tag:blogger.com,1999:blog-193518782024-03-14T18:42:47.244+08:00Epico's Home PageHome of a Computer Science Student.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.comBlogger28125tag:blogger.com,1999:blog-19351878.post-43145946794011521462017-12-13T10:15:00.000+08:002017-12-13T10:23:32.570+08:00i18n FAD Pune 2017Last month we had our first i18n Fedora Activity Day (FAD) in Pune, from 20th to 22th November. In Red Hat, Pune office, we hack together for 3 days. It is a really great experience that we collaborate together again.<br />
<h2>
Day 1</h2>
First we have self introductions and the plans for this FAD.<br />
<br />
The first presentation is given by Tagoh-san for Fonts and Rendering in Fedora. Following his topic, I shared my debugging experience about pango, and summarized the status of default Chinese fonts of Adobe Source Han fonts asked by Jens Petersen.<br />
<br />
Later, Takao Fujiwara explained IBus 1.6 plans and emoji demo. We discussed about some items of IBus plans. Then I suggested to collect API changes, and discuss more about changed API. And I plan to implement ForwardKeyEvent method for Qt 5 in this FAD.<br />
<br />
Mike Fabian gave demos about ibus-typing-booster and emoji-picker. Then Rafal presented his work on locales. Sundeep and Parag showed transtats and transtats-cli demo.<br />
<h2>
Day 2</h2>
Suggested by Mike FABIAN, I reported `ls -lh` date/time issues to coreutils upstream. Now I got some feedback about it. It is a known issue but seems hard to fix it. The date/time column in English column contains 12 characters, if we add one character for CJK. The column will become too long; if we remove some space characters from date/time string, it will cause some problem in scripts such as cut, sed, awk, etc.<br />
<br />
I spend the most time of the day with Takao Fujiwara to implement ForwardKeyEvent method. In the beginning I rebuilt the qt5-qtbase source rpm, but it takes too long time to finish building. Helped by Takao Fujiwara, I start to copy the binary library to the system, it makes debugging faster. Later we discussed the details about the patch, peer review helps to write the patch quickly.<br />
<h2>
Day 3</h2>
Today we have Fedora 27 release party and nice cake.<br />
<br />
The Qt patch seems to work now, we begin to check whether ibus-hangul and ibus-typing-booster work with the ForwardKeyEvent method. The results seems work well, and we submited the patch for upstream review.<br />
<br />
I had really nice experience during visiting Pune. Thanks Parag and Sundeep for organizing this event! Thanks to Fedora Council for approving this FAD budget and Bex for helping us with all kinds of budget issues!Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com2tag:blogger.com,1999:blog-19351878.post-70672650171538482502015-11-23T14:33:00.000+08:002015-11-25T11:17:31.391+08:00G11N FAD ReportFrom Nov 1 to Nov 3, we joined G11N FAD 2015.<br />
URL: https://fedoraproject.org/wiki/FAD_G11N_2015<br />
<br />
During 3 days, we have various activities, such as discussions and hackathons, etc.<br />
<br />
For the first day, Jens Petersen gives a warm welcome and introduction.<br />
And after some other introductions, we know about other G11N teams better.<br />
<br />
In the afternoon of day one, with Takao Fujiwara and Daiki Ueno we <br />
discussed ibus future development, such as the differences between ibus and fcitx.<br />
Ueno-san told me about the gnome input method support status for wayland.<br />
And we discussed whether can improve ibus wayland support.<br />
<br />
For the second day, Mike Fabian told me there are two ways to package glibc locales:<br />
1. with locale archive;<br />
2. with locales in many folders;<br />
<br />
During second day, we use the builtin systemd bootchart module and Michael Meeks' bootchart projects to compare the boot time with two ways to package glibc locales.<br />
This time we mainly tested a few locales installed.<br />
<br />
For the third day, we continued with glibc sub-packaging benchmarks.<br />
This time we tested with full locales installed.<br />
The results seems a bit random, but the performance seems not different much.<br />
<br />
Finally we have good outcomes from this FAD, and have a nice photo. :)<br />
<br />
<a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5t2HZWe8BxGA81UFGnlCKlS63kSWL5aweTJ7-z373o4QZL1lrQKQNiyUktB1VNAepXEnZdx-5qS2kDBFx50VdALp19uUuuuWNsCnmwd_C6b9VnY7DvzLxUAD4w6IrJAVipg4Omw/s1600/post.JPG" imageanchor="1"><img border="0" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5t2HZWe8BxGA81UFGnlCKlS63kSWL5aweTJ7-z373o4QZL1lrQKQNiyUktB1VNAepXEnZdx-5qS2kDBFx50VdALp19uUuuuWNsCnmwd_C6b9VnY7DvzLxUAD4w6IrJAVipg4Omw/s320/post.JPG" /></a>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-21857090732166702252011-09-19T18:22:00.002+08:002011-09-19T18:24:04.090+08:00How to highlight #if 0 in emacsIn vim, the editor can highlight the #if 0 ... #endif block, which emacs lacks. I did some digs in the web, and find a solution, with some slight modifications.
<br/>
Here are the <a href="http://machine-cycle.blogspot.com/2011/02/highlight-if-0-else-endif-in-emacs.html">Original Post</a>.
<br/>
Just add the below code snippet to your ~/.emacs:
<br/>
<pre>
(defun my-cpp-highlight ()
"highlight c/c++ #if 0 #endif macros"
;; (interactive)
(setq cpp-known-face 'default)
(setq cpp-unknown-face 'default)
(setq cpp-known-writable 't)
(setq cpp-unknown-writable 't)
(setq cpp-edit-list '(("0" font-lock-comment-face default both)
("1" default font-lock-comment-face both)))
(cpp-highlight-buffer t))
(add-hook 'c-mode-common-hook 'my-cpp-highlight)
</pre>
<br/>
And your emacs will highlight the #if 0 ... #endif block for you.:)
<br/>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-88027191063559260252010-10-18T10:55:00.012+08:002010-10-18T15:39:18.359+08:00SunPinyin、Novel Pinyin和iBus-Pinyin展开全面合作SunPinyin、Novel Pinyin和iBus-Pinyin社区决定长期合作,联手创建libpinyin子项目。libpinyin致力于为中文拼音输入法提供智能整句输入的算法核心,将试图合并novel-pinyin和sunpinyin中的智能拼音整句输入算法部分,争取实现目前SunPinyin和Novel Pinyin的功能超集。<br /><br />在不久的将来,libpinyin将为ibus-pinyin和sunpinyin提供智能拼音整句输入功能的支持。<br /><br />Project Wiki: <a href="http://github.com/libpinyin/libpinyin/wiki">http://github.com/libpinyin/libpinyin/wiki</a><br />Project SCM: <a href="http://github.com/libpinyin/libpinyin">http://github.com/libpinyin/libpinyin</a><br />Mailing List: <a href="http://groups.google.com/group/libpinyin">http://groups.google.com/group/libpinyin</a><br /><br />另见SunPinyin的官方声明:<a href="http://yongsun.me/2010/10/">http://yongsun.me</a>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com1tag:blogger.com,1999:blog-19351878.post-20684010722032566492010-06-29T13:24:00.001+08:002010-06-29T13:25:15.777+08:00novel-pinyin 0.2.5 released.Here are the new change log for novel-pinyin:<br />* Fixes train factor.<br />* Fixes file open/close bug in utils.<br />* Fixes a bug in winner tree,<br />which may potentially lower the correct rate.<br /><br />And this package is uploaded to sourceforge today.<br />The source control system for novel-pinyin has been switched from CVS to SVN. If you want to check out the newest code, please use the new SVN repository.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1935181067335615942009-11-20T16:49:00.005+08:002010-09-26T14:16:20.998+08:00博客网址镜像由于blogger.com被GFW防火墙屏蔽掉了,为了方便大家访问,novel-pinyin的blog已经被镜像到了sourceforge的WordPress上。<br />同时新申请了一个mailing list (novel-pinyin-devel@lists.sourceforge.net)专为 novel-pinyin 开发和讨论使用。<br />本人博客的最新内容请见<a href="http://alex-epico.blogspot.com/">http://alex-epico.blogspot.com/</a>。<br />由此引起的不便,敬请谅解,谢谢。Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-65139069304660798672009-07-06T16:02:00.016+08:002009-07-06T18:37:40.050+08:00How to prepare a Mozilla Building and Debuging Environment.The following information is about Firefox 3.0 and mozilla-xulrunner 1.9.0 on SUSE Linux. Please following the steps below:<br /><ol><br /><li>Grab MozillaFirefox 3.0.x and mozilla-xulrunner 1.9.0.x packages, and extract their contents to directory MozillaFirefox and mozilla-xulrunner190.<br /><li>Use quilt to apply patches to mozilla-xulrunner190:<br /><ol><br /> <li>use quilt -v setup mozilla-xulrunner190.spec to setup the xulrunner 1.9.0 source tree.<br /> <li>change directory to mozilla-xulrunner190/mozilla, using quilt push -a, to apply all xulrunner 1.9.0 patches.<br /> <li>use quilt import to import all patches in MozillaFirefox packages, as we put Firefox patches in two seperate packages containing different patches.<br /></ol><br /><li>Put a mozconfig in directory mozilla-xulrunner190/mozilla, contains the following informations:<br /> <pre><br /> . $topsrcdir/browser/config/mozconfig<br /> mk_add_options MOZ_OBJDIR=@TOPSRCDIR@/objdir-ff-debug<br /> ac_add_options --enable-debug<br /> ac_add_options --disable-optimize<br /></pre><br /><li>Type the command "make -f client.mk build" in directory mozilla-xulrunner190/mozilla.<br /><li>Wait for compiling finished, go to directory objdir-ff-debug/dist/bin, using "./firefox" to run the built Firefox 3.0.<br /></ol>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1965645343777250802009-07-06T15:28:00.009+08:002010-01-20T18:15:12.370+08:00In SUSE Mozilla Team now.Since February 2009, I was transfered from i18n team to Mozilla team in Novell.<br />Although it is a tough task, but it is also interesting to me.<br /><br />Recent months I am very busy, thanks for the help from <a href="http://www.rosenauer.org/blog/">Wolfgang</a>. Later maybe I can spend more time on making patches. :)<br /><br />As I spend lots of time about how to debug Firefox bugs on SUSE Linux. I will write some series of articles to ease others to join developing Mozilla Firefox on SUSE Linux.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-75623202966073558462009-03-08T21:25:00.003+08:002009-03-08T21:55:50.450+08:00Make Intel 82845G/GL/GE card 3D Acceleration work on openSUSE 11.1.This weekend I decided to update my home PC from openSUSE 10.2 to openSUSE 11.1.<br />After finishing the update, when X server starts, the OS stuck.<br />Re-configure my card using sax2, got vesa driver, without any 3D acceleration.<br /><br />Using my card vendor id and device id searching on google, found a bug on launchpad, and it recommended to add this to xorg.conf in device section:<br /> Option "NoAccel" "true" # works around LPB 304871 - probably disables the next two.<br /><br />And the driver prints "failed to init memory manager", using git clone to get intel driver code, find the message in code:<br /> if (!xf86InitFBManager(pScreen, &(pI810->FbMemBox))) {<br /> xf86DrvMsg(pScrn->scrnIndex, X_ERROR,<br /> "Failed to init memory manager\n");<br /> return FALSE;<br /> }<br /><br />And google xf86InitFBManager, found it was related to XAA. Reading man intel, find:<br /> Option "AccelMethod" "string"<br /> Choose acceleration architecture, either "XAA" or "EXA".<br />So I tried to remove "NoAccel", and add the following line to xorg.conf:<br /> Option "AccelMethod" "XAA".<br /><br />Last more than 20 changes to xorg.conf, failed to start Xserver, but for this time, the Xserver starts normally, and running glxgears, the performance is the same as before. <br />Oh, I got the 3D acceleration back again.<br /><br />After the Xserver is OK, I checked the src/i810_driver.c file again. I found that it doesn't include exa.h. So it seems not support EXA, but by default, the EXA is chosen instead of XAA, so this is the cause of the problem.<br /><br />PS: my card id:<br />Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device [8086:2562] (rev 01)Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-24788139960974096462009-03-08T20:53:00.007+08:002009-03-08T22:11:53.587+08:00Switch Editor from Vim to Emacs.Recently I begin to maintain a new project in SUSE Linux, glad to return the C++ world. The new project is tough, but more challenging, so the last month is very busy.<br /><br /> On Feb 10, I am trying to switch editor to emacs. In the past, I use emacs for my own project, and use vim to make patches for SUSE Linux. But I still feel uncomfortable on mode editor like vim, already get used to mode-less editor. But I didn't know how to use emacs to browse function definitions in large projects before, finally I settled this.<br /><br /> First, turn cua-mode on, by which you can use Ctrl-X, Ctrl-C, Ctrl-V, for cut, copy, paste, and Ctrl-Z for undo.<br /><br /> Then use ctags or etags to generate TAGS files.<br /> And enable emacs to browse multiple function definitions in C++ by using etags-select.el.<br /> Here are the URL:<br /> <a href="http://www.emacswiki.org/emacs/EtagsSelect">http://www.emacswiki.org/emacs/EtagsSelect</a>.<br /><br /> Now browsing C++ code is easy, but the code completion still does not work. It doesn't matter I don't rely on code completion much now.<br /><br />PS: command examples for generating TAGS file.<br />$ctags -eR .<br />or<br />$find \( -name '*.h' -or -name '*.cpp' \) -print | etags -Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-60032964365787734392008-08-25T16:13:00.008+08:002008-08-25T18:26:23.471+08:00Sorry, for the bugs of novel-pinyin 0.2.3 release.As the first release of 0.2.x series, I put the novel-pinyin on sourceforge. But later I withdraw the package, because some serious bug has been found.<br />During the Beijing Olympics, I finally released the novel-pinyin 0.2.3 package.<br />Thank <a href="http://lymanrb.blogspot.com/">lyman</a> for feedbacking the bug in initializing code.<br />As novel-pinyin has been released, the fix code is relatively small, so I decide to release the fix as a seperate patch.<br />using the following command in novel-pinyin-0.2.3 directory:<br />patch -p2 < ../../urgent-patch-fix-novel-pinyin-first-load.patch<br /><br />PS:<br />顺便提一句,输入法的中文名称变为了新智能拼音,英文名称为Novel Pinyin不变。Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-3130246532930355022008-08-08T14:08:00.011+08:002008-08-13T12:34:58.250+08:00novel-pinyin 0.3.x wishlistTODO Items:<br />1.Modify pinyin large table to merge scim-pinyin phrase lib into gb_char.table.<br />2. Write phrase to token conversion. (phrase_large_table)<br />3. Write n-gram segment to bootstrap phrase generation. (replace current mmseg.)<br />4. Larger corpus learning.<br />5. Entropy-based n-gram prune.<br />6. Add professional phrase libraries support.<br />7. Better fuzzy pinyin support.(like ms-pinyin)Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-3508698938651263582008-08-08T13:51:00.010+08:002008-08-13T12:34:45.815+08:00novel-pinyin 0.2.3 releasedDone Items:<br />1. Import the entire scim-pinyin phrases as corpus.<br />2. Better HMM parameter adjusts.<br />3. Better candidates adjusts.<br />4. Add version check.<br />5. Add data file corruption detection.<br />6. Protect against integer overflow.<br /><br />Todo Items:<br />A input pad module for temporarily input Chinese characters by strokes lookup.<br />(Maybe this can be done in Hacker Week.)Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-4055301907975960102008-05-14T10:33:00.012+08:002008-05-14T11:03:56.578+08:00novel-pinyin 0.2.x wishlistAs the first version of novel-pinyin has been released, some feedback has been received.<br />The next version of novel-pinyin will try to finish the following todo tasks:<br />1. Model Modification. Modify the P(P|W) from k/n to C(P,W)/C(W).<br /> (C(P,W) stands for counter of pinyin and word combination,<br /> C(W) stands for word counter.)<br /><br />2. Dynamic adjust phrase positions according to bi-gram possibilities.<br /> As in HMM model training process, the frequency adjusted is very small(1 or 6).<br /> To magnify the position changes, replace unigram with bi-gram when possible.<br /><br />3. Versioned Data File Format.<br /> As data file format will be changed in next release. So I will add a version file in<br /> ~/.scim/novel-pinyin, to indicate file format version.<br /> When different version has been detected, the files of old version will be flushed.<br /><br />Optional:<br /> skim integration.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com2tag:blogger.com,1999:blog-19351878.post-18885176706830624112008-02-19T12:15:00.002+08:002008-02-19T12:17:21.081+08:00novel-pinyin 0.1.0 internal testYou can get newest novel-pinyin 0.1.0 from the following url:<br />http://download.opensuse.org/repositories/home:/wupeng/<br /><br />The source code in sourceforge.net misses the data file, so it will not run.<br />Please use the rpm on the above url.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-41108786079994293492008-02-14T14:08:00.004+08:002008-02-19T12:17:50.624+08:002008 New Year!我自己的输入法Novel Pinyin终于跑起来了,还有一些bug,不过影响不大。现在我就在用我自己写的输入法,写自己的博客。<br />下周开始在同事中测试新的输入法。<br />首先,在这个周末,要把rpm在openSUSE Build Service上做出来。Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-46007486732260669962007-12-20T10:53:00.000+08:002007-12-20T11:10:03.211+08:00Finish segment & training part, upload it to sourceforge.Upload new novel-pinyin code to sourceforge, currently finished segment & training part.<br /><br />In this place, I use a modified interpolation method to ease implementation.<br />The parameter optimization part is done in research prototype.<br />So the code in novel-pinyin is relatively simple, just use parameters computed from prototype.<br /><br />The word segment use shortest path algorithm to segment words, and prepare the data to training part.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1089124265107632852007-12-13T17:02:00.000+08:002007-12-13T17:14:51.510+08:00AIGLX on OpenSUSE 10.3My Notebook has a 945 GM Graphics Card, which support AIGLX.<br />Default 3D Desktop on openSUSE will use XGL. But I want other OpenGL applications can benefit from<br />hardware acceleration. So I switched to AIGLX.<br />Refer to http://en.opensuse.org/AIGLX.<br /><br />And I use X.org config file from <a href="http://www.linuxquestions.org/questions/linux-software-2/compiz-fusion-on-opensuse-10.3-aiglx-intel-592150/">here:</a><br /><pre class="bbcodeblock" dir="ltr" style="border: 1px inset ; margin: 0px; padding: 3px; overflow: auto; width: 98%; height: 322px; text-align: left;">Section "Device"<br /> Identifier "** Intel i810 (generic) [i810]"<br /> Driver "i810"<br /> VideoRam 262144<br /> Option "DRI" "true"<br /> Option "XAANoOffscreenPixmaps" "true"<br />EndSection<br /><br />Section "ServerLayout"<br /> Option "AIGLX" "true"<br />EndSection<br /><br />Section "DRI"<br /> Mode 0666<br />EndSection<br /><br />Section "Extensions"<br /> Option "Composite" "Enable"<br />EndSection</pre><br />For Java Application, set AWT_TOOLKIT=MToolkit, from <a href="http://linuxmint.com/forum/viewtopic.php?p=7043">here</a>:<br />export AWT_TOOLKIT=MToolkit to avoid gray window.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-88105305944143827292007-11-30T11:18:00.000+08:002007-11-30T14:04:41.760+08:00Upload new novel-pinyin source code to sourceforge.netFinished n-gram storage code, and update code on http://novel-pinyin.cvs.sourceforge.net/novel-pinyin/novel-pinyin/.<br /><br />On currently implementation, I modified the P(P|W) from ratio which is computed from scim-pinyin, to k/n, k indicates k matched pinyins, n indicates total n pinyins for word W.<br />I don't know how it will influence the result of HMM correct rate, hopes it will not be bad.<br /><br />Tested this approach on research prototype, the correct rate is lower. And test with manual input, it don't work too bad.<br /><br />And counting the computing complexity, it seems that the speed is sufficient.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-44182433574169664502007-11-14T14:13:00.001+08:002008-02-19T12:18:47.020+08:00Novel-Pinyin Ver1 TODO List<span style="font-size:130%;">Storage:</span><br /><ul><li> N-gram file</li></ul><span style="font-size:130%;">Training:</span><br /><ul><li> Training using parameters in prototype system.</li></ul><span style="font-size:130%;">Lookup:</span><br /><ul><li> Original Lookup with Candidate Selection.</li></ul><span style="font-size:130%;">Self-Learning:</span><br /><ul><li> Learn User Sentence when Commit String.</li></ul><span style="font-size:130%;">novel-imengine:</span><br /><ul><li> Special Table support rewrite from scim-pinyin in c.</li><li> Scim UI Config Module.</li></ul>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-13930220502585748422007-10-16T10:28:00.000+08:002007-10-16T14:32:55.355+08:00Federico in BeijingNow Federico has come to Beijing. He is a great hacker and good man.<br />Here we together looking at the pango CJK performance issue.<br /><a href="http://www.gnome.org/%7Efederico/news-2007-10.html#pango-cjk-1">http://www.gnome.org/~federico/news-2007-10.html#pango-cjk-1</a><br /><br />And He went to the Great Wall:<br /><a href="http://www.gnome.org/%7Efederico/news-2007-10.html#15">http://www.gnome.org/~federico/news-2007-10.html#15</a>Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-87742872985255351252007-09-11T09:56:00.000+08:002007-09-11T13:48:10.553+08:00Planet SUSE blogFrom now on, I will post my blog in English, Because of the planetsuse.org.Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1167129173225021182006-12-26T18:27:00.000+08:002006-12-26T18:32:53.236+08:00让openSUSE 10.2支持包含中文文件名的U盘先是在hal的源码中找了半天,最后终于在gnome-volume-manager的代码找到了设置iocharset=utf8的方法。<br />在gnome-terminal下运行:<br />gconftool-2 -s /system/storage/default_options/vfat/mount_options -t list --list-type string [shortname=winnt,uid=,iocharset=utf8]<br />就可以解决问题了。Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1159862060234454432006-10-03T15:52:00.000+08:002006-12-27T09:29:06.120+08:00论文雏形基本搞定非常努力的工作了半年,看了N多资料。<br />终于,基于Hidden Markov Model的拼音整句输入法原型,可以正常工作了。<br />过一阵儿,在Linux下就可以用整句输入法了。<br />效果很好,不过,离成品还有一段距离。<br />呵呵,兴奋中。Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0tag:blogger.com,1999:blog-19351878.post-1159861953689717252006-10-03T15:42:00.001+08:002006-10-03T15:57:48.770+08:00十一回家休息!!!终于又回家了。<br />经过一番修改,我的Blog终于在IE6种正常显示了。<br />要把模板中的第一句话去掉,并把开头改成如下格式:<br /><html><br /><head><br /><meta equiv="content-type" content="text/html; charset=UTF-8"><br />...<br />-----------------------------------------------------------------------------------<br />就可以在IE6中正常显示了。呵呵Alex Epicohttp://www.blogger.com/profile/12195433461061993850noreply@blogger.com0