Replace std::hash_map/hash_set with std::unordered_map/unordered_set on trunk (C++11)GCC 15 removed std::hash_map/std::hash_set (GNU extensions) from libstdc++.Replace all 628 occurrences across 4
Replace std::hash_map/hash_set with std::unordered_map/unordered_set on trunk (C++11)GCC 15 removed std::hash_map/std::hash_set (GNU extensions) from libstdc++.Replace all 628 occurrences across 477 files with std::unordered_map/std::unordered_set, updating #include <hash_map>/<hash_set> toAffects: vcl, crashrep, extensions, dbaccess, scripting, sw, io, rdbmaker,codemaker, unodevtools, and many other modules.
show more ...
Replace std::auto_ptr with std::unique_ptrOwnership transfers that auto_ptr did silently are now explicitstd::move. Friends that gave auto_ptr access to a protected destructorname std::default_de
Replace std::auto_ptr with std::unique_ptrOwnership transfers that auto_ptr did silently are now explicitstd::move. Friends that gave auto_ptr access to a protected destructorname std::default_delete instead.
Drop register and declare the copies that C++11 deprecates
Add ODF 1.4 and 1.4 Extended save formats
ODF 1.2 conformance and ODF 1.3 supportThe "ODF 1.2" save mode writes only what the ODF 1.2 schema and rules allow; "ODF 1.2 Extended" stays the default. New "ODF 1.3" and "ODF 1.3 Extended" modes
ODF 1.2 conformance and ODF 1.3 supportThe "ODF 1.2" save mode writes only what the ODF 1.2 schema and rules allow; "ODF 1.2 Extended" stays the default. New "ODF 1.3" and "ODF 1.3 Extended" modes write the ODF 1.3 forms of what AOO wrote as extensions. Calc gets the ODF 1.3 formula changes. number:min-decimal-places and number:fill-character are read, and min-decimal-places is written in the 1.3 modes.Protection keys are written as SHA-256, as ODF 1.2 describes; older versions cannot unprotect such a sheet.
About dialog: load the AOO5 about5.png image, falling back to about.png
Win10 msvc trunk (#498)Build trunk on Windows with a modern MSVCconfigure.ac required VS2008, but the bundled dependencies had moved pastwhat VC9 can compile -- boost 1.84 needs C++11, Python 3.
Win10 msvc trunk (#498)Build trunk on Windows with a modern MSVCconfigure.ac required VS2008, but the bundled dependencies had moved pastwhat VC9 can compile -- boost 1.84 needs C++11, Python 3.11 needs VS2017 ornewer -- so the configured compiler could not build the configureddependencies, and no combination of the two worked. Every other platform hadalready moved; Windows was the one left pinned.This branch accepts a VS2015-or-newer toolset and the Windows 10 SDK, thenfixes what a modern compiler and SDK find: export lists, entry points, CRTand manifest handling, and source conformance across the tree. Everydependency builds, including Python, which trunk has never built on Windowsat all. The CLI/.NET binding is ported from Managed C++ to C++/CLI and itsbridgetest passes.configure, bootstrap and build.pl now run to completion, andinstsetoo_native produces openoffice450.msi and openoffice450sdk.msi.coinMp has been updated to the latest Version.What this does not claim: the office has not been started, andtesttools/source/bridgetest still fails getCaughtException() in the x86 C++bridge. Python is built without _ssl, _ctypes, _bz2, _lzma and _sqlite3, andPlatform SDK v7.0 is still required alongside the Windows 10 SDK. No otherplatform's build is changed.Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>Claude-Session: https://claude.ai/code/session_01UWgzQY2r1XwFvgeLPFWPsi
crashrep/fpicker/cui: Win64 (x64) source fixesBackport the Win64 source fixes from origin/windows-amd64 (dmake-erareference); meaningful hunks only -- the reference branch carries heavywhitespace
crashrep/fpicker/cui: Win64 (x64) source fixesBackport the Win64 source fixes from origin/windows-amd64 (dmake-erareference); meaningful hunks only -- the reference branch carries heavywhitespace churn -- and each compiles unchanged on x86 (dmake-backportable).- crashrep soreport.cpp (8740daf50e): GWL_ -> GWLP_ + *LongPtr, dialog procs BOOL -> INT_PTR, CONTEXT.Eip/Ebp -> Rip/Rbp (guarded by INTEL/X86_64), DWORD -> DWORD64 for SymGetSymFromAddr64.- fpicker FileOpenDlg + WinFileOpenImpl (73ea0802dd, b077179428): same GWL_ -> GWLP_ subclass cluster, OFN hook-proc returns -> UINT_PTR, lCustData/lResult LONG -> LONG_PTR/LRESULT pointer-truncation fixes.- cui optdict.cxx (93b3eead64): 0LU -> sal_uLong(0LU), resolving a 0-vs-NULL overload ambiguity (int-type vs pointer overload) that only bites once sal_uLong widens on x64.Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>(cherry picked from commit 25104ebbcafb3621565a09fbe219747935cb24c9)(cherry picked from commit d14d772a5bd66969dee8b8ef850cba39d09a196f)
Fix Wrong order of First/Last Name in Hungarian UI:https://bz.apache.org/ooo/show_bug.cgi?id=105342
Cleanup
Fix botton -> bottom
Use 'zeros' instead of 'zeroes', at least in UI
misc: fix spelling (#461)* misc: fix spelling* Regenerate ignored words list codespell.txt
Update copyright year to 2026
Typo run (#435)* Fix some typos* Regenerate ignored words list codespell.txt
Make dialog wider
Cleanup resource files
chore: whitespace clean up for `.src` files (#429)Enforced 3 hooks for `.src` files- end-of-file-fixer- mixed-line-ending- trailing-whitespace
Better use a GroupBox instead of a weird Button
pre-commit whitespace cleanup for `.component` files (#398)Enforced 3 hooks for `.component` files:- end-of-file-fixer- mixed-line-ending- trailing-whitespace
12345678910>>...12