revert to default OS theme for progressbar (fix #1071) #1091
Conversation
|
Dunno why, but the default is now even readable on Win7 as the text stays black, no matter if the Window is active or not ... I'm fine with the OS default then! |
| @@ -425,17 +422,17 @@ void BitcoinGUI::trayIconActivated(QSystemTrayIcon::ActivationReason reason) | |||
| void BitcoinGUI::toggleHidden() | |||
| { | |||
| // activateWindow() (sometimes) helps with keyboard focus on Windows | |||
| if(isHidden()) | |||
| if (isHidden()) | |||
There was a problem hiding this comment.
Why add spaces here? Most if the ifs() are like this, and IMO that's fine.
There was a problem hiding this comment.
There are at least 5 places in where it is if () instead of if(). What
do you prefer? I'll use that in the future, if we don't get a new
standard coding style, so that you and I use the same as we often work
on the same code parts.
There was a problem hiding this comment.
I always use this coding style for if/expressions:
if(a && !b && ((c + d) < e))
{
}
else [if(...)]
{
}
You can recognize other people's code if it is different :-) So I prefer it that way, but won't reject pull requests on it unless it's an absolute mess otherwise.
My point was that here's no need to change this function in this patch. We like minimal changes for small things, as they make merging (of other patches, less rebases) easier.
|
The dominant bitcoin coding style is overwhelmingly "if (", with a space. Two seconds of grep would show that... Let us continue to be consistent with the rest of the code. |
|
Maybe in the core. Not in the Qt code. Whatever the common coding style is, commits shouldn't be cluttered with unrelated changes. |
revert to default OS theme for progressbar (fix #1071)
|
Why was this merged? This is how it looks now for me: http://bitcoin.sipa.be/bitcoinqt.png |
|
Sipa: See discussion on #1071. We intend to only override the style in specific cases. In most cases the OS default simply works. |
revert to default OS theme for progressbar (fix bitcoin#1071)
* fix dstx acceptance * two more fixes: - do not re-verify dstx once accepted - fix dstx prioritization * return early for dstx conditions, more log verbosity
Fuzzer update
As per TODO list and further comments to PR bitcoin#1091 plus further improvements: - Disable the 'fuzz_tester' that breaks with a hang and a crash on purpose when fuzzing all cases at once. - For the 'list_tests' command line argument, also print the test number used for fuzzing everything at once. This is to have a cross reference to interpret fuzzer results obtained this way. - Updates to the fuzzer documentation, add some more info on how to start the fuzzing using AFL - Ignore std::ios_base::failure errors from CDataStream, also when reading fails to prevent some spurious crashes, also update mechanics so that run() is only called when reading the data succeeded. - Fix a potential buffer overrun in fuzz test case of ParseFixedPoint(..) - Allow to get the output from the fuzzed functions or objects (in the latter case, the reserializations of the objects). This is to aid in building a 'fuzz test case suite' that will allow to compare the results of interesting fuzzed test cases between subsequent version of the source code and might help in spotting inadvertent changes to the fuzzed functions when these assumed invariants break. Of course, there is no automatic interpretation of the results. This is enabled by prepending the fuzz test case with a '+'. If only a '+' is given as the sole command line argument, the generic switch{...} fuzz test case will be instructed to produce output. - fix to allow input for specific tests with less than 4 bytes test vector
…scanning 063eb8c Don't store state every block, when syncing or rescanning (dexX7) Pull request description: A major bottle neck during the initial synchronization, rescanning or processing new blocks is due to storing the consensus state to disk. This is done, so the state can easily be rolled back in case of a block reorganization. However, during the inital synchronization or when rescanning and reparsing historical blocks, it is only needed in case of a shutdown or crash, but can be skipped otherwise. Also: store state every 5000 blocks during initial synchronization or rescanning. Tree-SHA512: 48606bcac8549cac82882bcf64465bfa7bdc2a05075a23e89df24d98df0e584250478820918068100f2b69d2e9ee9ac3fb28a745c5309b1b2792d16167fa9d6e
As there were a few readability issues with certain OS themes and our self-defined progressbar style, this reverts the bar to OS defaults.
Fixes #1071.