noham/archived-Reclass
1
0
Fork 0
mirror of https://github.com/NohamR/Reclass.git synced 2026-05-10 19:59:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: fixed-width float formatting, fix test_32bit_support on Linux CI

Browse Source 
Float values now use a fixed 7-char body (digits.decimals + f suffix)
that adapts decimal places to the integer magnitude. Removes the
variable-width 'g' format and sign-space prefix.

Set QT_QPA_PLATFORM=offscreen for test_32bit_support so it no longer
crashes on headless Linux CI without an X display.
This commit is contained in:
IChooseYOu
2026-03-01 14:02:40 -07:00
committed by IChooseYou
parent ed8a44917b
commit cae599a0c6
3 changed files with 67 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
src/format.cpp
View File

@@ -80,38 +80,38 @@ static QString rawHex(uint64_t v, int digits) {
return QString::number(v, 16).rightJustified(digits, '0');
}
QString fmtInt8(int8_t v) { return hexVal((uint8_t)v); }
QString fmtInt16(int16_t v) { return hexVal((uint16_t)v); }
QString fmtInt32(int32_t v) { return hexVal((uint32_t)v); }
QString fmtInt64(int64_t v) { return hexVal((uint64_t)v); }
QString fmtInt8(int8_t v) { return QString::number(v); }
QString fmtInt16(int16_t v) { return QString::number(v); }
QString fmtInt32(int32_t v) { return QString::number(v); }
QString fmtInt64(int64_t v) { return QString::number((qlonglong)v); }
QString fmtUInt8(uint8_t v) { return hexVal(v); }
QString fmtUInt16(uint16_t v) { return hexVal(v); }
QString fmtUInt32(uint32_t v) { return hexVal(v); }
QString fmtUInt64(uint64_t v) { return hexVal(v); }
QString fmtFloat(float v) {
// Fixed 7-char body: digits + "." + decimals + "f"
// Negative values get a '-' prefix (8 chars total), positive stay 7.
if (std::isnan(v)) return QStringLiteral("NaN");
if (std::isinf(v)) return v > 0 ? QStringLiteral("inff") : QStringLiteral("-inff");
// 6 significant digits — covers full single-precision range
QString s = QString::number(v, 'g', 6);
float av = std::fabs(v);
if (av >= 100000.f)
return v < 0 ? QStringLiteral("-99999+f") : QStringLiteral("99999+f");
// If 'g' chose scientific notation, reformat as plain decimal
if (s.contains('e') || s.contains('E')) {
s = QString::number(v, 'f', 8);
if (s.contains('.')) {
int i = s.size() - 1;
while (i > 0 && s[i] == '0') i--;
if (s[i] == '.') i++; // keep at least one decimal digit
s.truncate(i + 1);
// body = digits + "." + decimals + "f", target exactly 7 chars.
// Start with max decimals, reduce if integer part is wide or rounding overflows.
for (int dec = 4; dec >= 0; dec--) {
QString body = QString::number(av, 'f', dec);
body += (dec == 0) ? QStringLiteral(".f") : QStringLiteral("f");
if (body.size() == 7) {
if (v < 0.f) body.prepend('-');
return body;
}
}
if (!s.contains('.'))
s += QStringLiteral(".f");
else
s += QLatin1Char('f');
return s;
// Rounding pushed past 99999 — use overflow cap
return v < 0 ? QStringLiteral("-99999+f") : QStringLiteral("99999+f");
}
QString fmtDouble(double v) {
QString s = QString::number(v, 'g', 6);
@@ -333,7 +333,7 @@ static QString readValueImpl(const Node& node, const Provider& prov,
int count = sizeForKind(node.kind) / 4;
QStringList parts;
for (int i = 0; i < count; i++)
parts << fmtFloat(prov.readF32(addr + i * 4)).trimmed();
parts << fmtFloat(prov.readF32(addr + i * 4));
return parts.join(QStringLiteral(", "));
}
case NodeKind::Mat4x4: {
@@ -342,7 +342,7 @@ static QString readValueImpl(const Node& node, const Provider& prov,
QString line = QStringLiteral("row%1 [").arg(subLine);
for (int c = 0; c < 4; c++) {
if (c > 0) line += QStringLiteral(", ");
line += fmtFloat(prov.readF32(addr + (subLine * 4 + c) * 4)).trimmed();
line += fmtFloat(prov.readF32(addr + (subLine * 4 + c) * 4));
}
line += QStringLiteral("]");
return line;