fix(7407): added check for x1a in fread.c to avoid segfault

[manmita]

Closes #7407

Adding ch ==eof to end_of_field to fix seg fault at fread
Removed single lined condition code from end_of_field

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.99%. Comparing base (b472d86) to head (868e5be).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7570   +/-   ##
=======================================
  Coverage   98.99%   98.99%           
=======================================
  Files          87       87           
  Lines       16729    16733    +4     
=======================================
+ Hits        16561    16565    +4     
  Misses        168      168           

☔ 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.

[ben-schwen]

The idea of the code above was that the complicated single line OR statement gets easier to grasp, so we would favor removing line 352 and add an case for if (*ch == '\x1A') return true;

[manmita]

Got it, in the file that code was written after the long single line one so it was dead code, hence removed it but I'll do the reverse,

keep those lines and remove this one

[ben-schwen]

Suggested change

	return *ch == sep || *ch == '\x1A'|| ((uint8_t)*ch <= 13 && (ch == eof || eol(&ch))) || (commentChar && *ch == commentChar);
	if (*ch == sep) return true;
	if (ch == eof) return true; // Check eof first to avoid reading past #7407
	if ((uint8_t)*ch <= 13 && eol(&ch)) return true;
	if (!commentChar) return false;
	return *ch == commentChar;

While I couldn't come up with another example than 0x1A it seems safer to check for eof instead of checking for 0x1A specifically

[manmita]

yeah, adding eof should also work, the bug was that ch==eof was in and with ch<=13 but x1A is greater

[ben-schwen]

Maybe something like

fread(text=) could segfault when reading text input ending with a \x1a (ASCII SUB) character after a long line,

[ben-schwen]

We only use { in tests if they make the tests clearer and concise.

Setting up the str with txt = paste0("foo\n", strrep("a", 4096 * 100), "\x1A") is fine.
For the test maybe test for nchar(fread(txt)) which ensure that the whole string is read?