Fix SA1629 "... should end with a period" in generated xml docs

[georg-jung]

Probably not adding too much value, but easy to change too :)...

Also, <remarks><para>... is generated like:

/// <para><see href="https://docs.microsoft.com/windows/win32/api/iphlpapi/nf-iphlpapi-sendarp">Learn more about this API from docs.microsoft.com.</see></para>

while the end of <summary> is generated as:

/// <para><see href="https://docs.microsoft.com/windows/win32/api/iphlpapi/nf-iphlpapi-sendarp#parameters">Read more on docs.microsoft.com</see>.</para>

thus, beeing SA1629 compliant.

[ghost]

All CLA requirements met.

[jnm2]

This reverses #295 and is visually irritating to me. I've been thinking about contributing a fix for SA1629. Either way, SA1629 doesn't run on generated source from what I remember.

[georg-jung]

I think you're right, I didn't notice it directly on generated source code, but when experimenting with an editable copy of the generated code directly in a project. Where the . should be is probably a question of taste - I don't have strong feelings about it. I just sent the PR because I randomly noticed that the code that is generated doesn't stick to the default StyleCop ruleset and it's easy to fix.

Either way, we currently have the point inside the <see> in one place and outside in the other (see above).

[jnm2]

Nice, I would have liked to have fix that if I had seen it.

[AArnott]

Thanks for the submission. I typically strive for StyleCop compliance in the generated code. Even though those analyzers don't typically run on generated code, copying from generated code into the user's project is a supported scenario, and we don't want that to require a ton of fixups after in order to fix stylecop warnings.
That said, I know we're probably far from that sort-of goal (of being stylecop compliant).

I'd generally prioritize avoiding analyzer warnings over catering to what @jnm2 finds "visually irritating" though (no offense intended). So I think we should take this. If @jnm2 succeeds in getting a change made to SA1929 such that either way will not produce a warning, I guess we can switch back at that point.

[jnm2]

Hey, no worries. The fact that no one else seems to be bothered by the period being a different color is a sign that I should consider it's just a "me" thing and not switch things back even after SA1929 is updated.

[jnm2]

Hey, <a href instead of <see href would not have had this conflict with SA1629. It's also shorter. <a started working at the same time in the IDE that <see href started working.

PR to update SA1629 is at DotNetAnalyzers/StyleCopAnalyzers#3371.