Check this commit message https://phabricator.kde.org/R32:f56cdda54b7f88b119f2c7cfb2f534b4fe74478f
- Queries
- All Stories
- Search
- Advanced Search
Advanced Search
May 5 2020
Jan 17 2020
I was debugging the code locally and here's what I found: simple wildcard like * is being replaced with cool stuff like \\A(?:[^/]*)\\z.
As you can see, this pattern excludes forward slashes, which may be a good thing for a file path, but it is exactly the source of the problem in this case.
So here's my debug code, which was tested by me and it works:
Jan 16 2020
In D26205#595428, @ahmadsamir wrote:I still fail to see what's broken, I tested with the system kwallet (5.65 on tumbleweed) and I don't see any difference between it and a build from a git checkout; i.e. falkon doesn't show the saved passwords after restarting it. I've already submitted another diff to fix the doubly-anchored pattern issue.
Jan 15 2020
otherwise, we could just revert this patch
QMap<QString, QByteArray> entries; if (m_wallet->readEntryList("*", entries) != 0) { qWarning() << "KWalletPasswordBackend::initialize Cannot read entries!"; return; }
In D26205#594867, @ahmadsamir wrote:A test case of what is broken would be appreciated, to try and fix/debug the issue.
This changeset breaks the app, making it impossible to get the list of entries from outside. Tested against Falkon browser with KDE integration plugin. See my inline comments here
Jan 3 2020
Is this discussion relevant? Just to have a reference of what is wrong with KMail
Apr 27 2018
In D12115#254882, @adridg wrote:What you could do, is to ask on the distributions@k.o list, to find out who ships what. FreeBSD, for instance, has /usr/local/lib/libavformat.so.57.83.100 so that looks like it would be ok. Debian stable, on the other hand ..
In D12115#254843, @bruns wrote:Before your changes, the code works for quite old versions of libavformat, after your change it requires libavformat > 56 (?), so you are changing the requirements.
In D12115#254831, @bruns wrote:Are you planning to do more cleanups? There is quite some more stuff, e.g. in https://phabricator.kde.org/source/k3b/browse/master/CMakeLists.txt
I think a proper version check for libavcodec/libavformat should be added to the toplevel CMakeLists.txt first, as you are relying on it here.
Apr 26 2018
Apr 13 2018
Apr 11 2018
Apr 2 2018
After my research on libswresample topic I started to think that it's not a good solution for this situation, quite the opposite even. It still requires format conversions which renders the code to a much more complex state. So no advantages here really.
Mar 27 2018
Final revision.
UPD: not really :D
Hello. I decided to rework the patch since some important changes happened:
Mar 25 2018
Check if it's safe to deallocate the buffer and to move the pointer
delete -> delete[]
Mar 24 2018
Rebased to the current master
I'll probably need the commit access to commit these changes
Nov 15 2017
pinentry-kwallet is not a replacement for pinentry-qt or pinentry-gtk binaries (that's what I was thinking). It just kind of works on top of them. Too bad this fact is not documented anywhere.
I realized that there's no need for these changes. Actually they can even break something. Also the project is not a part of KDE Apps https://www.mirbsd.org/kwalletcli.htm
Sorry for the noise
Oct 8 2017
In D8202#153340, @kossebau wrote:None of these are in performance-relevant code, so I don't believe it's worthwhile adding ifdefs; just revert this patch whenever 5.5 support is dropped.
Possible problems: the person bumping min qt version to > 5.5 will not remember about this. And the same for the next person "fixing" this non-const calls.
So while ifdef adds some another code variant which makes test coverage worse, it still help cleaning up the code on version bumps and prevents compat-breaking "fixing". So I would opt for ifdef.