From 065c4b443bcbeee02c6c6bb18cb1cc651e3fcf2b Mon Sep 17 00:00:00 2001
From: Tomohiro <gpuppur@gmail.com>
Date: Fri, 29 Aug 2025 04:56:46 +0900
Subject: [PATCH] fixes #25125 (#25126)

`strutils.formatSize` returns correct strings from large values close to
`int64.high`.
Round down `bytes` when it is converted to float.
---
 lib/pure/strutils.nim      | 43 +++++++++++++-------------
 tests/stdlib/tstrutils.nim | 63 ++++++++++++++++++++++++++++++++++++--
 2 files changed, 82 insertions(+), 24 deletions(-)

diff --git a/lib/pure/strutils.nim b/lib/pure/strutils.nim
index c218ac1c53..1a5d4d5d6f 100644
--- a/lib/pure/strutils.nim
+++ b/lib/pure/strutils.nim
@@ -74,6 +74,7 @@ import std/parseutils
 from std/math import pow, floor, log10
 from std/algorithm import fill, reverse
 import std/enumutils
+from std/bitops import fastLog2
 
 from std/unicode import toLower, toUpper
 export toLower, toUpper
@@ -2639,37 +2640,35 @@ func formatSize*(bytes: int64,
   ## * `strformat module<strformat.html>`_ for string interpolation and formatting
   runnableExamples:
     doAssert formatSize((1'i64 shl 31) + (300'i64 shl 20)) == "2.293GiB"
-    doAssert formatSize((2.234*1024*1024).int) == "2.234MiB"
+    doAssert formatSize((2.234*1024*1024).int) == "2.233MiB"
     doAssert formatSize(4096, includeSpace = true) == "4 KiB"
     doAssert formatSize(4096, prefix = bpColloquial, includeSpace = true) == "4 kB"
     doAssert formatSize(4096) == "4KiB"
-    doAssert formatSize(5_378_934, prefix = bpColloquial, decimalSep = ',') == "5,13MB"
+    doAssert formatSize(5_378_934, prefix = bpColloquial, decimalSep = ',') == "5,129MB"
 
-  const iecPrefixes = ["", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi", "Yi"]
-  const collPrefixes = ["", "k", "M", "G", "T", "P", "E", "Z", "Y"]
-  var
-    xb: int64 = bytes
-    fbytes: float
-    lastXb: int64 = bytes
-    matchedIndex = 0
-    prefixes: array[9, string]
+  # It doesn't needs Zi and larger units until we use int72 or larger ints.
+  const iecPrefixes = ["", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei"]
+  const collPrefixes = ["", "k", "M", "G", "T", "P", "E"]
+
+  let lg2 = if bytes == 0:
+      0
+    else:
+      when hasWorkingInt64:
+        fastLog2(bytes)
+      else:
+        fastLog2(int32 bytes)
+  let matchedIndex = lg2 div 10
+  # Lower bits that are smaller than 0.001 when `bytes` is converted to a real number and added prefix, are discard.
+  # Then it is converted to float with round down.
+  let discardBits = (lg2 div 10 - 1) * 10
+
+  var prefixes: array[7, string]
   if prefix == bpColloquial:
     prefixes = collPrefixes
   else:
     prefixes = iecPrefixes
 
-  # Iterate through prefixes seeing if value will be greater than
-  # 0 in each case
-  for index in 1..<prefixes.len:
-    lastXb = xb
-    xb = bytes div (1'i64 shl (index*10))
-    matchedIndex = index
-    if xb == 0:
-      xb = lastXb
-      matchedIndex = index - 1
-      break
-  # xb has the integer number for the latest value; index should be correct
-  fbytes = bytes.float / (1'i64 shl (matchedIndex*10)).float
+  let fbytes = if lg2 < 10: bytes.float elif lg2 < 20: bytes.float / 1024.0 else: (bytes shr discardBits).float / 1024.0
   result = formatFloat(fbytes, format = ffDecimal, precision = 3,
       decimalSep = decimalSep)
   result.trimZeros(decimalSep)
diff --git a/tests/stdlib/tstrutils.nim b/tests/stdlib/tstrutils.nim
index db9fb80c44..a21a337bd5 100644
--- a/tests/stdlib/tstrutils.nim
+++ b/tests/stdlib/tstrutils.nim
@@ -789,12 +789,71 @@ bar
   block: # formatSize
     disableVm:
       when hasWorkingInt64:
+        doAssert formatSize(1024 * 1024 * 1024 * 2 - 1) == "1.999GiB"
+        doAssert formatSize(1024 * 1024 * 1024 * 2) == "2GiB"
         doAssert formatSize((1'i64 shl 31) + (300'i64 shl 20)) == "2.293GiB" # <=== bug #8231
-      doAssert formatSize((2.234*1024*1024).int) == "2.234MiB"
+        doAssert formatSize(int64.high) == "7.999EiB"
+        doAssert formatSize(int64.high div 2 + 1) == "4EiB"
+        doAssert formatSize(int64.high div 2) == "3.999EiB"
+        doAssert formatSize(int64.high div 4 + 1) == "2EiB"
+        doAssert formatSize(int64.high div 4) == "1.999EiB"
+        doAssert formatSize(int64.high div 8 + 1) == "1EiB"
+        doAssert formatSize(int64.high div 8) == "1023.999PiB"
+        doAssert formatSize(int64.high div 16 + 1) == "512PiB"
+        doAssert formatSize(int64.high div 16) == "511.999PiB"
+      doAssert formatSize(0) == "0B"
+      doAssert formatSize(0, includeSpace = true) == "0 B"
+      doAssert formatSize(1) == "1B"
+      doAssert formatSize(2) == "2B"
+      doAssert formatSize(1022) == "1022B"
+      doAssert formatSize(1023) == "1023B"
+      doAssert formatSize(1024) == "1KiB"
+      doAssert formatSize(1025) == "1.001KiB"
+      doAssert formatSize(1026) == "1.002KiB"
+      doAssert formatSize(1024 * 2 - 2) == "1.998KiB"
+      doAssert formatSize(1024 * 2 - 1) == "1.999KiB"
+      doAssert formatSize(1024 * 2) == "2KiB"
+      doAssert formatSize(1024 * 2 + 1) == "2.001KiB"
+      doAssert formatSize(1024 * 2 + 2) == "2.002KiB"
+      doAssert formatSize(4096 - 1) == "3.999KiB"
       doAssert formatSize(4096) == "4KiB"
+      doAssert formatSize(4096 + 1) == "4.001KiB"
+      doAssert formatSize(1024 * 512 - 1) == "511.999KiB"
+      doAssert formatSize(1024 * 512) == "512KiB"
+      doAssert formatSize(1024 * 512 + 1) == "512.001KiB"
+      doAssert formatSize(1024 * 1024 - 2) == "1023.998KiB"
+      doAssert formatSize(1024 * 1024 - 1) == "1023.999KiB"
+      doAssert formatSize(1024 * 1024) == "1MiB"
+      doAssert formatSize(1024 * 1024 + 1) == "1MiB"
+      doAssert formatSize(1024 * 1024 + 1023) == "1MiB"
+      doAssert formatSize(1024 * 1024 + 1024) == "1.001MiB"
+      doAssert formatSize(1024 * 1024 + 1024 * 2) == "1.002MiB"
+      doAssert formatSize(1024 * 1024 * 2 - 1) == "1.999MiB"
+      doAssert formatSize(1024 * 1024 * 2) == "2MiB"
+      doAssert formatSize(1024 * 1024 * 2 + 1) == "2MiB"
+      doAssert formatSize(1024 * 1024 * 2 + 1024) == "2.001MiB"
+      doAssert formatSize(1024 * 1024 * 2 + 1024 * 2) == "2.002MiB"
+      doAssert formatSize(1024 * 1024 * 4 - 1) == "3.999MiB"
+      doAssert formatSize(1024 * 1024 * 4) == "4MiB"
+      doAssert formatSize(1024 * (1024 * 4 + 1)) == "4.001MiB"
+      doAssert formatSize(1024 * 1024 * 512 - 1025) == "511.998MiB"
+      doAssert formatSize(1024 * 1024 * 512 - 1) == "511.999MiB"
+      doAssert formatSize(1024 * 1024 * 512) == "512MiB"
+      doAssert formatSize(1024 * 1024 * 512 + 1) == "512MiB"
+      doAssert formatSize(1024 * 1024 * 512 + 1024) == "512.001MiB"
+      doAssert formatSize(1024 * 1024 * 512 + 1024 * 2) == "512.002MiB"
+      doAssert formatSize(1024 * 1024 * 1024 - 1) == "1023.999MiB"
+      doAssert formatSize(1024 * 1024 * 1024) == "1GiB"
+      doAssert formatSize(1024 * 1024 * 1024 + 1) == "1GiB"
+      doAssert formatSize(1024 * 1024 * 1025) == "1.001GiB"
+      doAssert formatSize(1024 * 1024 * 1026) == "1.002GiB"
+      # != 2.234MiB as (2.234 * 1024 * 1024).int.float / (1024 * 1024) = 2.23399...
+      # and formatSize round down the value
+      doAssert formatSize((2.234*1024*1024).int) == "2.233MiB"
       doAssert formatSize(4096, prefix = bpColloquial, includeSpace = true) == "4 kB"
       doAssert formatSize(4096, includeSpace = true) == "4 KiB"
-      doAssert formatSize(5_378_934, prefix = bpColloquial, decimalSep = ',') == "5,13MB"
+      # (5378934).float / (1024 * 1024) = 5.12975...
+      doAssert formatSize(5_378_934, prefix = bpColloquial, decimalSep = ',') == "5,129MB"
 
   block: # formatEng
     disableVm: