Adapt db366f5c060b6b27ed8d340bc22fba30a017dcf2

Lower the C++ standard, as required on this branch

Ignore meaningless attributes

Stop recognizing the db:java-classpath attribute, and drop JavaDriverClassPath
and SystemProperties when importing data source settings.

Assisted-By: Claude Opus 5.5 <

Ignore meaningless attributes

Stop recognizing the db:java-classpath attribute, and drop JavaDriverClassPath
and SystemProperties when importing data source settings.

Assisted-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry picked from commit 430d1577c639ad8ba4ece164a179111734112fd9)

show more ...

svx, avmedia: put the media question when the URL is set, not while painting

Avoid multiple popups.

Keep the checks where they are, but have the question put earlier, in a
place that does not paint

svx, avmedia: put the media question when the URL is set, not while painting

Avoid multiple popups.

Keep the checks where they are, but have the question put earlier, in a
place that does not paint: when a media object gets its URL -- on insertion,
while the document loads, through the API -- ask the document that holds the
object. Its LinkManager keeps the answer, and the checks made later while a
view paints get it without a question. A copy takes the URL without asking
again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

show more ...

Fix acceptance of local files

Assisted-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry picked and adapted from commit bb3a5a855831b861943c55afcadf2a8cf42eff01)

Accept local files

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
(cherry picked and adapted from commit 06d7e6fcc50631716e4dd8909f5d630a5b1f52c5)

Prep for possible RC2

Update copyright year to 2026 - We did this _late_

svtools: check the WMF private-escape text length and advance array

Limit the character count of a PRIVATE_ESCAPE_UNICODE record to what a
String can hold, and ignore an advance array that has fewer

svtools: check the WMF private-escape text length and advance array

Limit the character count of a PRIVATE_ESCAPE_UNICODE record to what a
String can hold, and ignore an advance array that has fewer entries than
the text has characters, as if it were missing.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

show more ...

vcl: check bounds in the font subsetter

Make the CFF charstring interpreter's stack and array checks run-time
checks, and limit its output to the buffer it writes into.

Check the TrueType parser's

vcl: check bounds in the font subsetter

Make the CFF charstring interpreter's stack and array checks run-time
checks, and limit its output to the buffer it writes into.

Check the TrueType parser's table, glyph and cmap subtable offsets and
lengths against the tables they refer to.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

show more ...

svtools: one shared policy for following URLs out of document content

Each module that follows a URL named by document content -- linked graphics,
background images, instance data, media -- decided

svtools: one shared policy for following URLs out of document content

Each module that follows a URL named by document content -- linked graphics,
background images, instance data, media -- decided for itself whether to
check anything, and the check that does exist was written out inline at the
call site.

Add svtools/linkpolicy.hxx as the one place that decides, and apply it at the
load sites in editeng, forms, sfx2, sw, avmedia and svgio. A relative
reference and a file on this machine load as before; anything else asks the
running document, which is the question the LinkManager already asks.

Two places on those paths assumed success and no longer do: a reference that
cannot be resolved, and a graphic that is not there.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 5d0d2251b2545c6f290a9e81744e43c48ab6ba17)

show more ...

Read the catalog and schema from the new table object, not the table
descriptor used to create it. This avoids an exception when the
table descriptor's CatalogName property returns VOID instead of
a

Read the catalog and schema from the new table object, not the table
descriptor used to create it. This avoids an exception when the
table descriptor's CatalogName property returns VOID instead of
a string.

Patch by: me

(cherry picked from commit 6c6152cef718882dd13b1f34cd72a257d7ce94ca)

show more ...

java.lang.Boolean can't be cast to com.sun.star.uno.Any.
In Java, UNO's "Any" is java.lang.Object.

Patch by: me

(cherry picked from commit 7826359d221fc59714335fac155fb85bfeea0326)

Fix null pointer dereference

(cherry picked from commit c8c07ca898e41fafebb34bcd6a71572aca680e5d)

filter: idxf: port the input-validation hunks missed by the #491 backport

PR #491 was landed on AOO41X as 3c5a7184f8 ("cherry pick and adaptation
of merge commit c082241a9b"), but three hardening hu

filter: idxf: port the input-validation hunks missed by the #491 backport

PR #491 was landed on AOO41X as 3c5a7184f8 ("cherry pick and adaptation
of merge commit c082241a9b"), but three hardening hunks of the series did
not survive the adaptation:

* DXFBoundaryPathData::EvaluateGroup, group 93: the vertex count taken
from the file was used for `new DXFVector[n]` without checking the
reader status or the sign. A negative or unread count reached
operator new[]; a huge one threw std::bad_alloc out of the filter.
Check GetStatus() and n >= 0, and turn bad_alloc into a reader error.

* DXFHatchEntity::EvaluateGroup, group 91: identical problem for the
boundary-path count driving `new DXFBoundaryPathData[n]`.

* DXFGroupReader::ReadI: the hand-rolled do/while accumulated digits
without any overflow or termination check, so it also ran once on a
non-digit before testing. Parse with strtol over the digit run and
reject the group unless the converted span ends exactly where the
scan did.

Both allocation sites are reachable from HATCH entities in an untrusted
DXF, which is what makes these worth carrying over rather than leaving to
the next full sync.

Also fixes the "assoiciative" comment typo on nAssociativityFlag, so the
file matches trunk.

No functional change beyond the above: after this commit the idxf sources
on AOO41X are content-identical to the #491 head (96b0524be2), modulo the
trailing-whitespace cleanup trunk applied to the licence headers.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

show more ...

Bz126990 file corruption at shutdown trunk (#493)

Test confirms fix: https://lists.apache.org/thread/m2plf5f2rtkmy7rdvnlbt5vjxk263m6d
merging the fix
* ucb/ucpfile: flush file data to the medium bef

Bz126990 file corruption at shutdown trunk (#493)

Test confirms fix: https://lists.apache.org/thread/m2plf5f2rtkmy7rdvnlbt5vjxk263m6d
merging the fix
* ucb/ucpfile: flush file data to the medium before reporting a write successful

Issue 126990; see main/sfx2/source/doc/readme.md for the full analysis.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 0387c05dadfcf647c52ce960f2e148e65ac63c00)

* sfx2: replace the document atomically instead of rewriting it in place

Issue 126990; see main/sfx2/source/doc/readme.md for the full analysis.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 6a1923dc3753b5aa0d063f41324dde9fe8ec0f33)

* update license

---------

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
(cherry picked from commit 210da5364ee5715a5b2c0345cd5580a3bfc88049)

show more ...

Limit paths for text caches for embedded databases

(cherry picked and adapted from commit b81185736b1dc1d0d31d3c6399b1242bd42a93a5)

Check lengths

Also fix the allocation check which tested ppStream rather than *ppStream

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UWgzQY

Check lengths

Also fix the allocation check which tested ppStream rather than *ppStream

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UWgzQY2r1XwFvgeLPFWPsi
(cherry picked from commit a076163db497445461c819d3ae9afe052eb01982)

show more ...

sw: ww8: validate PLCF length

Add an inline WW8PLCF::IsValidLength() predicate and call it at the
entry of both WW8PLCF::ReadPLCF and the WW8PLCFspecial constructor.

Add GoogleTest_sw_ww8plcf cover

sw: ww8: validate PLCF length

Add an inline WW8PLCF::IsValidLength() predicate and call it at the
entry of both WW8PLCF::ReadPLCF and the WW8PLCFspecial constructor.

Add GoogleTest_sw_ww8plcf covering the boundary values.

Thanks to Yukihiro Nakamura (nakamurabeer@gmail.com)

Co-authored-by: Yukihiro Nakamura (nakamurabeer@gmail.com)
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
(cherry picked and adapted from commit 8604ca21f9283f811f592bd95847a6002173ce6e)

show more ...

Reject further invalid URLs

Assisted-By: Claude Opus 4.8 <noreply@anthropic.com>
(cherry picked from commit 181421139242694b309751fb666406eddc203c50)

Restore compatibility with provided files

Certain documents are slightly invalid ZIP's: their entries have a data
descriptor, but their "flags" field say otherwise. Remain compatible
with our past

Restore compatibility with provided files

Certain documents are slightly invalid ZIP's: their entries have a data
descriptor, but their "flags" field say otherwise. Remain compatible
with our past bugs.

Thanks to Czesław Wolański for reporting, and Peter Kovacs for proposing
this fix!

(cherry picked from commit 57cd559d060c509755c8b0f6cac469b512911542)

show more ...

More checks

(cherry picked from commit 66ab9c1188ef75c84225c856a87bc93d575247ca)

VCL: when loading PNG's, ensure we do not reuse an object that was created on the stack

Also, add a guard to ZCodec against unwanted reusal of variables.

(cherry picked from commit 56e4d38c33928bad

VCL: when loading PNG's, ensure we do not reuse an object that was created on the stack

Also, add a guard to ZCodec against unwanted reusal of variables.

(cherry picked from commit 56e4d38c33928bade1ec98e3d1fd9ebea94cde1d)

show more ...

Merge pull request #491 from apache/bugfix-dxf-filter

Bugfix dxf filter

(cherry pick and adaptation of merge commit
c082241a9b512a7cac07a6b4facb9ec65953be75)

Additional check

(cherry picked from commit 040baae65323da00a837571412055ec32f31a97c)

Fix check for enough bytes remaining

Issue found by Damjan Jovanovic. Thank you!

(cherry picked from commit e8b69c8c40ed70ab02f2db79c62a8803ffec612a)