Can you include .3MF to the list of re-compressible formats? Its structure is similar to MS Office 2007 documents and Open Document Format. It is a ZIP Deflate archive with XML data and some JPG, and/or PNG pictures inside. Otherwise, if I try to compress .3MF it bearly makes it smaller unless I recompress .3MF to the Store setting then it makes it a lot smaller.
Wish they all would move to 7zip ZSTD in the first place so that the optimized file size with FileOptimizer would be 50% of the ZIP Deflate version. And there would be no extra compression needed :)
I noticed that the option to add the optimize archive function to the context menu is missing on Windows 10.
Opening each archive with the interface in order to click it becomes tedious with many files.
Same for others functions like Remove Archive Encryption
Solved Different Hash Algorithns
-
I know CRC32 was the standard for checking the integrity of compressed files but it is not really used anymore. Can you add ones that are used now MD5 or SHA?
-
@brian-gregory said in Different Hash Algorithns:
@bigdog I think It’s just showing you the value from the directory in the archive. If you stored a MD5 or SHA1 there instead of a CRC32 you’d have a new archive format that nothing else would understand.
I guess they could have used them on their own new PA format. But it’s too late for that now. PA2 or PAX anyone?Indeed, what is used for data check is defined by the format. So using anything outside format specs would make it unreadable by everyone else.
But main reason CRC is used is because it is faster and has same function for this purpose as MD5 or SHA, so using something else would just be slower and would create larger archives - it would make no sense.
However this is only true for data check in archives. When using encryption, we use SHA2… or for instance in PA format for data deduplication, a variant of MD5 is used.
So everything has its purpose.
In 2017 we sped up zip compression by adding faster cpu/hardware optimized crc algorithm that uses cpu instructions, together with some other optimization this made zip compression of large files significantly faster than before (even up to 2x).
thanks for the feedback!
-
@bigdog said in Different Hash Algorithns:
I know CRC32 was the standard for checking the integrity of compressed files but it is not really used anymore. Can you add ones that are used now MD5 or SHA?
Hello bigdog, where exactly do you want them added?
-
In the window that shows the files in an archive where you have a column for CRC, that could be re-purposed for those newer hash codes.
-
@bigdog I think It’s just showing you the value from the directory in the archive. If you stored a MD5 or SHA1 there instead of a CRC32 you’d have a new archive format that nothing else would understand.
I guess they could have used them on their own new PA format. But it’s too late for that now. PA2 or PAX anyone? -
@brian-gregory said in Different Hash Algorithns:
@bigdog I think It’s just showing you the value from the directory in the archive. If you stored a MD5 or SHA1 there instead of a CRC32 you’d have a new archive format that nothing else would understand.
I guess they could have used them on their own new PA format. But it’s too late for that now. PA2 or PAX anyone?Indeed, what is used for data check is defined by the format. So using anything outside format specs would make it unreadable by everyone else.
But main reason CRC is used is because it is faster and has same function for this purpose as MD5 or SHA, so using something else would just be slower and would create larger archives - it would make no sense.
However this is only true for data check in archives. When using encryption, we use SHA2… or for instance in PA format for data deduplication, a variant of MD5 is used.
So everything has its purpose.
In 2017 we sped up zip compression by adding faster cpu/hardware optimized crc algorithm that uses cpu instructions, together with some other optimization this made zip compression of large files significantly faster than before (even up to 2x).
thanks for the feedback!
-
I assumed (maybe incorrectly?) that those hash values were being calculated in real time when opening (or verifying) the archive. If that was the case it should not matter if that is not the standard from the days of old as the program is just displaying additional data.
Thanks