Undid linter

Note -- this patch also contains some stuff the linter picked up.

Ok, this looks exceptionally well crafted to me. Approved.

Added some comments.

Overall this looks good and I agree with this entire class and with the usage and code. See my in-line comments.

In D470#8257, @deadalnix wrote:

Then review it.

Added some useful output to end of configure script to indicate which sanitizers, if any, are enabled

Note: it works. Here is some sample output.

In D474#8239, @deadalnix wrote:

Can you try to introduce errors on purpose in the code to check that the sanitizers are properly used ?

Modified to be minimally required expression syntax to have all types auto-promoted to int64

I can't apply the patch.. :/

In T38#1292, @schancel wrote:

Do we specifically want to use buildbot? Or is that being used loosely? How many different build configurations do we need?

I can think of something ultra-simple that meets these requirements:

Ok, well, then I'll close this revision.

@freetrader so you're decidedly against changing copyright notices to include us?

Oops.. a few files already had the (c) message so it was duplicated. Removed duplicates.

What da fuq arcanist?!?!

Question was also asked by me in D466

Better formatting (also implemented deadalnix's recommendations)

Thanks for the feedback. I just did a quick pass to change it over -- but your readability feedback is good. I'll change it and update the revision.

ok

This looks ok to me.. I'm accepting.

Ok, I went over everything. Looks good to me.

I can't apply the patch.. fails against master.. what did you do, deadalnix?

Went over all the changes by eye carefully and also compiled. Works and looks fine.

looks good, accept

WTF.. arc messed this up. Why?!

Forgot one, thanks freetrader

I'll update

OK, perfect. Works as intended.

This looks correct to me and I can't think of any place you "missed a spot".

In D457#7844, @freetrader wrote:

No, you're right, -noprefcashpeer is not working as intended. I'd need to elaborate the condition you pointed at.
If we want to keep the option, that is.

I like the addition of the command-line argument. I can imagine use cases where someone might want this turned off explicitly (say, when synching up or whatnot). We should leave the command-line argument in.

Yeah it's just a good habit because some compilers can warn about format types not matching arguments if you tag the function as being a printf-style function.

In D452#7811, @freetrader wrote:

Nice, thanks. Accepted on code review basis, @CCulianu you are responsible if this breaks Mac users :-)

FYI since you guys don't have MacOS -- here is proof that this thing builds on Mac:

Ok, after talking to deadalnix about it, he convinced me the netsplit risk isn't so large and that we want to be strict about tx format -- so rejecting the TX (which leads to a ban) is what we're going to do.

In D442#7735, @thezerg wrote:

A ban is a big thing. Why not be tolerant for a month or two more to avoid risking a network split?

Looks perfect, and works on top of that (all extended tests pass).

Got rid of redundant check for MoneyValue (AmountFromValue already did that), expanded on comments

Oops, missed a set of braces

Added changes as per reviewer comments

Forgot 1 line

Added an inline comment.. will review more in-depth tomorrow (to make sure it doesn't break any testsm python tests, etc), but otherwise it looks correct

Looks good and it's what we need, especially for D444

Oops! I didn't realize D440 was part of this stack. You should have mentioned that. i totally didn't see it.

I think you need to update the txvalidationcache_tests -- right now C++ tests are failing with this patch, because the test case doesn't construct replay protected tx's.

Updated permissions on signrawtransaction.py to +x