Normalize line endings

Good catch but two things;

function _pimport_normalize_line_endings($s)

$s is not a variable name - it's a letter ;o) Use $string, $input, etc - it makes the code easier for other people to read. Meaningful variable names really do help with debugging, module names however - yeah they can be as random as you like ;o)

$s = str_replace("\r\n", "\n", $s);
$s = str_replace("\r", "\n", $s);

This doesn't need duplication, we could just as easily use;

$s = str_replace(array("\r\n", "\r"), "\n", $s);

It still reads fine.

Thanks,

Pobster

Looks good to me, will test it tonight and commit it (unless Marcus can test and commit it sooner?)

* Setting to reviewed and tested because I can see that it'll work but I'll still actually physically test it later! ;o)

Thanks,

Pobster

Apologies, I completely forgot to commit this, even though I did actually remember to test it!

I see that you've already committed it (and rightly so!) so marking this as fixed.

Thanks,

Pobster

I've just realised that this patch means that blank lines are removed before the file is processed which means the row counting doesn't work properly for reporting. I think it's more important to give accurate row numbers for problems than to save processing on the import.

I've taken out the code that strips the blank lines and comitted.

Just remove this one line;

+  // Don't allow out-of-control blank lines
+  $string = preg_replace("/\n{2,}/", "\n\n", $string);

Pobster
edit: Oh umm... READ Pobster, READ... You said that already ;o)