implement comment.char argument for fread

[ben-schwen]

Closes #856
Supersedes #4486

Needed this on machine without grep.
It should have no performance impact since we always check for comment.char before entering costly checking branch.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.12%. Comparing base (67129f0) to head (ba0c68a).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7375   +/-   ##
=======================================
  Coverage   99.12%   99.12%           
=======================================
  Files          85       85           
  Lines       16589    16637   +48     
=======================================
+ Hits        16444    16492   +48     
  Misses        145      145           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[MichaelChirico]

hm that's interesting. I don't know what I'd expect out of reading this file. E.g. data.table(a=1L, b=2L) also seems like a possible answer, and even data.table(a=c(1L, NA), b=c("2", "only comment")) could be justified too. Maybe let's just error if comment.char==sep?

[MichaelChirico]

Looks basically GTG. Main remaining thing I think is what to do about blank.lines.skip.

@eddelbuettel if you have any "real" files laying around you'd like to test on that would be most helpful :)

[eddelbuettel]

LOL @MichaelChirico -- while I wrote eight years ago

Bump. Needing this right now.

I am sorry to say that I have not been sitting here with baited breath and no, I do not have that file around. I do not even remember what the comment character may have been. I am sure the PR will be fine.

[MichaelChirico]

@Anirban166 it looks like some {pak} cacheing issue failing the atime GHA, any idea?

https://github.com/Rdatatable/data.table/actions/runs/18658671941/job/53193860953

[github-actions]

• HEAD=fread_commentChar stopped early for fread disk overhead improved in #6925
  

Generated via commit ba0c68a

Download link for the artifact containing the test results: ↓ atime-results.zip

Task	Duration
R setup and installing dependencies	6 minutes and 33 seconds
Installing different package versions	10 minutes and 2 seconds
Running and plotting the test cases	2 minutes and 40 seconds

[Anirban166]

Haven't seen that one before, but I think its a flaky error having to do with network/resource contention as it worked now that you re-runned (the r-lib/actions/setup-r-dependencies@v2 action that I'm using internally uses pak, so it would usually be an upstream error if one, but in cases of corrupted caches or missing files, we can use a different number for cache-version)

[MichaelChirico]

Thanks I'll try and notice if it recurs, indeed seems flaky. There were other failures related to missing branches, which clouded things. Seems restored to working now.

[MichaelChirico]

Great work!