In cases like this xml code:
<string name="app_name">"OpenKeychain"</string>
<!-- title -->
<string name="title_encrypt_text">"Encrypt"</string>
tostring() returns trailing whitespaces (including newlines). Which
aren't removed until the very end, after we try to remove enclosing
quotes. So strip right after tostring() too, since we never really care
about whitespaces anyway.
Up until now, this was done only if no overall version could be found. But in
tags, we would get stuck with an old version from an old tag and not see
subdir changes in new tags.
This finds maven repos of the format:
maven {
url 'http://foo.bar'
}
And checks if the repository is one that we allow. As usual, scanignore can be
used, or the list modified, if there are exceptions or more repositories to
allow.
* Move apks back from the archive repo (e.g. if builds are disabled thus the
number of apks on the main repo decreases)
* Also move apk signatures back and forth
This allows us to fetch strings like the following:
<string name="app_name">foo <xliff:g>bar</xliff:g></string>
Up until now, using .text would only return "foo ", but if we use .tostring()
with the text method, it converts everything into plain text for us resulting
in "foo bar".
8k is too small, e.g. when make -jN errors or when there are a lot of scanner
errors and warnings. 16k should be better. Also, use "[...]" to make it
obvious that the output is truncated.
Looks to me like we were using SHA1withRSA all along. Tested that everything
still works with a test repo.
As Hans reports, SHA1withRSA has been the default on Android tools for a long
time and it's supported on all Android versions.
Add asynchronous filereader, fix python3 lockups
with the current implementation of AsynchronousFileReader with Python 3 FDroidPopen deadlocks with 100% CPU-utilization
The code is from https://github.com/soxofaan/asynchronousfilereader
I hope the MIT-Licence makes no problems
See merge request !80
