Sanitize tss version number too

[mcr]

This replaces PR #578.
It is rebased after ther DOCS_RS changes.
Without this patch, the version number does not parse with VersionReq::parse.

(This depends upon #603, please merge that first)

[Superhepper]

I am not against this. But I would like for the sanitize of the TSS version to be opt-in by using environment variable or by some other means.

[mcr]

I am not against this. But I would like for the sanitize of the TSS version to be opt-in by using environment variable or by some other means.

okay, that sounds okay to me.
Please name the environment variable for me... :-)
I think that I also build tss statically, and I believe I had to patch it to make it all work without .so.
(when debugging, the wrong .so always winds up loading when you least expect it. That's neither here nor there, but I'm thinking about the documentation update on the environment variable)

[Superhepper]

Every other environment variable that has to deal with tpm2-tss is named something with TPM2_TSS_*. So something like TPM2_TSS_VERSION_ALLOW_DIRTY
TPM2_TSS_VERSION_DIRTY
TPM2_TSS_VERSION_ALLOW_PRE_VERSION
TPM2_TSS_IS_PRE_VERSION

Interesting I tried to build it with static linkage when I finished bundled. But it a bit more complicated so I left it for the future. But if it would be possible to have static linkage on all platforms I think that would be great.

[mcr]

-------- Jesper Brynolf ***@***.***> wrote: Interesting I tried to build it with static linkage when I finished bundled. But it a bit more complicated so I left it for the future. But if it would be possible to have static linkage on all platforms I think that would be great. Both swtpm and libtss need a few small patches to build completely statically.

[mcr]

I went with: $TPM2_TSS_VERSION_IGNORE_PRERELEASE
Should I put this into the README.md?
Or maybe a new BUILDING.md?