LS-proper-proper-fix-I-swear

[fredg1]

:,(

[heeheehee-kolmafia]

I can also generate a debug log with the exciting script

script;

I suggest being more methodical than "hey look script validation no longer fails on this specific case".

[fredg1]

I can also generate a debug log with the exciting script

script;

I suggest being more methodical than "hey look script validation no longer fails on this specific case".

I tried to be thorough; I looked at everything calling makeComment and makeToken, and saw that specific case.
It's just that I focused on the endIndex = line.length();, and didn't think of also looking into the endIndex = line.indexOf(";"); when I ruled out that the former was safe ;-;

[fredg1]

Please don't think I'm botching this ;A;

[heeheehee-kolmafia]

I think these are all the changes we need. Might be nice to have a second opinion, though.

@BadHorseMonkey?

[BadHorseMonkey]

I think these are all the changes we need. Might be nice to have a second opinion, though.

@BadHorseMonkey?

Let me fire up my Fredg directory and check out the branch...

[BadHorseMonkey]

I think these are all the changes we need. Might be nice to have a second opinion, though.
@BadHorseMonkey?

Let me fire up my Fredg directory and check out the branch...

builds for me. I ran

cli_execute( "validate VeracitySpacegate.ash" );
print( "veracity0-spacegate" + ": " + svn_exists( "veracity0-spacegate" ));

on 28 scripts in my scripts directory. I'm currently running my Autoscend evening turns in it. Nothing I know of is broken.

It occurs to me that if we're revamping the parser, one thing we'll want is a library of good and bad files to test with, which can prove that our parser is acting correctly end-to-end. It wouldn't be unit tests, but complementary to them. Functional tests are good at finding things like this.

[heeheehee-kolmafia]

Let me fire up my Fredg directory and check out the branch...

fyi you can fetch PR branches via git:

git fetch origin pull/164/head:new-branch-for-the-pr
git switch new-branch-for-the-pr

Anyways. I'll merge this fix, then. I couldn't find anything else when staring at the code.