Sort Lexographically Ascending: Bad Allocation

[stevemk14ebr]

Description of the Issue

With a sufficiently large text file (1.5 million lines of sha256 hashes, 64 characters each), line operation sort lexographically ascending will 100% of the time fail with a bad allocation warning and then notepad++ will crash with a data loss warning.

Steps to Reproduce the Issue

1. Open Big file
2. Sort
3. Don't Profit

Expected Behavior

Sorting to not crash

Actual Behavior

Crash

Debug Information

Notepad++ v8.1.1 (32-bit)
Build time : Jul 1 2021 - 14:04:28
Path : C:\Program Files (x86)\Notepad++\notepad++.exe
Command Line : "\wsl$\Debian\home\steve\source\repos\Flareup2\utility_scripts\sha256.txt"
Admin mode : OFF
Local Conf mode : OFF
Cloud Config : OFF
OS Name : Windows 10 Enterprise (64-bit)
OS Version : 2009
OS Build : 19043.1165
Current ANSI codepage : 1252
Plugins : ComparePlugin.dll mimeTools.dll NppConverter.dll NppExport.dll NPPJSONViewer.dll NppTextFX.dll

[chcg]

@felix-22
Is this also happening for the x64 version of N++?
Might be related to 32bit process limit of 2 GB.

[mshingote]

Sharing some observations for discussion.

In following file there is one virtual function sort which takes collection of lines as a copy.
...notepad-plus-plus\PowerEditor\src\MISC\Common\Sorters.h:74
virtual std::vector<generic_string> sort(std::vector<generic_string> lines) = 0;

One sorting case from following file
...notepad-plus-plus\PowerEditor\src\ScintillaComponent\ScintillaEditView.cpp

std::vector<generic_string> splitText = stringSplit(text, getEOLString());
...
const std::vector<generic_string> sortedText = pSort->sort(splitText);

Lines are passed as a copy for sorting, seems there is room for memory allocation optimization by referencing existing list of lines.

I may be wrong as I'm not fully aware of end to end flow.