nomennescio

To support #2577

> Cool, let us know how the experiments go -- i'm particularly concerned about adding startup delay since we are a lot slower than python/ruby/etc in that use-case due to...

> This isn't quite a direct comparison, but I want to be able to be used for scripting use-cases, and things like CGI scripts, etc. > > ``` > $...

I still need to rewrite existing code to support decompressing, even for non-compressed images. I can currently only compare on a Virtual Machine, so I'm not sure how representative any...

What I mean is that I have to rewrite the whole load_image to not use `FILE*` at all. But move all file operations upfront. Furthermore I need to check how...

> The embedded image doesn't help much for startup time unless it's a deployed image stripping out stuff that's unused, it's basically just appended to the binary. Depends on how...

As for compression, if it's optional, it would at least give the benefit to people to whom disk space occupied by images does matter (at ~0.3GB each it does matter...

clang generates warning for `fallthrough` attribute, ``` clang++ -c -Wall -Wextra -pedantic -DFACTOR_VERSION="0.99" -DFACTOR_GIT_LABEL="remotes/pull/2589/merge-5e0a441b8f78dc1fa2cd73e7d7ecb3bfaa32c26b" -m64 -O3 -std=c++11 -o vm/zstd.o vm/zstd.cpp 117 In file included from vm/zstd.cpp:4: 118 vm/zstd.c:2092:17: warning: use...

@mrjbq7 Here are preliminary results, using zstd internal allocation of a temporary buffer. ``` /media/sf_project/software/factor.fork$ time for i in {1..100};do ./factor -i=factor.image -e="";done real 0m24.924s user 0m8.939s sys 0m2.520s /media/sf_project/software/factor.fork$...

Basis + core compression factor : 6.6 ``` -rwxr-xr-x+ 1 192485800 Feb 15 18:58 factor.basis+core.image -rwxr-xr-x+ 1 29304829 Feb 15 18:59 factor.basis+core.image.compressed ``` ``` /media/sf_project/software/factor.fork$ time for i in {1..100};do...