Hello team,
I just now merged a cleanup of the content-encoding code [1]. While working on
this, I noticed that we (accidentally) used enormous amounts of memory
allocated for buffers used for the gzip, brotly and zstd decompression. It no
longer does this. I shrunk the allocations from 10MB to 16KB. I also made them
not done repeatedly during a transfer but they are reused during the
transfer's life time.
Anyway, since this is quite a drastic change I figured I should just make
everyone aware of this and if someone is using this feature it would be cool
if you could perhaps just do a quick performance tests or two, to make sure
that this updated version does not becomes a total hog for you.
Thanks!
[1] = https://github.com/curl/curl/pull/16079
--
/ daniel.haxx.se || https://rock-solid.curl.dev
--
Unsubscribe: https://lists.haxx.se/mailman/listinfo/curl-library
Etiquette: https://curl.se/mail/etiquette.html
