reflected in

go-diskfs/partition/gpt/partiton.go

Line 257 in 8ff8bc6

case size > 0 && start > 0 && end == 0:

if size equals blocksize (but it would create an invalid result for 0 < size < blocksize, with end = start-1).
The code will fail later on in a second check for end > start.

[deitch]

Are you suggesting that this:

 case start >= 0 && end > start && size == calculatedSize: 

should be

 case start >= 0 && end >= start && size == calculatedSize: 

To be honest, this line is a bit in error in that it doesn't do anything, as it is a lone case statement.

[therjak]

yes.

go-diskfs/partition/gpt/partiton.go

Line 150 in 8ff8bc6

case p.Size <= 0 && p.End > p.Start:

go-diskfs/partition/gpt/partiton.go

Line 253 in 8ff8bc6

case start >= 0 && end > start && size == calculatedSize:

go-diskfs/partition/gpt/partiton.go

Line 254 in 8ff8bc6

case size == 0 && start >= 0 && end > start:

should allow end == start

I assume

go-diskfs/partition/gpt/partiton.go

Line 152 in 8ff8bc6

case p.Size > 0 && p.End <= p.Start:

should be
case p.Size != 0 && p.End < p.Start:
or
case p.Size != 0 && p.End == 0

(btw p.Size <= 0 is the same same p.Size == 0 as it is unsigned but I guess that is more about preference)

[deitch]

Happy to take a PR on it.