On 2025/1/17 17:40, Neal Gompa wrote:
On Fri, Jan 17, 2025 at 2:17 AM Phillip Lougher
<phillip.lougher@xxxxxxxxx> wrote:
Neal Gompa wrote:
Thank you, I appreciate whatever you can do to make things better for
our use-cases. :)
Well it is obvious where your bias lies.
Well, I appreciate your work too! Nowhere have I said you aren't doing
good work with SquashFS. SquashFS has worked fairly well for Fedora's
live media for a long time. It's still the default for live media
creation in kiwi too. I have a lot of sympathy for you as SquashFS has
been the technology used for this sort of thing for 15 years and in
that time frame, you haven't gotten much help from anyone.
I don't think that is fair, take a small enhancement
UUID and LABEL request
https://github.com/plougher/squashfs-tools/issues/59
as an example, it was formally raised in 2019 (with 13
likes) and even some PR raised in 2023:
https://github.com/plougher/squashfs-tools/pull/264
until now it has no upstream response.
Another small enhancement: folks would like to add
POSIX ACL support (another small change) to SquashFS:
v1: https://lore.kernel.org/r/af77c1f80e2725c4cf1bf106d6add820b3b0eed5.1523276963.git.geliangtang@xxxxxxxxx
v2: https://lore.kernel.org/r/975b0f7acbb65445551ee374a2dd38d553ac2e6a.1523326310.git.geliangtang@xxxxxxxxx
v3: https://lore.kernel.org/r/cover.1533630854.git.geliangtang@xxxxxxxxx
v4: https://lore.kernel.org/r/cover.1548403955.git.geliangtang@xxxxxxxxx
v5: https://lore.kernel.org/r/cover.1548406694.git.geliangtang@xxxxxxxxx
None of them got response.
If digging into more in the kernel mailing list, Android
team also tried to improve SquashFS first rather than
use EROFS at first. So obviously SquashFS gets more
community development resource (even enterprise resources)
rather than EROFS.
Honestly, our team also considered improving SquashFS
format in 2017 to address its performance issue especially
for smaller chunks (-b <= 16384) for our high-performance
devices. but those are not minor changes, and the previous
feedback status scared us.
I didn't want to say too much in this reply (as I said,
currently I don't even care Fedora LiveCD case since some
feature people may care is missing like metadata
compression and multi-threaded CDC), but "By doing this
they deliberately reduce the compression and speed of
Squashfs in their tests. This IMHO is biased and
unethical. " really sounds unpleasant to me:
https://erofs.docs.kernel.org/en/latest/comparsion/dedupe.html
just would like to express the benefits of the compressed
data deduplication feature (like two version wikipedia
texts). Since EROFS doesn't support metadata compression,
so I turned it off and explicitly drop a comment below.
As even don't consider our previous paper, for 3rd-party
comparsion of these two FSes, there are many other
online materials, e.g.:
https://youtu.be/kLxM4FyiVpQ?t=2024
and
https://sigma-star.at/blog/2022/07/squashfs-erofs
Thanks,
Gao Xiang
--
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
