Am 05.05.2024 um 20:03 schrieb Konstantin Kushnir:
On Sun, 5 May 2024 11:22:52 +0200
Harald Oehlmann <[email protected]> wrote:
Am 04.05.2024 um 23:52 schrieb Konstantin Kushnir:
I am pleased to announce version 1.5.0 of cookfs.
Just to double check. Are you aware that zip virtual file system is now
included in TCL 9.0?
Thank you for the headup! I saw this information briefly, but have
already forgotten.
Unfortunatelly, I need better compression and lzma is the best choice
in 2024. Cookfs also offers the option to further improve compression
by using the solid achive mode. And it is not just a solid mode, but
a compromise between solid archive benefits and random data access
within archive.
Another killer feature of Cookfs is the parameters (compression, page
size, etc.) for different parts of file. The ability to use your own
commands for compression/decompression also adds flexibility.
I haven't dug into asynchronous mode enough yet, where further pages
are decompressed in the background and the application doesn't have to
wait for new data to be decompressed when it needs it. It seems that
this mode is now not working as the developers had planned. It will be
a nice performance boost if it works.
Thus, zipfs in core is a nice feature for combining multiple files into
a single file with a small compression bonus. But it can't be
considered a true compressed archive when other formats like
lzma/zstd/bzip2 available, and the zip format is more about portability
but not flexibility.
Is there anything in cookfs, what is not included in TCL 9 core ?
Then, we may port it to TCL ...
I am not sure about this. A feature of combining multiple files into
one is not something that many scripts require. Zipfs looks like a
sufficient tool to fulfill some basic possible requests, such as
creating a data package.
Cookfs contains advanced features for creating and working with data packages. This is more interesting for specialized applications such as tclkit systems. Thus, it is better to be as a module, rather than
taking up space in all the built interpreters.
Konstantin,
don't get me wrong. cookfs is great. In my 25 years of experience with
TCL, I have seen great projects coming and going. The "going" part is
the problem.
The best way to preserve something great is to get it into the core.
There are many compromises required.
But I invite you as a wizard to look what may be moved to the core, even
if it is a small piece.
Wizards normally tend to do their thing and that is ok. But if the core
may get a small piece, welcome !
Take care,
Harald
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)