Improve binding redirects performance#3239
Conversation
|
IIRC we used Cecil because of mono bugs. Maybe they are fixed now and we can try again |
|
Is this the same API as in Why do we need no new dependency here? Is this reflection loading the assembly or just reading it? /cc @viktor-svub |
|
seems similar to #3223 but it's very different in both intent and implementation:
I'd recommend to take #3239 (this), if/when/after known/proven safe on supported platforms, |
|
ok let's try it! |
|
@viktor-svub wow that's a nice summary! @mrinaldi I like the performance figures! |
|
This branch was lying around for quite some time waiting for tests to be promoted to a PR. Anyway, nice work @viktor-svub |
|
I'm not sure this was the correct approach: It seems this In any case we should indeed go the complete way and remove mono.cecil instead of having multiple dependencies. @viktor-svub Is it possible for you to update #3223? Or what do you think? |
|
I can update #3223, but it may take few days to find the time :) |
|
I'm not sure what exactly @matthid is asking @viktor-svub to update. It might be this, but I'll point out anyway in case it's not.
That said, I think we should aim to remove it completely in case we go with the microsoft package for the points @matthid already pointed out. @viktor-svub let me know if you need any help with that. |
This PR changes from
Mono.CeciltoAssemblyReaderto read the assemblies to find out if binding redirects is needed.paket installresults: