| Age | Commit message (Collapse) | Author |
|---|
|
If $HOME isn't canonical (e.g. on FreeBSD it's /home/user while the
canonical path is /usr/home/user), replacing the canonical version of $HOME
with $HOME means that we'll read back a different value than we wrote
in.
It might seem "equivalent" but it leads to surprises like
KRecentDocuments showing duplicates because /usr/home/user became
/home/user in the KConfig roundtrip (but not in the XBEL roundtrip).
This commit loses the replacement of /usr/home/user with $HOME on
FreeBSD, but I think an exact roundtrip is what we expect, rather than
stuff being modified under our feet. The alternative would be to
canonicalize everything in KRecentDocuments but users don't want to
see the /usr in front, I assume (so we would have to use a cache of
canonicalized path, for the removal of duplicates, awful performance
wise).
|
|
This commit is an alternative to the earlier incorrect and reverted
b3dc879e8b108c26c929bfbe551bcdf77f140e94. That commit contained several
mistakes and an algorithmic complexity increase:
1) the added conditions were inverted: should have been
`hasNonDeletedEntries` (without `!`);
2) KConfigPrivate::groupList() passed group instead of key.mGroup to
hasNonDeletedEntries();
3) The complexity of hasNonDeletedEntries() is O(entryMap.size()). Calls
to this function were added into loops that iterated entryMap.size()
times. So the overall complexity of groupList() increased from linear
to quadratic.
This fix collects `mGroup`s of non-deleted key entries instead of
`mGroup`s of group entries. The number of key entries can be much
greater than the number of group entries, so this fix hurts performance.
But at least the algorithmic complexity of groupList() stays linear.
Future commits can optimize the loops and make them almost as fast or
even faster than before this fix.
The `!key.mKey.isNull() && !entryMapIt->bDeleted` checks added in this
commit are consistent with the check in
KConfigPrivate::hasNonDeletedEntries(). KConfig::hasGroupImpl() forwards
its argument to hasNonDeletedEntries() with the following comment:
// No need to look for the actual group entry anymore, or for subgroups:
// a group exists if it contains any non-deleted entry.
BUG: 384039
FIXED-IN: 5.90
|
|
This was built with:
-DQT_MAJOR_VERSION=6 \
-DEXCLUDE_DEPRECATED_BEFORE_AND_AT=5.90.0 \
-DKF_DISABLE_DEPRECATED_BEFORE_AND_AT=0x055a00
Move the include(KDEInstallDirs) call before the first find_package(Qt*,
the former is what auto-detects the Qt version, and defaults to 5. This is
needed to be able to build against Qt5 by default.
All unit tests still pass.
|
|
This will output a warning if an invalid value is requested. The goal is to avoid
implicit conversion which might result in a zero-int value.
|
|
By using an enum we have a central place to provide docs for the most
common actions/restrictions. Also consumers can pass in type-safe enum
values instead of potentially undocumented strings that might contain typos.
Also it is better documents is a value is supposed to be authorized using
KAuthorized::authorize or KAuthorized::authorizeAction, in the case of "shell_access"
this was mixed up from time to time.
Considering that we do not want the parameter for the methods to become too long
I have deliberately decided against using `enum class`. Also we don't have
and usecases for the binary operators in combination with the newly added enums.
Task: https://phabricator.kde.org/T11948
|
|
This will become especially useful when moving state data from the config file
to a dedicated state data file.
Task: https://phabricator.kde.org/T12549
|
|
|
|
clang-tidy checks:
readability-isolate-declaration and readability-braces-around-statements
KF task: https://phabricator.kde.org/T14729
GIT_SILENT
|
|
This is the case where we expected to see Key[$d] in the config file,
and it was somehow broken. When saving, the key was omitted, so when
reloading the kdeglobals key was present again.
Detected when trying to write a unittest for a different patch...
I had to reshuffle the unittest a bit because testThreads calls
testSimple which didn't expect that the "[AAA]" group would actually
be deleted now.
|
|
GIT_SILENT
|
|
No impact on the rest of the test, but I was wondering if it might
break anything, better restore it just in case.
|
|
This reverts commit b3dc879e8b108c26c929bfbe551bcdf77f140e94.
This change breaks plasmashell startup and possible other apps as well.
Reverting so the root cause can be investigated without time pressure.
|
|
calling deleteGroup only deletes all entries but the group does still
exists in listGroups
this is somewhat irritating since calling exists on that group will
return false
with this patch the group does no longer exists in listGroup
|
|
NO_CHANGELOG
|
|
It's a QByteArray not a QString.
GIT_SILENT
|
|
I didn't consider the '#ifdef Q_OS_WIN' code paths (since that compiles
just fine locally).
GIT_SILENT
|
|
- Use a global var for a QString that's used many times
- Break up long-all-cap variable names, it makes it harder to read (and
I've fixed one typo in one of those ALLCAPS)
- Fix some clazy warnings, make global QString objects in unit tests
static (so that the compiler doesn't create symbols for them, it doesn't
matter in a unit test but KF code acts as a reference sometimes that
others copy from, tip from dfaure)
- Add TODO note about changing kconfig_compiler to generate C++ code that
uses multi-arg QString::arg(QString, QString, QString) instead of
QString::arg().arg()
- More const; more QString::at() instead of operator[] where appropriate
- Use a ternary where it makes the code more readable (and uses less lines :))
NO_CHANGELOG
|
|
NO_CHANGELOG
|
|
NO_CHANGELOG
|
|
Fix some clazy warnings, and some other minor code optimisations.
NO_CHANGELOG
|
|
I stored a file's lastModified() into KConfig, to compare with it
again on next start, and the loss of milliseconds made the code think
the timestamps didn't match, even when they did.
|
|
In Qt6 QTextCodec/QTextStream::setCodec is replaced with
QTextStream::setEncoding(QStringConverter::Encoding)
Also UTF-8 is the new default for QTextStream.
|
|
|
|
GIT_SILENT
|
|
Summary:
Fix the following bug, if an entry is set on HOME/.config/kdeglobals and on a specific config file like kcmfonts
When you hit restore defaults button and apply, value will be not wrote on the specific file, but then the value is the one from kdeglobals
This patch ensure we write the key to the specific configuration file on those case with an empty value. KConfig will take default value automatically
Test Plan:
Added a test to ensure flag is working
Tested using some KCM to ensure all is working fine
Reviewers: ervin, dfaure, meven, crossi, hchain
Reviewed By: ervin, dfaure, meven
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D28128
|
|
Summary:
Convert license headers to SPDX statements and add
license files as required by REUSE specification.
Reviewers: cgiboudeaux
Reviewed By: cgiboudeaux
Subscribers: ognarb, cgiboudeaux, kde-frameworks-devel
Tags: #frameworks
Maniphest Tasks: T11550
Differential Revision: https://phabricator.kde.org/D27601
|
|
std::cout/std::cerr
Summary: Port endl to "\n" as it's qt namespaced
Test Plan: autotest ok
Reviewers: dfaure
Reviewed By: dfaure
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D26860
|
|
|
|
Summary:
Allows
* projects linking to KConfigCore/Gui to hide deprecated API up to a
given version or silence deprecation warnings after a given version,
using
* -DKCONFIGCORE_DISABLE_DEPRECATED_BEFORE_AND_AT
* -DKCONFIGCORE_NO_DEPRECATED
* -DKCONFIGCORE_DEPRECATED_WARNINGS_SINCE
* -DKCONFIGCORE_NO_DEPRECATED_WARNINGS
* -DKCONFIGGUI_DISABLE_DEPRECATED_BEFORE_AND_AT
* -DKCONFIGGUI_NO_DEPRECATED
* -DKCONFIGGUI_DEPRECATED_WARNINGS_SINCE
* -DKCONFIGGUI_NO_DEPRECATED_WARNINGS
or
* -DKF_DISABLE_DEPRECATED_BEFORE_AND_AT
* -DKF_NO_DEPRECATED
* -DKF_DEPRECATED_WARNINGS_SINCE
* -DKF_NO_DEPRECATED_WARNINGS
* to build KConfigCore/Gui optionally with deprecated API excluded from
the build, using "EXCLUDE_DEPRECATED_BEFORE_AND_AT" cmake argument.
Test Plan:
Builds with EXCLUDE_DEPRECATED_BEFORE_AND_AT set to 0, 4.0.0, 5.0.0,
5.11.0, 5.24.0, 5.39.0, 5.42.0, CURRENT.
Reviewers: #frameworks, mlaurent
Reviewed By: mlaurent
Subscribers: mlaurent, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D24496
|
|
Summary:
In kconf_update, the ctime usage used to be about metadata change time
(buff.st_ctime, before it got ported to the misnamed created()).
I ported it to birthTime, because I think
date of birth is a more useful way to identify a file than
date of permission change which doesn't really matter to us.
But in practice, I can't help but wonder if mtime alone wouldn't be
enough.
For the QStringLiteral("%%1").arg(i) bit, I tested it in tst_qstring,
the first % is left untouched.
Test Plan: make && ctest
Reviewers: mdawson, arichardson, vkrause
Reviewed By: vkrause
Subscribers: pino, arojas, mlaurent, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D23815
|
|
Summary:
It is very unclear at this point what a valid use case for this feature
would possibly be. The old documentation only mentions $(hostname) as
an example, which can be done with $HOSTNAME instead.
Note that $(...) is still supported in Exec lines of desktop files,
this does not require [$e] anyway (and actually works better without it,
otherwise the $ signs need to be doubled to obey kconfig $e escaping rules...).
Test Plan:
ctest passes; various testcases with $(...) in desktop files,
directory files, and config files, no longer execute commands.
Reviewers: mdawson, aacid, broulik, davidedmundson, kossebau, apol, sitter, security-team
Reviewed By: mdawson, davidedmundson
Subscribers: ZaWertun, rikmills, fvogt, ngraham, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D22979
|
|
This enables KConfigWatcher to detect those changes, too.
Differential Revision: https://phabricator.kde.org/D20039
|
|
Summary: compile without foreach
Test Plan: Unittest Ok as previously
Reviewers: dfaure
Reviewed By: dfaure
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D19326
|
|
Summary:
commit 6a18528 introduced escaping of bytes >= 127 to ensure that
KConfig files are valid UTF8.
The simplistic approach with a cutoff results in many escaped bytes
where it is not required. Especially non-western configuration files
would have many escapes.
This commit fixes that by only escaping bytes that are not valid UTF8.
BUG: 403557
FIXED-IN: 5.56
Test Plan: ninja && ninja test
Reviewers: dfaure, arichardson, apol, #frameworks, thiago
Subscribers: rapiteanu, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D19107
|
|
|
|
|
|
|
|
Summary:
UserBase tells me that KDE configuration files are encoded in UTF-8.
https://userbase.kde.org/KDE_System_Administration/Configuration_Files
In practice some *rc files have bytes outside that encoding. In my home directory I found two files that are not valid UTF-8.
I searched with
find ~/.config/ -name '*rc' -exec file {} +
which gives these exceptions:
akonadiconsolerc: Non-ISO extended-ASCII text, with very long lines
kmail2rc: Non-ISO extended-ASCII text, with very long lines
In kmail2rc, the offending fields are
[AttachmentView]/State
[CollectionFolderView]/HeaderState
Both are QByteArray values saved from QHeaderView::saveState().
In the instance I found, the offending bytes were 0x81 and 0x84.
akonadiconsolerc had way more of these values. All seem related to saving widget state and hence probably QByteArrays.
The written QByteArray values look very strange. The bytes in the non-printable ASCII range are written as \xXX but the values above 127 are written literally.
The encoding is done by KConfigIniBackend::stringToPrintable. It does not currently escape bytes that are larger than or equal to 127.
Reviewers: dfaure, arichardson, apol
Reviewed By: apol
Subscribers: apol, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D17651
|
|
|
|
Summary:
Intention is not to create a registry like system, but to replace
KDElibs4Support::KGlobalSettings and to replace other system settingss
-> some app communication in a more generic way.
writeEntry gains an additional flag Notify which if set, will notify
clients of what has actually changed when we sync.
Rationale to put this into KConfig was so that we could have everything
batched and sychronised to the file sync and to get the fine detailed
exposure of what has actually changed which we don't get with a file
watcher.
Default behaviour remains identical without any broadcast messages.
As it is a new dependency it is purely optional and anything referencing
DBus is not in the public API. Our deployment on platforms without DBus
tend to be standalone applications anyway.
Test Plan: Attached unit test
Reviewers: broulik, dfaure
Reviewed By: broulik, dfaure
Subscribers: dfaure, broulik, zzag, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D13034
|
|
Summary: They were not being split properly.
Test Plan: Tests pass, including the new one.
Reviewers: #frameworks, dfaure
Reviewed By: dfaure
Subscribers: dfaure, anthonyfieroni, kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D14162
|
|
Summary:
We expand the path of kdeglobals to
~/.qttest/kconfigtests/../kdeglobals
which works during the test, but this cleanup is run both on startup and
after removing ~/.qttest/kconfigtests at which point the expansion
fails.
Test Plan: Ran test, looked at qttest dir. It had cleaned up
Reviewers: broulik
Reviewed By: broulik
Subscribers: kde-frameworks-devel
Tags: #frameworks
Differential Revision: https://phabricator.kde.org/D13032
|
|
|
|
Summary:
The CI has $HOME=/home/jenkins/ with a trailing slash, which leads to
FAIL! : KConfigTest::testPath() Compared values are not the same
Actual (sc3.readPathEntry("homepath", QString())): "/home/jenkins//foo"
Expected (HOMEPATH): "/home/jenkins/foo"
QDir::homePath() is too clean for our purposes, use $HOME on Unix.
Test Plan:
kconfigtest now passes with HOME=/home/dfaure/
|
|
QStringLiteral is latin1, not utf8.
|
|
|
|
The test checks that saving a read-only config file fails. But because root can write into read-only files, the test is failing when running by root. Check for uid when running the test and make it XFAIL if we are running as root.
REVIEW: 128916
|
|
Inside desktop files we want to reach also data, cache and config home
to create files inside these directories.
REVIEW: 127462
CHANGELOG: Add support for get QStandardPaths locations inside desktop files.
|
|
Due to https://codereview.qt-project.org/#/c/106473/, Qt 5.6 keeps null
characters in QByteArray -> QString conversions, which breaks this test as
one QByteArray contains nulls. Instead, convert the QByteArray to const
char * first, so QString stops at the first null.
The actual behaviour of KConfig is unchanged, as internally the conversion
always went through a const char *, which avoids creating QStrings with
null characters.
REVIEW: 126001
|
|
strings-made-from-floats/doubles.
Qt 5.5 writes out "123.567001" for 123.567f. No matter, what we want is the
roundtrip float-string-float to work, and it does.
|
