Strange (but positive) output

[alexmurari]

Hi, Ki!

After the last commits from 03/05/2016, after confusing a simple exe/dll solution, the output size of the executable assembly is relatively low: 94KB (before 175KB~, I didn't notice any difference in DLLs) and the runtime speed was greatly increased compared to the previous commits.

I do not use compressor.

Until now it's only great news: runtime speed increased, size decreased, everything's working fine.

I browsed the assembly with dnSpy and apparently all protections are there. Maybe I'm wrong because this version is outputing a assembly with half of the size of the previous version.

I use all protections except resources.

What happened?

[yck1509]

Hi, It seems strange, since I did not change much related to protections. Do you have some sample showing the difference between two version of output?

[alexmurari]

Go to the CI Server and download this version and this version.

Protect any console app with or without an class library. You'll se the difference in the output size in these two versions of ConfuserEx. Use this setting.

<rule pattern="true" inherit="false">
    <protection id="anti ildasm" />
    <protection id="anti tamper" action="remove">
    </protection>
    <protection id="constants">
      <argument name="mode" value="x86" />
      <argument name="decoderCount" value="10" />
      <argument name="cfg" value="false" />
      <argument name="elements" value="SPI" />
    </protection>
    <protection id="ctrl flow">
      <argument name="predicate" value="expression" />
      <argument name="intensity" value="75" />
      <argument name="depth" value="8" />
      <argument name="junk" value="true" />
    </protection>
    <protection id="anti dump" />
    <protection id="anti debug">
      <argument name="mode" value="antinet" />
    </protection>
    <protection id="invalid metadata" />
    <protection id="ref proxy">
      <argument name="mode" value="strong" />
      <argument name="encoding" value="expression" />
      <argument name="initCount" value="16" />
      <argument name="depth" value="6" />
      <argument name="internal" value="false" />
      <argument name="typeErasure" value="false" />
    </protection>
    <protection id="resources" action="remove"/>
    <protection id="rename">
      <argument name="renXaml" value="false" />
      <argument name="renPdb" value="true" />
    </protection>
  </rule>

[yck1509]

Hi, It turns out when I modified ObfAttrMarker, I made a mistake that all injected helpers are marked incorrectly so that they don't have any protection applied. 2b4f482 should fix them.