I’m hopeful that 2008 will be even better and hope everyone who reads this (yes, both of you) has a similarly hopeful outlook.
printf "t%d '%s' error%sn", $errorCount, $error, $errorCount == 1 ? '' : 's';
I know people don’t usually notice, but printf statements with ternary operators make proper pluralization easy to code, and the reports don’t irk me as much.
printf "For %s to %s %d of %d full.jpgs (%.2f%%) were missing. | ", sub{ sprintf "%04d-%02d-%02dT%02d:%02d:%02d", $_[5]+1900, $_[4]+1, $_[3], $_[2], $_[1], $_[0]}->(gmtime($start_time)), sub{ sprintf "%04d-%02d-%02dT%02d:%02d:%02d", $_[5]+1900, $_[4]+1, $_[3], $_[2], $_[1], $_[0]}->(gmtime($end_time)), $totalMissingFullJPGs, $totalFullJPGs, $totalFullJPGs > 0 ? $totalMissingFullJPGs / $totalFullJPGs : 0;
by Matt at January 05, 2008 06:04 PM
So, I did a full port upgrade a while back and it killed lots of stuff on my server. I got pretty much everything up and running rather quickly, but I was having problems with Apache and PHP which I simply set aside. Today I’ve finally done the finger work and figured out exactly what was wrong and, as you can see, fixed everything.
by Matt at January 05, 2008 05:49 PM
Came across this and had to read it twice to make sure I wasn’t seeing things…
http://blogs.telegraph.co.uk/ukcorrespondents/holysmoke/dec07/islam-resolution.htm
by david at January 01, 2008 09:36 PM
2007 turned out to be a fantastic year for me, despite a slowish start I’m hopeful that 2008 will be even better and hope everyone who reads this (yes, both of you) has a similarly hopeful outlook.
by david at January 01, 2008 09:32 PM
Yesterday I got my hands on a eBook reader, and I got mighty impressed. The Bookeen Cybook ePaper has the perfect size and is therefor easy to carry around. As I work on computers all day long, and have to stare on back lit LCD displays, I have been quite skeptical to eBooks. In fact, if the document I am to read is longer then 2-3 A4 pages, I prefer to print it out. I swear that I could feel my eyes relaxing when I read on the Cybook, it is the closest thing to reading paper that I have ever seen. Imagine being able to have all the Windows and MacOS X manuals with you when solving problems at work! Cool, isn’t it? Well, that is not the main reason why I want to buy it. I work at the university and at our department, we recently started making our scientific staff’s research/publications available as PDFs. Until now we have printed them and handed them to the students, and with making them available as PDF we hope to reduce cost and use of paper.
eBook Readers have suddently become amazingly popular, and we can thank Amazon for that. Their wireless reader Kindle became a success. Most likely cause of the possibility of buying books online directly from the device. It’s just that I don’t like Kindle’s design, as it is a book reader, I don’t want to have a thumbnail keyboard built in. The device is simply to big! Another problem is the inbuilt wireless card, it shortens the battery life. Cybook doesn’t have a keyboard and wireless card, and therefor achieves ultra-low power consumption. And what is better, it has a a thin, light form, and fits perfectly in your hands. The Cybook dimensions are very close to the ones of a paperback. As a paperback, the Cybook can be easily arranged in a hand bag. The size is 188 x 118 x 8.5 mm, and the weight is as low as 174 g.
The Cybook shows an outstanding black and white screen, utilizing a breakthrough technology: the Electronic Ink Display (E Ink®). The screen simply acts like a paper page. It’s amazing! The screen is not backlit, and it uses ambient light to enlighten the reading surface, so you cannot use it when it’s dark. On the other hand, the Cybook is readable in full sunlight which is not the case for almost all standard LCD displays.
eBooks are simple data files, and can be transfered with a USB cable from a host computer by drag and dropping your files. No software installation, no synchronization process, no file conversion, and best of all the Cybook is compatible with any kind of host computer (Windows®, Linux, Mac®). Can it get easier? The following file formats are supported: MobipocketTM .PRC, PalmDoc, TXT, HTML and PDF. What I am not totally sure of, is the Amazon’s Kindle format and Mobipocket the same? Amazon bought Mobipocket.com back in 2005, so why on earth do they call most of their eBooks Kindle Edition? It’s a little bit confusing cause it seems to be the same format. Is it to indicate that books are saved to fit Kindle’s screen resolution? The screen resolution are the same on both devices, so my question is can I read Kindle eBooks also on CyBook?
So where to get content? First of all you can save your files in all the supported formats. Many books are also available for free, and a good starting point is then MANYBOOKS.NET. The site is ran by my friend Matthew McClintock, you might know him for his BeOS SVG icons. Thousands of books are also available in the Mobipocket format (an encrypted and unencrypted format). According to Bookeen more then 200 resellers are using this format, so also the Norwegian Online Book Shop Digitalbok.no.
by frankps at December 30, 2007 10:42 AM
First is for Bug 152156, it allows proper Mozilla windows update/redraw when FilePicker is open.
Second, for Bug 217723, adds native file/mime type recognition.
How to test second one?
1)Try to open with current Mozilla some text or image file without extension (rename some file temporarily if it has extension). Mozilla fails at that attempt.
2)Try same with patch applied.
There is also third patch on hold - it adds filters to FilePicker, but I will wait checkin for Bug 152156
Just as one of my friends posted a Christmas photo of him on his blog, I thought I should do the same:
As every year I do the Christmas dinner, so also this year. I simply love making bigger dinners, and turkey is one of my favourite dinners, not only for Christmas. This year mom bought a beautiful 7 kg turkey. The biggest turkey that I have ever prepared here at home, with lots of stuffing. It almost didn’t fit in to the owen.
Another tradition is the Christmas present we send to our friends in Reykjavik, Iceland, and the presents they send us. Last year I blogged about 13 Jólasveinar (Yuletide lads). As you can see it is quite a collection of lads:
This year Jon and Fridur sent us the lad Gáttaþefur - Doorway Sniffer (December 22), and here is a short saying about the 11 th Icelandic jólasveinn:
Eleventh was Doorway Sniffer
- Who never had a cold,
Even though he had a funny
And enormous nose.The scent of Leaf Bread
He smelled in the hills,
And lightly, like the smoke,
He followed that scent.
The collection of Jolasveinarnir (Yuletide Lads) is almost complete, only three figures are now missing:
by frankps at December 28, 2007 08:09 AM
but I'm here with a purpose.
First, some background on me: I'm a software developer: MS Apps: VB, Access, Excel, etc. Web dev: PHP + MySQL, Cold Fusion, ASP
I've been learning C++ since the beginning of 2007, so I'm still a beginner in this regard.
I've joined this group because I need help and this is the only qualified group that can provide it.
As not to overwhelm you with details, here is what I'm doing:
I decided that I wanted to create a plugin for FireFox on Beos. After reading pages of reference material, I fould out I need to build XULRunner to get the Gecko SDK to create FireFox plugins.
Well I've tried to compile XULRunner twice. The second time I compiled it with what I thought was a fix to the error from the first compile. I was wrong, but I think I know this time. The error is: 'undefined reference to BFilePanel::xxxx in the file nsFilePicker.o' I believe the problem is the compiler is not including the Be libraries but I don't know hardly enough about the build process and the tools to correct it.
Can someone help me out? I'll stick with this until it's completed but I will need some hand holding.
Thanks in advance,
bekeeper
Hi,
Today I thought I would check out the status of bezilla.org, and to my surprise this domain just redirects to this blog. Nothing wrong with that per se, but I really wonder why it was taken down and what happened to the site there and its content.
I was also wondering, what happened to the tinderboxes that were donated a long time ago to the project? Where they ever put to any good use? If they are not being used, maybe we can find a good use for Haiku. Who is in possession of the hardware?
Cheers,
Norwegian has been flying Gardemoen - Budapest for years already, and have until now been alone on flying direct. Norwegian will also be flying Rygge - Budapest from upcoming spring. Their planes on the first route has always been more or less full, and the tickets have cause of this become a bit more expensive. But from next summer this might change, as the Hungarian low fair airline Wizz Air will start flying to Torp on June 26th. Here is some lines from their press release of today:
Wizz Air, the largest low cost airline based in CEE recently announced a series of developments from its bases in Hungary and Poland.
The airline said, that it would expand its Budapest operations next summer by adding 15 new weekly flights compared to summer 2007 or 27% more capacity. The airline will increase frequencies on most of its current routes, add Goteborg, Oslo-Torp and Venice-Treviso to its Budapest network and reintroduce the popular summer seasonal flights to Bulgaria (Bourgas and Varna), Spain (Barcelona-Girona and Palma de Mallorca) and Greece (Corfu, Rhodes and Heraklion).
Great to see that I will be getting cheaper tickets to Hungary from next year on, as according to Boarding.no prices will start at 199 NOK (incl. handling fees and taxes). Bus tickets will come on top of that.
by frankps at December 20, 2007 04:02 PM
Opera Software recently filed a complaint with the European Commission that was to aim at giving consumers a genuine choice of Web browsers. So far so good. The complaint describes how Microsoft is abusing its dominant position by tying its browser, Internet Explorer, to the Windows operating system and by hindering interoperability by not following accepted Web standards. Opera has requested the Commission to take the necessary actions to compel Microsoft to give consumers a real choice and to support open Web standards in Internet Explorer.
Microsoft of course defended themselves with the fact that alternative browsers could not only be installed on Windows, but also be set as default, but this seems not to be enough for Opera Software, as they want EU to force Microsoft to remove Internet Explorer as part of Windows (the OS). Funny how Opera hasn’t learned the lesson from the highly successful Windows XP N and Windows Vista N Editions. Not a single PC manufacturer has delivered a single PC with these editions, and in fact most EU citizens will have problem getting their hands on them as the demands are so low that shops don’t have them on their shelves.
When it comes to the tying the browser in to the OS, Opera is completely right. Internet Explorer has been integrated in to Vista in a way we have never seen any browser in to another operating system. And it has been successful, I dare say that Internet Explorer 7 on Vista is the most secure browser on the market today. When users have not turned off UAC (which they have been told to do by so many IT Professionals), Internet Explorer 7 will run with less privileges then a normal user. This is a Protected Mode that provides the safety of a robust Internet browsing experience while helping to keep hackers from taking over your browser, damaging your system and installing software. Internet Explorer Protected Mode protects users and their systems from malicious downloads by restricting where files can be saved without the users consent, and further Internet Explorer 7 in Windows Vista cannot modify user or system files and settings without user consent. Protected Mode requires the user to confirm any activity that tries to put something on your machine or start another program. By ensuring the user consents to these kinds of actions, the likelihood of automated and/or unwanted software installation is reduced. This feature also makes you aware of what a website is trying to do, giving you a chance to stop it and take time to double check the trustworthiness of the website.
Now this is not bad from an enduser’s perspective and should not let Opera Software demand Internet Explorer to be removed from Windows. What Opera Software should be heard with is the possibility to do the same with their browser as Microsoft has done with theirs, and for that Microsoft needs to corporate with other browser vendors.
In my eyes Opera Software did another mistake, as they through interviews (Wium Lie) have been focusing more on forcing Microsoft to support all major web standards that are commonly accepted within development communities. Internet Explorer 7 was a little bit better then 6, and Microsoft had already announced that version 8 would be released in not a to distant future. Well, yesterday Microsoft showed that IE 8 was passing the ACID2 test.
I’m asking myself what Opera Software have achieved with their PR stunt, and what do they think that they now will achieve with an EU trial? Even though I have never liked the bloated Firefox, Mozilla Foundation has shown that they are capable of competing with Microsoft on the desktop. A free advise for Opera Software, if you want to have any chance of competing with Internet Explorer and now Firefox, enable sysadmins to set settings through group policies with .admx files. For me group policy settings are far more important then voice, online storage of bookmarks and Ogg Theora support. They are all great features, but not something for the enterprise users. I don’t let my users get the option to use Opera, and with the latest security problems for Firefox, I have started to remove as many Firefox installations as possible. Safari is so far not an option. So in the end Internet Explorer will be the only browser that I let my Windows users have.
I guess it is just a matter of time before Opera Software will find that they are only a small browser vendor in the embedded market. Apple and Nokia seem to have great success with their webkit browsers. More and more companies get in to problems competing with open source projects.
by frankps at December 20, 2007 08:08 AM
Of course not, at least that is what most of us would say. Tonight I was watching the news on the Hungarian channel Duna TV (yes, some of you have seen me there before), and what did I see?
In Mindszent somebody managed to steal 6.000 m3 with sand from the local beach, and if we are to believe the city hall and the local police, they have not noticed anything special happening the latest days. No one has reported to them any unregular activities on the beach. At first I couldn’t do anything else then laugh! But think about it … Could the whole town have been sleeping? About 6-800 trucks with sand must have gone through the main street of this small town, and nobody have noticed? Unbelievable!
Here are the sad changes, in a before:
and after:
Sadly, as a friend of me commented at once she saw the pictures, this is a horrible environmental crime.
I can only hope that the ones responsible will be put on trial.
by frankps at December 18, 2007 07:25 PM
I have avoided opening a MacBook Pro for as long as possible, but in the end I simply had to. The laptops come with only one year warranty, and 5 out of the 6 machines we have of this model have had some kind of problems within +/- one year. Mine got a new motherboard 9 months after I bought it. On the machine on the picture, the hard drive seemed to have an error that made the machine beep quite frequently:
Changing the hard drive made me and my colleague remove about 20 screws! Why is it that people only talk about looks when they talk about design? And never about how practical the design is? Have a look at your laptop, and ask yourself where your hard drive is located and how many screws you need to remove to replace it? If your laptop is well-designed your answer would be 5-6 screws!
by frankps at December 13, 2007 05:37 PM
It’s finally here:
I can finally get rid of the stinking AIM client on Windows.
Simply go in to the Gmail Settings, under the tab Chat, enable your AIM account!
by frankps at December 04, 2007 10:23 PM
Out of sheer boredom (and a bit of interest), I tried building cairo. After satisfying the dependencies (freetype, fontconfig and libpixman), and fixing some small things, Cairo 1.5.2 builds, but does not link.
Update: Fixed. It was because the Cairo headers weren't imported as C-headers. Putting up an extern "C" guard fixed it.
I decided to do a little more investigation on the section of xpcom thread code that had entries for pretty much any platform except BeOS. I added a dummy BeOS section:
DisposeEventHandlerUPP(self->eventHandlerUPP);
RemoveEventHandler(self->eventHandlerRef);
+#elif defined(XP_BEOS)
+ /* Placeholder. need code to delete semaphores and port
+ */
+ #ifdef DEBUG
+ printf("pl_CleanupNativeNotifier HAS BEEN CALLED IN BEOS!!!!\n");
+ #endif
#endif
}
#if defined(_WIN32)
Just to see how often this bit of code is called by Firefox. Guess what? Fairly often - 14 times from launch to home page. So while I don't know what it does or if it's important to our port, I can at least say Firefox is trying to do something here.
It's possible that our unique window/thread structure and multiple restarts during launch make this code unwanted or unnecessary - we need the experts to weigh in on the subject.
Update:Well, we certainly don't want to delete ports at this point. I added code to do this and Firefox dies before the first window is ever created. It just shuts down and stops - no crash. So maybe I guessed right - if we delete ports here, we delete something needed for continued execution.
Isn't if fun to watch a non-dev try to figure out how code works? :)
tqh, did you ever file a bugzilla bug for your rewritten BeOS NSPR code?
Update: Found it - bug 300595
I’m surprised and can still not really believe that it could happen. I have always enjoyed staying in Germany, and one of the reasons was that I found the towns and villages that I visited to be clean, and that the local communities took care of their surrounding areas. A few days ago, my image changed completly!
Last year 20 villages in Hungary, two in Csongrád county, got delivered a total of 4.000 tons of rubbish from Germany! According to the county police of Bács-Kiskun Megye the trash (especially electronics and waste from private households) has been unlegally imported in a period from July to December last year. The trash has been brought to Hungary from Bavaria and Baden-Württenberg.
How can Germany be trading trash? Where is their ethics? Will EU take action? Of course, as always, it takes two to tango! A few Hungarians have become filthy rich …
 |
| Trash placed outside the small town Mindszent. Photo: Tésik Attila |
by frankps at November 29, 2007 07:43 PM
... my hosting service seems to be corrupting the .zip files somehow. I have uploaded the files using Netpenguin and Firefox. Results are the same either way... a file that unzipped and produced a usable Firefox locally is corrupted upon upload/download. Either it won't unzip or the resulting files don't run. ARRGH!
I discovered this problem AFTER updating the BeBits entry. I then discovered BeBits does not have a mechanism for rolling back to a previously published version. For now, the Firefox entry on BeBits contains a notice but no links to files.
I'll have to contact my hosting company in the morning. I'm really sorry about this!
Update 2: Now available on BeBits. Thanks to everyone who helped out with download/testing!
Now I’m not a vegetarian, so this obviously doesn’t apply to me! How often have you heard your mom say don’t play with the food? Here’s a bunch of people doing a serious effort on getting their moms upset, and they are good at it too:
Footage of the vegetable orchestra during shopping, preparations and some clips of one of our live shows.
From their website you can read they perform music solely on instruments made of vegetables. Using carrot flutes, pumpkin basses, leek violins, leek-zucchini-vibrators, cucumberophones and celery bongos, the orchestra creates its own extraordinary and vegetabile sound universe.
As so often, I picked up the news about this special orchestra at Joanne Colan’s Rocketboom, my favourite webcast.
by frankps at November 25, 2007 11:24 AM
A little while back, a bug turned up in Thunderbird's dogbert migration code. When support for Mac OSX was added, it appears ifdefs were inserted so code only executed for Unix and not OSX. Turns out some of that code is also needed for BeOS. I spent some time with LXR looking at the 1.8 branch and came up with a number of candidates where such conditionals occur. I'm posting the list here so everyone interested can look at it. Fyysik wanted me to put it in a safe place so he can look at it in his spare time. I hope this is a safe place :)
http://lxr.mozilla.org/mozilla1.8/source/xpcom/threads/plevent.c#342
http://lxr.mozilla.org/mozilla1.8/source/xpcom/threads/plevent.c#667
http://lxr.mozilla.org/mozilla1.8/source/xpcom/threads/plevent.c#954
http://lxr.mozilla.org/mozilla1.8/source/xpcom/threads/plevent.c#1346
--------comment says "nothing to do on other platforms" but I wonder...
Update: tqh, please look at line 945 and beyond. We might be missing a function.
http://lxr.mozilla.org/mozilla1.8/source/xpcom/threads/plevent.c#1734
--------comment says "extra function for unix" do we also need it?
http://lxr.mozilla.org/mozilla1.8/source/xpinstall/src/nsInstallFolder.cpp#610
--------we don't create an actual installer so maybe this isn't important?
http://lxr.mozilla.org/mozilla1.8/source/toolkit/xre/nsUpdateDriver.cpp#94
http://lxr.mozilla.org/mozilla1.8/source/toolkit/mozapps/update/src/updater/updater.cpp#116
--------these two are in updater code but looks like BEOS is missing. need to fix it if we ever put in updater
http://lxr.mozilla.org/mozilla1.8/source/browser/components/migration/src/nsOperaProfileMigrator.h#117
http://lxr.mozilla.org/mozilla1.8/source/browser/components/migration/src/nsOperaProfileMigrator.h#128
http://lxr.mozilla.org/mozilla1.8/source/browser/components/migration/src/nsOperaProfileMigrator.cpp#1090
http://lxr.mozilla.org/mozilla1.8/source/browser/components/migration/src/nsOperaProfileMigrator.cpp#1111
--------may be needed if migrating from old version of Opera for BeOS
http://lxr.mozilla.org/mozilla1.8/source/mail/components/migration/src/nsDogbertProfileMigrator.h#72
http://lxr.mozilla.org/mozilla1.8/source/mail/components/migration/src/nsDogbertProfileMigrator.cpp#141
http://lxr.mozilla.org/mozilla1.8/source/mail/components/migration/src/nsDogbertProfileMigrator.cpp#183
--------more things to look at in Dogbert profile migrator
http://lxr.mozilla.org/mozilla1.8/source/xpfe/appshell/src/nsXULWindow.cpp#976
Update: We can ignore this. Enabling it for BeOS means, among other things, the main window always starts at the top-left corner, rather than saving the window position from the last shutdown.
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/public/npapi.h#470
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/public/nsplugindefs.h#275
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/src/nsPluginNativeWindow.cpp#65
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/src/nsPluginHostImpl.cpp#2674
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/src/nsPluginHostImpl.cpp#3906
--------last one has to do with Java so we may not see this code. Need to check the others to see if they're needed for BeOS
http://lxr.mozilla.org/mozilla1.8/source/modules/plugin/base/src/ns4xPlugin.cpp#1941
--------two instances in this file. BeOS is handled separately in the 1st one, not present in this one.
http://lxr.mozilla.org/mozilla1.8/source/modules/libreg/src/VerReg.c#120
--------lots of references in this file. This is the first.
http://lxr.mozilla.org/mozilla1.8/source/mailnews/import/comm4x/src/nsComm4xProfile.cpp#51
--------another bit of migration code to look at
http://lxr.mozilla.org/mozilla1.8/source/layout/svg/base/src/nsSVGImageFrame.cpp#568
http://lxr.mozilla.org/mozilla1.8/source/layout/generic/nsObjectFrame.cpp#1902
--------platform specific printing code. could be important - if we ever get printing to work :)
http://lxr.mozilla.org/mozilla1.8/source/layout/generic/nsObjectFrame.cpp#2346
--------plug-in stuff. since there aren't many plug-ins for BeOS right now (are there any?) we wouldn't have seen this.
http://lxr.mozilla.org/mozilla1.8/source/layout/generic/nsObjectFrame.cpp#2346
There is solution how to use VLC to watch YouTube from BeZilla.
Inspite article describes Firefox tricks for that, it works also with SeaMonkey. Only difference is to install SeaMonkey-version of GreaseMonkey
Thanks, xeD!
Crap. It appears that it is impossible to open text files from disk in branch browsers. Our file-type handling is crap.
There was old bug about it, but today I filed two new, about partial solution of global problem.
One is to get first proper mime-guessing from extensions - I have solution for that, and second, forgot number - about adding TXT extension and PLAIN_TEXT type to default mime-table in mozilla exthandler.
After fixing redraw problems under nsFilePicker, started to implement Open And Save types/filters, existing for other OS-es, but missing in Mozilla for BeOS.
Today finished implementation polishing, filtering now works nicely but...
It seems almost stable - seems crashing only if to change type of filter very fast.
Seems that there is threadsafety problem here.
Array with filters belongs to main (mozilla) thread, while FilePanel is running in it's own thread.
Looks like such situation causes crashes at segment violation.
So trying to refactor code to more safe version.
P.S. Also noticed, when tinkering with all that, that we have problems at Open
- Mozilla don't respect (or just ignores) in some cases BeOS file type info, and thus complaining that, for example, text file is "octet-stream". And asks what to do with it. Wondering if bug on such topic was open already
Update: Added bug 404982 , submitted patch draft (applies against clean branch source). Seems no crashes with it till now, but will check more.
Seems so here, in my local source tree, as side effect of nsAppShell workflow investigation:
here is bugzilla comment on topic
I can provide interested people, tigerdog especially, with code changes to test it in their builds.
But till now things look like working perfectly, windows behind FilePanel are updating, modal mode works OK, no crashes when pushing "browse" buttons on pages once, twice etc, files are properly selected.
Build of Seamonkey-BONE with fixed FilePicker and appshell is available
aswell as small update lib, which can be used (I hope) with branch FireFox-BONE.
P.S.Looks like there is problem with small Firefox updates now, as available builds seems built with big componentlib instead bunch of easy replaceable (and interchangeable between all Mozilla apps) libs. So if you don't see lobwidget_beos.so in your system - you should wait for next FF or TB test build:(
Current source snapshot of my nsFilePicker.
I'm interested in testing it with old, standard app shell.
If it don't work with standard appshell, get new sources for appshell here.
If anyone besides me uses the Thunderbird email client, there's an updated build on BeBits. Thunderbird 2.0.0.9 was officially released on 14 November (I missed it by a few days.) I also "cheated" a little and included the cursor fix and gfx patches. They're not in the 2.0.0.9 code but both are going to be added to the branch and fix known and obvious problems, so I included them anyway.
For me, bezilla tends to hand on secure sites (with CVSem locked) if I use DIRECT CONNECTION to Internet in browser settings.
But does it very rare, if at all, if I use explicitly my provider's cache proxy set in browser.
Can someone who experience such https hangouts more regularly than me to veify this observation?
After applying the gfx patch in bug 343987, I've seen a couple of instances when screen elements didn't paint correctly. For instance, I left a very long download running last night with the download manager window on top of the main window. When I clicked the main window, the portion previously covered by the download window did not repaint. Also when preparing for a download, the dialogue box "open with/save" radio button did not paint when selected. I'll test some more but I'm a little concerned.
Update: call my experience a fluke. I have not seen the problem again.
Started to investigate our old app-shell used in branch.
Noticed that GetNativeEvent and then DispatchNativeEvent are called in cycle when modal, and you guess - native (FilePicker) window is open.
Implemented GetNativeEvent which was previously pure stub and then rewrote a bit DispatchNativeEvent to work in pair with Get.
Will see if I can to rewrite some other parts, to allow it to be more smoother.
When that is ready and I'm more familiar with all that mess, will try to implement there some tqh's ideas from his trunk patches. E.g. get rid of ports.
Hi there, it's been a while.
This weekend I dug up my old computer which contained the firefox source. I couldn't resist playing with my filepicker patch. Just
to let you know how things have been going:
I'm pretty sure that attachment 244713 [details] is _the_ way to fix this issue in
Firefox 2.
There are two issues:
As BeOS is quite bad in situation when real RAM is over, I think that Mozilla's fame Pavlov utility RAMback is really worth to test, especially on computers with < 256 MB RAM.
"RAMBack will cause Firefox to issue an internal notification to free up memory that is otherwise held for performance purposes. Additional items will be hooked into this notification in the future."
BUT! But:(
"Works with - Firefox: 3.0a8pre – 3.0.0.*"
So, inspite it looks like very essential for us, we cannot use it atm - maybe there is some smart enough among readers to try to backportport it to 2.0.* versions?
Update Johndrinkwater fixed installer to accept all versions since 0.96+, you can try it from here
I've now surfed with Firefox in Haiku, and it worked fairly well. I didn't manage to log in and post here though. The big issues seems to be the widget-code, especially which view has focus/mouse handling, getting drawing notifications from OS and the destructor code.
For instance if you move some part of browser offscreen and then back it will not redraw properly. Mouse seems to be locked to certain views, and it did crash on nsWindowBeOS destructor at least one time.
Not sure when I'll start really looking into those things though.
I have just been notified by a colleague of mine that I am to hold a speech on this year’s Microsoft TechEd in Barcelona. OOOppppps!!!
Just to make things clear, I am not … Something must have gone horrible wrong on the Microsoft Tech Ed site…
by frankps at November 12, 2007 09:49 AM
2.0.0.X's argv handling:
http://lxr.mozilla.org/mozilla1.8.0/source/toolkit/xre/nsNativeAppSupportBeOS.cpp#56
Trunk:
http://lxr.mozilla.org/seamonkey/source/toolkit/xre/nsNativeAppSupportBeOS.cpp#113
From bug filed over at Haiku: Firefox crashes with "-safe-mode" switch.
I've created bug 403406 to implement the cursor fix I placed in the experimental version. This seems to work well without having an undue performance hit. Ultimately, maybe we create some sort of global variable to reflect the current cursor instead of relying on mCursor. For now, this makes Firefox (and probably Thunderbird) look much better.
Somewhere in last weeks Gmail started for me to be troublesome with our Mozilla builds. No, not SSL hangouts, I didn't experience https problems here for long long time. But looks like with addition of lot of AJAX crap at Gmail default page. Now it tends to consume 100% of CPU even if you do nothing in opened gmail page, making Bezilla much less responsive. May be not so noticeable on modern machines, but if you have ProcessController or Pulse visible on desktop - place attention there.
Probably I will try to look in our ancient nsAppShell code for some reforming, using our experience with new AppShell, but unsure atm if it helps. Maybe there are really ugly loops now in gmail JS code.
I've joined the people who are trying to get Firefox working under Haiku by installing Haiku on my PC. While I have no network under Haiu (yet), we seem to be at least a little bit closer after Marcus O fixed a memory corruption problem in Haiku's BPolygon:AddPoints.
If you are interested the bug is here: Haiku bug 1427: Firefox crashes on start.
Both SeaMonkey and Firefox sometimes deadlock for me with O(zero) CPU usage
on Dual CPU 550 MHz, but never do such grip on faster single-CPU machine.
Audacity is a great application - free, fully featured and easy enough to use. Or it was.
Ubuntu now has 1.3.3 in it’s packages and this version appears to remove the option to import tracks! The most useful option and the one that every podcast creator relies upon for creating their final audio is gone! Missing. Denied. Have they gone mad?
Apparently the version that Ubuntu is packaging is a variant called CleanSpeech, which adds some useful features, but it does so at a price that seems far too high. If the ability to have multiple tracks in a single instance of Audacity exists, please tell me as it’s now stopping the release of the next Feathercast episode, in fact it’s stopping any further Feathercasts!
To say I’m annoyed and confused is an understatement.
UPDATE: I downloaded the Windows version of 1.3.3 and there were all the features that I expected to find. Thinking it was a simple case of different versions I downloaded the source tarball and built it on Ubuntu to find exactly the same version that the package delivered! It’s very annoying as I don’t currently have a version of Audacity that will allow me to create podcasts at AC US next week. I’ll try and get an older version working as an interim measure, but of course that requires an older version of WxWidgets… *sigh*
I wish the Audacity team would get their act together.
UPDATE #2: It’s possible to get back the old Audacity that you know and love! Under Preferences > Batch there is an option that controls whether Cleanspeech is used. Simply remove the tick for this option (and restart if using 1.3.3 but not required with the code in CVS) and all is well. Why the option defaults to On is a question only the audacity team can answer.
by david at November 07, 2007 11:50 AM
OK, so we get it - the “Jesus Phone” is about to be unleashed on a bunch of poor unsuspecting idiots ready to part with a serious amount of money, but do we really need the adverts to be aired so much? Purlease.
Although, I am looking forward to the end of world hunger once enough get sold
Oh, but the environmental cost of mass producing them seems to be rather a high price to pay…
by david at November 06, 2007 10:16 PM
I created Zeta and R5 builds from the latest branch, plus tqh's experimental NSPR and a tiny fix to eliminate the problem of the wrong cursor type sometimes being displayed. I reordered the files according to tqh's instructions below, then zipped them up. If anyone is interested in trying one, please contact me directly by email d o u g ( a t ) s h e l t o n f a m i l y (d ot) o r g. I'd like to know who has a copy, just to make sure we get appropriate feedback and to insure this version doesn't get to the general public without some level of documented trial.
If it seems successful, we can look at trying to get these changes incorporated in branch code.
What a hopeless company! Their websites are so full of PR and marketing that they’re useless for finding information of any use. When you call them they answer reasonably quickly, but leave you hanging on and are unable to help you. If they weren’t the best of the providers (hardly a ringing endorsement) then I’d change provider…
My current state of bewilderment comes from my attempts to sign up to use Vodafone to pay for Wifi access when travelling. The registration site came up OK, asked me to register, then crashed with a highly generic error code. This was annoying, but when I was told that the only way I could register was to visit a shop or use a windows PC I was stunned. So much for Vodafone being a technology company!
Today, sitting at a Vista powered laptop and about to travel again I thought I’d go to the website and register. Surely this would be easy? No. Apparently the site doesn’t exist - according to the Vodafone website searches at least.
After 2 calls to the customer support lines (the first was disconnected for some reason), they eventually managed to tell me how to find it. The next problem was that my abortive initial attempt managed to create an account, albeit not one that could do anything! How this was possible using an Ubuntu PC but actually registering for the service wasn’t is something that only they can answer.
After calling again to get my password reset it now tells me that I’ve registered. We shall see later on today when I’m next in the airport.
by david at October 25, 2007 12:40 PM
I’ve been playing with websites for a long time now and have used the usual variety of technologies along the way. However, I’d always resisted using any of the “frameworks” as when I’d originally looked they didn’t seem to offer me much and required me to learn a new language or technology. A few months ago this changed.
I’d been trying to find the time to create a website and had been tackling it using PHP. This made sense for my server already had PHP and I’ve been using it for a long time so knew my way around. Going was slow though as writing stuff from scratch takes time and time was the one thing i didn’t have. Bemoaning this fact to someone online led to them persuading me to try django. Initially I was reluctant - it’s a framework after all - but the fact that it’s python and python has been one of the things I keep meaning to spend more time with swayed me far enough that I downloaded it and started playing.
Several months later I’m a convert! It makes writing sites easier than I’d thought they could be and I really like the way it does things. It was a little odd to start with, but that soon passed as I started to spend more time with it. The biggest problems have been some of the documentation and my general lack of python knowledge. I should clarify and say that I’m not knocking the documentation that’s available, which is generally of a very high standard, but at times there is a lack of examples and explanation that I found puzzling. The #django IRC channel on freenode is very helpful and one of the most welcoming I’ve come across.
While the initial site I started working on still isn’t finished, another site is approaching the point that could be considered complete - flickrVotr! This is a small site that aims to allow people from flickr to hold competitions in an easier manner. It still has a few bugs, but it’s basically all there and starting to work well enough for me to advertise it here. It needs a revamp on the looks front, but then which of my websites doesn’t? If you have a look and come across any problems please let me know!
Finally, a huge thank you to Ian Holsman who is providing me with the hosting while I get my own server up to speed for mod_python.
by david at September 30, 2007 03:38 PM
I bought an SB800 flash quite a few months ago to complement by aging SB24 and to add the new features that make it ideal for off camera work. I wasn’t overly convinced how much I’d use it, but there was a deal on the SB800 that day and I figured it was worth the extra. How right I was! It’s now become a standard part of my bag wherever I go with the camera (which is pretty well everywhere these days). It’s been used in all sorts of places where I’d never have thought of using flash and has resulted in some very cool pictures. Along with a lot of fun, it’s forced me to think about my pictures in a new and different way. Basically it’s been well worth the money!
As I’ve grown more and more used to using the SB800 I’ve started to find places that I wanted another flash, places where the aging SB24 with it’s short off camera lead just doesn’t work. The obvious solution was to get another SB800, so I’m now the proud owner of 2 of these great flashes. work is pretty hectic at present so I’ve not had as much chance to play with them as I’d have liked, but hopefully soon. Will the addition mean better pictures? I don’t know, but I hope so. It’ll certainly give me more flexibility and opportunities. Watch my flickr stream for the answers!
by david at September 29, 2007 09:48 PM
(10:48:20 AM)reinsnitz: you got a date?
(10:48:25 AM) gildncranz: yep.
(10:48:30 AM)reinsnitz: :D
(10:48:33 AM)reinsnitz: can we come and watch
(10:48:40 AM)reinsnitz: just kidding
(10:48:49 AM)reinsnitz: more like spy
(10:48:53 AM)reinsnitz: but I am kidding still
(10:49:18 AM)reinsnitz: I saw the place to take her last night
(10:49:22 AM) gildncranz: As entertaining as that would be AFTER THE FACT, it would be torturous to go through.
(10:49:23 AM)reinsnitz: but they won’t open for a couple weeks :(
(10:49:33 AM) gildncranz: Well, damn.
(10:49:40 AM) gildncranz: We’re going out tomorrow night!
(10:49:49 AM)reinsnitz: at about I70 and C470
(10:49:51 AM)reinsnitz: on the west side
(10:50:19 AM)reinsnitz: by Home Depot and Kholes
(10:50:22 AM)reinsnitz: Jack in the Box
(10:50:34 AM) gildncranz: You’re obsessed.
(10:51:00 AM)reinsnitz: no I’m not
(10:51:08 AM)reinsnitz: I just know what the BEST DARN FAST FOOD IN THE WORLD is
(10:51:21 AM)reinsnitz: 2 Taco’s and a Jumbo Jack please
(10:52:13 AM) gildncranz: I would just like to point out that you are a lucky, lucky man that you met chelle.
by Matt at September 24, 2007 10:09 AM
In recent weeks I’ve been taking more and more pictures using the external flash. Mostly it worked great, but from time to time I noticed that it didn’t seem to work! Dark images were the result. Two shots, both with identical settings would give one light and one dark image - very odd! The problem, as it turns out, is with my “mental model” of what I was doing.
The SB800 is a pretty powerful flash so when using it indoors I’ve normally got the power down at 1/16th or lower. This gives me excellent recharge times and usually more than enough light. When moving outdoors things change a bit when I need more power. Thinking about it rationally as the power goes up the recharge times will get longer - which if you add in batteries that are past their peak will equate to longer waits. Of course having grown so used to the “instant” recharge that’s available at low power settings this didn’t cross my mind until it was pointed out today.
The results from the last session were fine until I increased the power of the flash - at which point I (as model and director) should have built in some longer pauses.
Thankfully, some of the pictures didn’t need the extra power
Hopefully having learned this lesson the hard way I won’t forget it in a hurry…
by david at September 23, 2007 04:40 PM
My foray the other week back into the coding world of python has done two things for me:
by Matt at September 21, 2007 01:48 PM
No, not the type of flashing that is covered by “Breach of the Peace, Indecent Exposure”, but rather the camera style of flash that has become popular thanks to Strobist.com. Armed with a few small flash units and some cameras we headed for the ruins of Old Chapel on the hillside facing Edinburgh city as the sun slowly set. It was light for longer than I expected it to be and so we had time to setup and play with the strobes and the twilight.
There were only three of us, so when we arrived we all headed off in different directions with our cameras - meaning that I was lacking a model! Without my IR remote I was forced into self timer and taking pictures of myself!
As the sky darkened Scott donned his leather mac and climbed onto some nearby rocks for a series of “heroic style” pictures!
It was a good learning experience and points the way to the next meetups for this activity. The whole set can be viewed here!
by david at September 21, 2007 10:13 AM
Boulder is known for it’s athletic inhabitants, but there’s a seriously geeky streak running under the surface: In RadioShack Sucks I mention JB Saunders, the Brewer’s Association is based out here, ton’s of tech giants have offices in the area, and, I just learned tonight, Spark Fun is just up the diagonal.
A couple of months ago I decided that some of my projects really didn’t require a computer; lots of the stuff that I want to do involves simple logic based on simple simple sensor data, such as a lagering control for a chest freezer. The answer to this problem seemed simple: learn to program a micro-controller and build circuit boards. After doing some research I decided the Arduino was the chip to go with. Unfortunately I couldn’t find a local supplier with a pre-built Diecimila boards, I did, however discover that Spark Fun Entertainment is a good resource and I got on their waiting list. Tonight I got an email saying the boards are ready and I promptly filled up my cart. At checkout I noticed a local pickup option tagged “For Colorado residents only”. What? I found that they’re just north of Boulder. Wow, I could pickup the board and hit JB Saunder’s on the way home.
Sweet.
by Matt at September 20, 2007 09:59 PM
After a somewhat busy time at work I’ve been on holiday this week. It’s been a great week so far with the weather playing ball and staying pleasant. The sun has even been present the last few days! (Yes, the head is slightly warm and glowing again ) Rather than travel any great distance I’ve been staying in Scotland and spending the week with Rosie in and around Perth.
We’ve been making the most of the weather and have been out and about every day this week. Highlights have been walking at the Birks of Aberfeldy and a visit to “Rumbling Bridge” besides Dunkeld with it’s great waterfalls. (The pictures are geotagged if you’re curious where these places are).
Rumbling Bridge
Birks of Aberfeldy
Today saw us walking in Kinoull Park and ending up at the fantastic folly that overlooks (and dominates) the M90/A90 junction just outside Perth.
We’re heading down to Edinburgh tomorrow and hopefully the weather will stay nice. It’s going to be hard to go back to work next week!
I’ve continued working with the excellent Django for a couple of website projects, with one managing to limp to a limited publication this week that has resulted in a lot more eyes looking and bugs to fix. Hopefully it’ll be stable enough to announce publicly soon.
by david at September 19, 2007 09:27 PM
I decided to look over my LibraryThing Catalog and found that there were now several duplicates. Last December I manually entered sixty or seventy ISBNs into LibraryThing and know that they are all represent the correct edition, so I went through comparing the different versions of specific books in my catalog. Every single book that was misrepresented was imported from Google. It looks like Google doesn’t actually store the ISBN which was originally imported, it stores a representative ISBN.
This observation lead to my realization that it makes sense. LibraryThing is designed to be used to catalog personal libraries. Users should be able to keep track of each and every book they have, even if they have multiple editions of the same book. Google’s “My Library”, on the other hand, isn’t supposed to be a catalog for a library, it’s supposed to represent an electronically searchable collection of hardcopy books—most ISBN changes refer to editions which changed the cover or other minor details without actually changing the content (Google probably doesn’t even differentiate between revisions of manuals, but I’d have to check some of my O’Reilly’s stuff to either confirm or refute that). Yes, it behaves similarly to a library, but it’s not generally concerned with specific editions. I am saddened that it doesn’t at least keep track of my specific ISBNs, but I’m not really surprised.
What this means, especially combined with LibraryThing’s increased success at recognizing ISBNs (not to mention that you can manually enter the information related to unrecognized ISBNs) that if you want to maintain both a LibraryThing Catalog and Google Books My Library simultaneously—without entering each book into each system individually—you should enter all the books into LibraryThing, export the catalog, and then import it into Google Books.
by Matt at September 18, 2007 10:14 PM
Over the weekend I scanned a few hundred of my books and uploaded the ISBNs to Google; in the process I found about a third of the ISBNs failed to load. I payed attention to which books loaded and which failed, but there doesn’t really seem to be a pattern as to which went through and which didn’t. At times newer books failed to be entered, and that’s understandable, however when I scanned my boxed set of “The Hobbit and the Lord of the Rings” The Hobbit and The Two Towers both entered while The Fellowship of the Ring and The Return of the King both failed. There have been several other instances where portions of sets, or other closely related books, failed.
As a test I tried several of the ISBNs in LibraryThing and not a single one failed. Interested in having a single online database of all my books I pointed LibraryThing at my Google RSS and waited, and waited, and waited (importing multiple books at a time is a slow process involving waiting for hundreds, or thousands, of other books to clear the queue ahead of yours). This morning when I found my books had finally loaded I noticed that one of the books had failed.
ISBN 0573620776 apparently isn’t in the databases of Amazon.com, The Library of Congress, or Amazon.co.uk, however Google has it (along with images). I’m most surprised that it’s not in The Library of Congress by ISBN (although a search for “Childhood” by Thornton Wilder does return a hit). Curious how well LibraryThing will continue to perform I’ve pointed LibraryThing at my Google Library RSS again and am now waiting for it to check on 108 new ISBNs.
Additionally I’m editing my :CueCat python script to work with LibraryThing as well as exporting Google or outputing to a file. Part of the reason I’m doing this is that LibraryThing keeps track of failed imports—for how long exactly, I haven’t tried to check—listing exactly which ISBNs it couldn’t load. Google, however, simply states the number of ISBNs it failed to import.
by Matt at September 17, 2007 07:06 PM
Boss: I love what you’ve done with this script, but I’d like you to check for command-line variables, prompt for variables if needed, and perform error checking.
Me: Sure, no problem.
<edit script />
<test script>
[[: command not found
</test>
Me: WTF?
<check code />
<double check code />
<bash -version>
GNU bash, version 1.14.7(1)
</bash>
Me: NO!
Me: Is this really necessary? This is bash 1.14 for crying out loud.
Boss: Yes, you have to make it idiot proof.
Me: Can’t we just make our hiring process idiot proof?
Boos: What makes you think you’d be working here?
Me: Touche.
by Matt at September 14, 2007 04:12 PM
Has somebody complained about the copyright status of something you posted on the internet?
Has it been removed because of this complaint?
Do you contest the claims?
File a Counter Notification Letter.
by Matt at September 14, 2007 01:11 PM
by Matt at September 12, 2007 01:19 PM
Powered by Planet!
Last updated: January 08, 2008 11:00 AM