diff options
Diffstat (limited to 'blog/post/2016-09-24.html')
| -rw-r--r-- | blog/post/2016-09-24.html | 230 |
1 files changed, 230 insertions, 0 deletions
diff --git a/blog/post/2016-09-24.html b/blog/post/2016-09-24.html new file mode 100644 index 0000000..0de4d43 --- /dev/null +++ b/blog/post/2016-09-24.html @@ -0,0 +1,230 @@ +<html><head> +<meta charset="utf-8"> +<meta name="viewport" content="width=device-width,initial-scale=1"> +<meta name="theme-color" content="#000000"> +<title>Chrisoft::Blog</title> +<script type="text/javascript" src="/panel.js"></script> +<script type="text/javascript" src="/blog/footnoter.js"></script> +<script type="text/javascript" src="/blog/aes-js.js"></script> +<script type="text/javascript" src="/blog/scrypt.js"></script> +<script type="text/javascript" src="/blog/sha256.js"></script> +<script type="text/javascript" src="/blog/decryptor.js"></script> +<link rel="stylesheet" type="text/css" href="/common.css"> +<link rel="stylesheet" type="text/css" href="/panel.css"> +<link rel="stylesheet" type="text/css" href="/theme0a.css" id="theme0a"> +<link rel="stylesheet" type="text/css" href="/theme0b.css" id="theme0b"> +<link rel="stylesheet" type="text/css" href="/theme1a.css" id="theme1a"> +<link rel="stylesheet" type="text/css" href="/theme1b.css" id="theme1b"> +<link rel="stylesheet" type="text/css" href="/theme2a.css" id="theme2a"> +<link rel="stylesheet" type="text/css" href="/theme2b.css" id="theme2b"> +<link rel="stylesheet" type="text/css" href="/theme3a.css" id="theme3a"> +<link rel="stylesheet" type="text/css" href="/theme3b.css" id="theme3b"> +<link rel="stylesheet" type="text/css" href="/blog/blogext.css"> +<script> +function ol() +{ + window.onresize=function() + { + if(window.innerWidth<768) + setupevents(); + else unsetevents(); + } + window.onresize(); + _decryptonload(); +} +function loadTheme(){ + var thm=document.cookie.replace(new RegExp("(?:(?:^|.*;\\s*)thm\\s*\\=\\s*([^;]*).*$)|^.*$"),"$1"); + if(thm.length<2||'0123z'.indexOf(thm[0])==-1||'abz'.indexOf(thm[1])==-1)thm='zz'; + var ent=""; + var d=new Date(); + if(thm[0]=='z') + { + var m=d.getMonth()+1; + if(m>=3&&m<6)thm='0'+thm[1]; + else if(m>=6&&m<9)thm='1'+thm[1]; + else if(m>=9&&m<12)thm='2'+thm[1]; + else thm='3'+thm[1]; + } + if(thm[1]=='z') + {if(d.getHours()>=18||d.getHours()<6)thm=thm[0]+'b';else thm=thm[0]+'a';} + ent=`theme${thm}`; + var R=new RegExp('theme[0-4][ab]'); + for(var i=0;i<document.styleSheets.length;++i) + { + if(R.exec(document.styleSheets[i].ownerNode.id)!==null&&document.styleSheets[i].ownerNode.id!=ent) + document.styleSheets[i].disabled=true; + else document.styleSheets[i].disabled=false; + } + var thmcolor=""; + switch(thm[0]) + { + case '0':thmcolor=thm[1]=='a'?'#f59dda':'#2f0933';break; + case '1':thmcolor=thm[1]=='a'?'#9df59d':'#090933';break; + case '2':thmcolor=thm[1]=='a'?'#edb47b':'#1f1205';break; + case '3':thmcolor=thm[1]=='a'?'#a0cdfa':'#051933';break; + } + document.querySelector("meta[name=theme-color]").setAttribute('content',thmcolor); +} +loadTheme(); +</script> +</head> +<body onload="ol()" style="overflow-x:hidden;"> + <div id="panel" class="TText"> + <ul id="panellist"> + <li><a href="/"><h1>Chrisoft</h1></a></li> + <li><a href="/blog"><h2>Blog</h2></a></li> + <li><a href="#"><h3 id="title">2-in-1</h3></a></li> + <li><span>Tags</span> + <ul id="tagslist"> + <li><a href="/blog/list/devel/">devel</a></li><li><a href="/blog/list/qmidiplayer/">qmidiplayer</a></li><li><a href="/blog/list/c++/">c++</a></li><li><a href="/blog/list/qt/">qt</a></li></ul> + </li> + <li id="tocouter"> + <span>Table of Contents</span> + <ul id="tocroot"> + <li><a class="toctarg" href="#tocanch0">UPD 2016-10-28</a></li><li><ul class="tocnode"><li><a class="toctarg" href="#tocanch1">BUT THAT'S NOT THE END OF THE STORY.</a></li></ul></li><li><a class="toctarg" href="#tocanch2">UPD@2016-12-28</a></li></ul> + </li> + <li style="margin-left:-0.5em"><a id="prevp" href="2016-06-21.html">Prev post</a></li> + <li style="margin-left:-0.5em"><a id="nextp" href="2017-01-09.html">Next post</a></li> + </ul> + </div> + <div id="content"> + <h2 id="titleh" class="TText" style="font-wight:normal;">2-in-1</h2> + <div id="datetags" class="TText" style="margin-bottom:1em;">2016-09-24<br>#devel #qmidiplayer #c++ #qt</div> + <hr><div id="article" class="TText"> +<article> +<p> +The visualization plugin for QMidiPlayer in Linux has suffered from a dbus-related bug for a long time: +after you have opened the visualization window, close it and try to bring it up again, it will either +result in several lines of dbus warning which suggests there's a bug in the program (on Debian sid), +or simply a crash (on Arch Linux). +</p> +<p> +Well, I have to admit the design of QMidiPlayer's visualization architecture is pretty weird: +the visualization plugin spawns its own thread and do almost everything in it. After doing some +research on the Internet, I realized that everyone suggests that Qt and SDL <i>shouldn't</i> be +used together, "use QtOpenGL instead". However I thought that will hang the main UI thread and +I just don't believe in the myth. +</p> +<p> +So I started my investigation in early May. I even opened an related issue on github. Unfortunately nobody +offered help. So I had to solve the issue myself. First I had a look into the dbus warnings: +</p><pre style="white-space: pre-wrap;">process <PID>: arguments to dbus_connection_unref() were incorrect, assertion "connection->generation == _dbus_current_generation" failed in file ../../dbus/dbus-connection.c line 2822. +This is normally a bug in some application using the D-Bus library. +</pre> +<p></p> +<p> +Qt definitely uses dbus. After spitting these into stderr, QMidiPlayer crashes instantly in the +QDBusConnection thread. Googling the warning turned out nothing. Then I came across with this blog post +from a KDE developer:<br> +<a href="https://blogs.kde.org/2009/03/26/how-crash-almost-every-qtkde-application-and-how-fix-it-0">https://blogs.kde.org/2009/03/26/how-crash-almost-every-qtkde-application-and-how-fix-it-0</a> +</p> +<p> +Unfortunately, after applying the "fix", the bug didn't change even a little bit. So I had to move on. +</p> +<h3 id="tocanch0" class="tvis">UPD 2016-10-28</h3> +<p> +After several months' idleness of the investigation due to the lack of idea, I suddenly decided to have +a look at all dbus calls when QMidiPlayer crashes. Then I found the following dbus call: +</p><pre style="white-space: pre-wrap;">path=/org/freedesktop/ScreenSaver; interface=org.freedesktop.ScreenSaver; method=Inhibit; +string "My SDL application" +string "Playing a game" +</pre> +<p></p> +<p> +Apparently SDL is also using dbus to do some dirty work -- disabling the screen saver! So I appended +SDL_EnableScreenSaver in SMELT's init function. Unfortunately, nothing happened again. However I've +guessed out the cause to the problem: <b>both Qt and SDL try to initialize a dbus connection in <i>one</i> +process, which is not allowed. So I have to get rid of either one of them.</b> +</p> +<p> +Therefore I began to port the SMELT engine to GLFW. After two days of work, the port was finished and it worked +like a charm -- I have solved the problem! +</p> +<h4 id="tocanch1" class="tvis">BUT THAT'S NOT THE END OF THE STORY.</h4> +<p> +Driven by the unstoppable merriness after fixing a long-existed bug, I started to build GLFW version of +SMELT for Windows. The process was painful but finally I managed to do it successfully. After that +it's time to compile the visualization library. A beautiful DLL file, sized 1096192 bytes, was produced. +This build is almost 4x smaller than the previous failed mingw build and I regarded this as a omen of +fortune. +</p> +<p> +I put the DLL file into the plugin folder of QMidiPlayer and launched the application. No errors occured +as the mingw build once caused and the plugin showed up correctly in the plugin manager. I was so excited +that I turned on the plugin immediately and I couldn't wait to announce the result to my friends. However +when I hit the visualization I was greeted by a déjà vu scene. +</p> +<img src="//filestorage.chrisoft.org/blog/img/qmpcrash.png" alt="" width="70%" decoding="async"> +<p> +And this nasty little dialog from the Windows 2000 era. +</p> +<img src="//filestorage.chrisoft.org/blog/img/couldntbewritten.png" alt="" width="40%" decoding="async"> +<p> +So another tale of a battle between a fierce bug and me is due to happen. <s>Stay tuned and watch the fight!</s> +</p> +<a id="n1" href="#note1" class="note">[1]</a> +<h3 id="tocanch2" class="tvis">UPD@2016-12-28</h3> +<p> +So the bug has been (mostly) fixed by the following two commits!<br> +<a href="https://github.com/chirs241097/QMidiPlayer/commit/b79c4b7e3cab3711e87ba9e28fa8423a84ea7efa">b79c4b7e3cab3711e87ba9e28fa8423a84ea7efa@QMidiPlayer</a><br> +<a href="https://github.com/BearKidsTeam/SMELT/commit/b13f72857af93489f535b84d62882f681dc84a73">b13f72857af93489f535b84d62882f681dc84a73@SMELT</a><br> +</p> +<p> +Basically the most tough part to debug is caused by the stupidest typo... +</p><pre style="padding-left:20px;border:1px white solid;">@@ -62,7 +62,7 @@ bool SMELT_IMPL::smInit() + else if(i==0x80000004) + memcpy(cpuName+32, CPUInfo, sizeof(CPUInfo)); + } + - while(*cpuName=' ')++cpuName; + + while(*cpuName==' ')++cpuName; + smLog("%s:" SLINE ": CPU: %s\n", SYS_SDL_SRCFN,cpuName); + free(loced); +</pre> +<p></p> +<p> +This naïve mistake screwed up the memory and this should be the cause that made the debugger report different stack traces in different build modes... +(that reminds me of the experience of screwing up the call stack, in which the backtrace was filled up by "0x0000000000000000 in ??()"...) +</p> +<p> +After having fixed this, it was much easier to debug the application. Soon afterwards the first working visualization plugin for QMP on Windows +came out. +</p> +<p> +Finally I have to make more complaints about the develop environment for a cross-platform C++ application developer in Microsoft Windows. +It's bloated, complicated and confusing. I wonder whether Microsoft is deliberately making it difficult in order to force more developers +to migrate to C#/.Net. What a big ambition Microsoft is bearing! +</p> +<article> +</article></article></div><br><hr> + <div class="TText" id="notediv" style="font-size:80%;"><span class="TText"><a id="note1" href="#n1">[1]</a>: + Current status of the bug in Windows:<br> + Release build QMidiPlayer+Release build plugin=crash at somewhere in visualization.dll.<br> + Release build QMidiPlayer+Debug build plugin=crash at somewhere in Qt5Core.dll.<br> + Debug build QMidiPlayer+Debug build plugin=crash instantly on launch at plugin initialization.<br> +I have totally no idea about what to do next. +<br></span></div> + <div id="insanch" style="height:3em;"></div> + <div id="footer" style=""> + <div id="pagesw" class="TText" style="width:100%;height:0.5em;"></div> + <div style="text-align:center;" class="TText"> + Proudly powered by SSBS <reduced style="font-size:70%;">(the static stupid blogging system)</reduced> 2.5 + <br> + Content licensed under CC BY-SA 4.0. <span id="purgep" style="display:none;font-size:70%;">This page has passphrase(s) stored. Click <a href="javascript:_purgep()">here</a> to purge.</span> + </div> + </div> + <div id="cmdbuf" class="TText" style="transition:500ms;padding:1em;font-size:2em;color:white;position:absolute;background-color:rgba(0,0,0,0.6);left:50%;top:50%;transform:translate(-50%,-50%);pointer-events:none;opacity:0;"> + </div> + </div> + <div id="decryptui" style="display:none;opacity:0;color:white;z-index:1000;position:fixed;left:0;top:0;width:100%;height:100%;background-color:rgba(0,0,0,0.4);transition:opacity 0.5s;"> + <div id="decryptdlg" class="TText" style="padding:10px 20px;position:absolute;left:50%;top:50%;transform:translate(-50%,-50%);background-color:rgba(0,0,0,0.6);"> + <div id="keyhint" style="margin-bottom:8px;"></div> + <div style="margin-bottom:8px;">Key: <input id="keyinp" type="text" style="color:#fff;"></div> + <div style="height:2.25em;"> + <button id="btndecrypt" onclick="decryptor(decid,document.getElementById('keyinp').value);" style="position:absolute;left:20px;">Decrypt</button> + <button onclick="hidedecryptui();" style="position:absolute;right:20px;">Cancel</button> + </div> + + </div> + + +</div></body></html>
\ No newline at end of file |