Hints.txt

Some non-obvious aspects of RESearch usage are explained here.

- Literals like '\n' DO NOT work in plain text search! Either use regexp, or enter needed
  symbol directly using Ctrl-Q.

- '\n' and others do work in replace string even with plain-text search, UNLESS
  "[x] Use escapes in plain-text replace" is turned off.

- Single-line search in both editor and panels, both regexp and plaintext, is done
  by extracting lines from text sequentially, WITHOUT newline symbols - therefore to search for
  the word 'end' at the end of string, use regexp 'end$', not 'end\n'!

- Several-line search in editor is done by internally concatenating lines with a single '\n',
  so it is to be used instead of '\r\n' always, UNLESS you turn on "[X] Use real end-of-line sequences".

- Several-line search from panels is done in the same way - use a single '\n'.

- Multi-line search/replace from panels feeds the whole file to RegExp engine, therefore
  actual newline symbols are to be used. '\r?\n' is good for most cases.

- Or, you can use '\R' and don't bother about newline sequences at all.

- Multi-line replace from panels has some file size limits: 256Mb for x86 and 1Tb for x64
  If file exceeds it, several-line replace is used. Also, if file is below the limit but buffer
  for the original and decoded file cannot be allocated, file is processed in chunks without
  overlap

- If '[ ] Dot matches newline also' is not set, you can use '\C' as 'Absolutely any symbol,
  including newline'

- If zero-width string is found, characters surrounding it are highlighted.

- If UTF-16LE/BE or UTF-8 byte-order mark (BOM) is found while working from Panels,
  no other encodings are tried!

- Without '[ ] Use all character tables', panel operations use the following:
  1. If BOM is found, that encoding is used
  2. In Far 1.x (ANSI) version, OEM codepage is used
  3. In Far 2/3 (Unicode) version, codepage selected as default is used (OEM or ANSI)

- In Replace from panels, only default codepage is used.