Compressed resource optimization #93
Conversation
|
This is related to #89 |
pkg/archive/archive.go
Outdated
| @@ -55,9 +55,10 @@ type Entry interface { | |||
| Path() string // Absolute path to the entry in the archive. | |||
| Length() uint64 // Uncompressed data length. | |||
| CompressedLength() uint64 // Compressed data length. | |||
| CompressedAs(compressionMethod uint16) bool // Whether the entry is compressed using the given method. | |||
There was a problem hiding this comment.
Isn't the 8 value for deflate only relevant in the context of a ZIP? Theoretically we could use Archive for a CBR/RAR file as well.
To remove any ambiguity, what do you think of adding an explicit enum with Deflate case in the toolkit instead of using ints?
Having CompressionMethod() CompressionMethod might be more convenient too.
There was a problem hiding this comment.
Technically, the latest version of the ZIP spec supports other compression formats as well. I think the correct term would be that the EPUB standard only supports modes 0 and 8. I think it's worth keeping the door open for other compression modes for other container formats that we might support, or whose specs may change to support other methods in the future.
I've added your suggested CompressionMethod enum
This enhancement allow for resources to be read in their compressed form, which usually applies to most if not all resources in any ZIP-based format (primarily EPUB). This saves CPU time because any modern web browser is capable of decompressing a
deflate-encoded resource, so the raw contents of the ZIP can be directly proxied to any web client. Closes #89