Age | Commit message (Collapse) | Author |
|
After commit cf333a878ceed18d0343520a2c65be69fc433b1f, condition
RequestHandler::cProcessed is set regardless if the request was
processed in RequestHandler::ExecuteCmdLineRequests. This allows
PipeIpcThread::execute to avoid infinite wait for the condition
object that is never signaled. But this makes the latter function
to send PROCESSING_DONE to the calling process. So the secondary
process gets no hint that the request actually failed.
This change adds a boolean field to signal if the request really
was processed. In the case of refused processing (i.e., when
pGlobal.is() is false, or when !pGlobal->AreRequestsEnabled() is
true, like when shutting down), the secondary process would get
empty response to its request in PipeIpcThread::enable, instead
of expected PROCESSING_DONE, and thus the process will eventually
exit with non-0 error code, which would allow to detect this in
scripts.
Change-Id: Id91ea28025e3f3ab60c7049f02ed9fc226d2a654
Reviewed-on: https://gerrit.libreoffice.org/65610
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
When determining if a file is an XML file for siging, we need to
read the manifest file to get an accurate detection. In case when
we were signing in the GUI the manifest file was read when the
storage was set. When we didn't sign over the GUI, the manifest
was never read: the code was only present in the GUI code -
"documentsignaturesdialog.cxx" so the detection was wrong and
isXML returned "true" for current.xml.
With this we move the manifest reading to DigitalSignatureManager,
where the manifest is read when needed.
Change-Id: If45a32af6410bc5f7c5afdb976b182bd69ab7d6b
Reviewed-on: https://gerrit.libreoffice.org/65600
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: I6272ae92b7e948680fb7241c387eb205adbbea01
Reviewed-on: https://gerrit.libreoffice.org/65588
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I6dd52a9116d7d6f692e6f1a080c7169a92b265ea
Reviewed-on: https://gerrit.libreoffice.org/65463
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
To not increase the size of the main org.libreoffice.LibreOffice app further,
the plan was to realize this as an org.libreoffice.LibreOffice.Help extension.
Ideally, this would be a localized extension, so that, by default, only a
relevant subset of the extension would be downloaded and installed. (But see
below.)
There are multiple technical problems, as discussed at <https://github.com/
flathub/org.libreoffice.LibreOffice/issues/35#issuecomment-447295308> "Add
integrated LibreOffice Help offline":
* LO can't pass a file URL with query part to xdg-open, so uses a temporary
wrapper .html file that redirects to the target URL. But for the flatpak case
this wrapper can't be in /tmp (which isn't visible from outside the flatpak
sandbox), but is instead stored in a new temp dir under $XDG_CACHE_HOME (which
is always set for flatpaks and /is/ visible form the outside) that is
removed on LO exit.
* The file URL stored in the temp file must be rewritten from the internal path
(/app/libreoffice/help/...) to the path as seen outside the flatpak sandbox.
While the path for the org.libreoffice.LibreOffice /app is stored in
/.flatpak-info, the external path for the org.libreoffice.LibreOffice.Help
extension is different and not stored there. So use a hack trying to
construct that path from what information is available in /.flatpak-info.
* But the help content consists of locale-specific and shared files, and those
reference each other with relative links. But a localized flatpak extension
cannot contain shared files, it can only contain per-language sub-dirs. And
if the shared help files were kept out of the extension, as part of the app
itself, the relative paths among these files inside the flatpak sandbox would
differ from those outside of it, so would be broken when viewing the content
in the external browser. A solution would either (a) need to somehow rewrite
the content of all the help files being served from LO to the external
browser, or (b) replicate the shared help files in all the extension's per-
language sub-dirs (and if some localization uses en-US content as a fallback
for only part of its translated content, e.g., in the case of media files,
would need to also replicate that en-US content), or (c) use a non-localized
extension that always contains the content for all localizations.
For now, I chose the third approach. This makes the
org.libreoffice.LibreOffice.Help extension relatively large (the current
/app/libreoffice/help tree has a size of ca. 100MB), so that I decided to not
have it downloaded and installed automatically ("no-autodownload": true in
solenv/flatpak-manifest.in). (I checked with Flatpak 1.0.6 that if the
extension should be changed to a localized one with the same name in the future,
updating from an older version would work. If the old extension was not
installed, just the relevant localizations of the new version will be downloaded
and installed. If the old extension was installed, the full set of
localizations of the new version will be downloaded and installed.)
(As also mentioned at <https://github.com/flathub/org.libreoffice.LibreOffice/
issues/35#issuecomment-447295308>: "A second, minor, nuisance is that, for
security reasons, an `xdg-open file:///...html` call from a flatpak leads to an
intermediate popup dialog letting the user chose which application to use to
open the URI, while e.g. an `xdg-open http:///...html` is allowed to go directly
to the user's preferred browser. So accessing help content from LO flatpak
would present that popup dialog first, forcing the user to select a browser to
proceed.")
Change-Id: I35f5a23947dd551dc1b9bff1dd2abd6710073b5f
Reviewed-on: https://gerrit.libreoffice.org/65451
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: Ic305e5305890fd1efa3a3130e5216f9c672870e5
Reviewed-on: https://gerrit.libreoffice.org/65175
Tested-by: Jenkins
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
|
|
Change-Id: Iaa962fe5a590ef16e710fdd49d02d564f10f0f9f
Reviewed-on: https://gerrit.libreoffice.org/65188
Tested-by: Jenkins
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
|
|
Separate generic code from Impress / Draw specific code
and implement shape selection handling in case of Writer.
This is an internal filter, so it can be called only from
the code.
Change-Id: I807e04a0949530d6029037bb964c10c80197ff33
Reviewed-on: https://gerrit.libreoffice.org/65174
Tested-by: Jenkins
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
|
|
...by reverting part of 6f50961e69406a17d6ec998956a6b33208b1001b "remove more
rtl::OUString and OString prefixes"
Change-Id: I50d37f437bb6f760c0d3990faaf71084913520f0
Reviewed-on: https://gerrit.libreoffice.org/65169
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: Idf2ebb552065fa600a4010e83a24a0f13651c029
|
|
Will handle it another way instead that isn't so visible and annoying.
This reverts commit 5f0dda8e9f2e35e1fdb5750089a70543478fa8a1.
This reverts commit 4f1a341bc312673b8d92e4474cc39bc3ab69aa5b.
Change-Id: If456d2792a5f2793ee6337898aaa4a55637f8522
|
|
Change-Id: I4021ba6090bd3f42b0eb5ea1ec9bbf05d1594b35
Reviewed-on: https://gerrit.libreoffice.org/65054
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I3efea7e47ad18a1dd39e81929aac319fe46357f0
Reviewed-on: https://gerrit.libreoffice.org/65032
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: If49c33e271426ff5c2d0c6cc4010670f797bdd38
Reviewed-on: https://gerrit.libreoffice.org/65001
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ib97d0da4c52215009dd2142908168c68ea0aebcb
Reviewed-on: https://gerrit.libreoffice.org/64939
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
Change-Id: I9c720c109893fac671006ad7b3b2c58b397cc744
|
|
Combines the LanguageTag::Create() and Translate::get() (or nget())
functionality.
Also add an ABI test for the Kit class in the LibreOfficeKit unit
test, and a test for the translation function.
Change-Id: I78c48a8bbb434d6f204869290822830288022a53
Reviewed-on: https://gerrit.libreoffice.org/64882
Tested-by: Jenkins
Reviewed-by: Tor Lillqvist <tml@collabora.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: Ie87d27dd2c385a63349e0b322fd067ba03d2d152
Reviewed-on: https://gerrit.libreoffice.org/64479
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
which benefits LOOL since we can delay creating the image until
we know the dpi setting of the display we are going to write to.
Achieved by
perl -pi -w -e "s/\bImage\s*\(\s*BitmapEx\s*\((\w+)\s*\)\s*\)/Image\(\1\)/g" $( git grep -lw "BitmapEx" )
followed by
git grep -nP '\bImage\s*\(\s*BitmapEx\s*\('
followed by commenting out the BitmapEx(OUString) constructor and seeing
what needed adjusting.
Change-Id: I3224e11937d720fa484b0d659d25673a9e809267
Reviewed-on: https://gerrit.libreoffice.org/64760
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
|
|
...after 7ffdd830d5fb52f2ca25aa80277d22ea6d89970b
"HAVE_CPP_ATTRIBUTE_FALLTHROUGH is always true now"
Change-Id: I54e5ff4e036a6bb3e5774d1c0524158aae18e937
Reviewed-on: https://gerrit.libreoffice.org/64800
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Unlike C++, C doesn't allow reference-to-pointer types,
and we do have C code that wouldn't compile with ref-to-ptr.
Had to change to ptr-to-ptr, which is the proper way of
having output arrays.
For the same reason, we cannot use new/delete, rather we
must use malloc/free.
Another (lesser) issue was that we used the renderShapeSelection
API to echo back an array we give it as prefix. This made
the API unecessarily complex (in undocumented ways) and
forced the implementation to both worry about user-data
and managing the input memory. This logic is best moved
to the client and the API simply returns the output data.
Speaking of returning data, the API now returns the size
of the array it allocated and wrote to, so the client
can do a simple check on the return value directly.
Change-Id: Ida216c10d5b37efd1e0861e26b72cabb25c568e6
|
|
Change-Id: Ic1b30a89ae513b3b3ec71853830ef7600801dee0
|
|
It works for Impress only now.
Change-Id: I95e3e37ae7df49b567108f6d6467038b715e886d
|
|
Change-Id: I9fb8366634b31230b732dd38a98f800075529714
Reviewed-on: https://gerrit.libreoffice.org/64510
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: If3b121e44ceec8be668820a7e8a9e35c15d547ff
Reviewed-on: https://gerrit.libreoffice.org/64295
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
When triggering rechecking of signature status, an assert fails and
brings down the program because we don't hold a SolarMutex. The
rechecking of signature status needs to broadcasts.
Change-Id: Ib500b06d2e191465df1e0da2097ddb1b6b06725a
Reviewed-on: https://gerrit.libreoffice.org/64292
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: I1ae105067c109a32f4acc34749b425d9e4e26514
Reviewed-on: https://gerrit.libreoffice.org/64291
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: I651b7f202fa52ff5f5357a11aa72c43eb7dc7f95
Reviewed-on: https://gerrit.libreoffice.org/64102
Tested-by: Jenkins
Reviewed-by: Bartosz Kosiorek <gang65@poczta.onet.pl>
|
|
This reverts commit 7eaaf96e17c14ac49f98da9722a06c0c95f0c5c0.
Change-Id: I559b8817db6f04b44d225346914ef385f7e5048c
Reviewed-on: https://gerrit.libreoffice.org/64251
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
Tested-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: If1b2e04872eb0dd6725802c1709a9085f4cd8c91
Reviewed-on: https://gerrit.libreoffice.org/64141
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
and tweak the methods in check.hxx to make them more flexible when
called with
dc.Class(xxx ? "foo" : "bar")
Change-Id: I881fe628f22121ced4d8849715d6b1c92b092da1
Reviewed-on: https://gerrit.libreoffice.org/64207
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ib22c4f628042458cab937c27224505489c67da2e
Reviewed-on: https://gerrit.libreoffice.org/64024
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
Tested-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
found with
git grep -n -A4 'if.*!.*empty' | grep -B3 -P
'(\bfor)|(\bwhile)|(\bdo)'
Change-Id: I582235b7cf977a0f9fb4099eb306fdb4a07b5334
Reviewed-on: https://gerrit.libreoffice.org/64169
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I185527fcc4481ab2bd97ad81849ee38c6fa50919
Reviewed-on: https://gerrit.libreoffice.org/64167
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
while the extension manager is querying via a dialog. Reuse
the isBusy for this.
Change-Id: I97572404ae296e87fd44711bf1e978bd5bad6280
Reviewed-on: https://gerrit.libreoffice.org/64155
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: Ice2eb8c5994bf2ccb88972332ca4a1d3ed41752a
Reviewed-on: https://gerrit.libreoffice.org/63826
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Change-Id: I1ac4c52ea51503373644bc58dcd4395c69f1a675
Reviewed-on: https://gerrit.libreoffice.org/64007
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Fixes the rendering of spreadsheets in the iOS app. (The cell area was
rendered at half the scale of the row and column headers.)
Actual code change only in desktop/source/lib/init.cxx, but update
related comments elsewhere to mention CoreGraphics, too, and not just
cairo.
Change-Id: Ife99c6a2d58e592cfea3b4ed1ab09c19fba77e72
|
|
and put back original SvTreeListBox a11y factory use
Change-Id: I4ad8ce29d8fed6ec5d44e9a1d641919a89226b79
Reviewed-on: https://gerrit.libreoffice.org/63501
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
check for calls to constructors, and extend the list of types we check
for unnecessary temporary creation
Change-Id: Ia2c1f202b41ed6866779fff5343c821128033eec
Reviewed-on: https://gerrit.libreoffice.org/63472
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: Iabe571aa8f00492902c499094bea8365a3e17fca
Reviewed-on: https://gerrit.libreoffice.org/63623
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Being a GUI application on Windows (with related flag in the executable header
- see https://blogs.msdn.microsoft.com/oldnewthing/20090101-00/?p=19643/), OS
would detect the subsystem before launching the application, and won't attach
the parent console or redirected output handles from it to the application.
Also, different hacks to reattach the GUI application to the console later are
unreliable on different Windows versions, and work improperly (the output goes
to the console after the launch command has already returned, which is wrong
in batch files). This makes it extremily difficult to do CLI operations with
LibreOffice on Windows, with error codes/warnings/messages/output missing or
going to wrong consoles.
Making an executable for CUI subsystem, on the other hand, makes Windows to
allocate a console before starting it when the program is run by itself. This
makes the console window to appear on screen unconditionally, even if it's
hidden later when the program has started. This flashing is undesirable.
But we use a wrapper executable on Windows, called soffice.exe, which is what
actually launched by user, and which runs soffice.bin. This allows us to make
soffice.bin the proper console application, and thus make it capable to behave
properly in CLI scenarios, while avoid the console flashing when run from the
soffice.exe (which would suppress the console creation using DETACHED_PROCESS
creation flag to CreateProcessW).
Also creating a new wrapper for console (soffice.com) allows to use command
lines which omit explicit executable extension (no ".bin"), like this:
"C:\Program Files\LibreOffice\program\soffice" --help
which allows to continue using multiple available help resources unchanged,
since .com extension is tried prior to .exe by Windows' cmd.exe.
Change-Id: I089d0f30f860da6cfc781b4383f6598a08a4d238
Reviewed-on: https://gerrit.libreoffice.org/63572
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I5b7f503e32c10b5a844817c040ab0dde27bf48f7
Reviewed-on: https://gerrit.libreoffice.org/63666
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Otherwise they would be passed to the other running instance, and
processed there, without outputting the data to the callng console
Change-Id: I5cd9d11b567dfed06e19e9a2c4c29bc82868c064
Reviewed-on: https://gerrit.libreoffice.org/63577
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
a) use GtkTreeStores for GtkTreeViews
b) ironically can't store GtkTreeStore contents in .ui apparently
c) set show_expanders for all non-trees and unconverted cases
d) on-demand subtrees
Change-Id: I3c1036a222daba2c129b1a22ffeb3fe35005ae31
Reviewed-on: https://gerrit.libreoffice.org/63336
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: If18c80fc64e55d797953e24e40e5d5e62bd9c625
Reviewed-on: https://gerrit.libreoffice.org/63453
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Change-Id: I0bf922c5f0135d32f66a756c9e6af0bb22575a75
Reviewed-on: https://gerrit.libreoffice.org/63532
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
u8 literals incompatibly change their type (as implemented by recent Clang
trunk)
Change-Id: Ia4f7b91f5d86656a056303d2754981ab2093a739
Reviewed-on: https://gerrit.libreoffice.org/63494
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: I25660634dcb9ebb31292275e31ed7a047e4ddd25
Reviewed-on: https://gerrit.libreoffice.org/63474
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
Change-Id: I0aecc459cf448c03dd8c17a486cc66f541f6868d
Reviewed-on: https://gerrit.libreoffice.org/63322
Tested-by: Jenkins
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
|