Version 0.51
This is mostly a bug fixed version. But you’re now able to get the list of hits right from the start. So I removed the “Go” button because it’s not necessary anymore.
The “article not found” bug for the English edition is also history.
Image support is in and a packager for images is also available (OS X and Windows) but the spider, which downloads the images, is not ready yet. But the indexer spills out a list of used image names now.
So if you’re focusing problems (mainly the users of the English edition) update to this version. Otherwise stick with the previous one. Over at Goodle code I’ve added a list of changes.
Februar 26th, 2008 at 21:38
Woot!
Already up and running on my pod
it’s a pleasure to see this app evolve!
Thanks alot!
Februar 26th, 2008 at 22:03
Hello,
Quote:
“So I remove the “Go” button because it’s not necessary anymore.”
But I need the “Go” button, e.g for searching “pi”
Thank you for the great project!
Februar 26th, 2008 at 22:10
Corber-
you do need the button? Try it. You won’t need it. As soon as you enter “p” you get the first 25 articles beginning with “p”. And after you’ve entered “i” the article name “pi” leads the list.
It amazing how fast the binary search algorithm is. Every time you add a letter over 1 million (in the German edition) or 3 million titles (English one) are searched. I’m still impressed about that.
-Tom
Februar 26th, 2008 at 22:42
Hi Tom,
Just updated and checked that ‘Manchester Central FC’ article I mentioned before, and it now works fine. I think ‘animal’ had been mentioned as a non-working article as well and that’s fine now.
Looking forward to the image thing working. It looks like I will need to re-index for that, and I unfortunately neglected to keep the original file from the Wikipedia site, so I’m off to download that again!
Thanks for a great app!
Jason.
Februar 26th, 2008 at 23:17
Hi Tom
A small bug when searching Chinese articles: The hist list doesn’t show up automatically until I enter an additional “Leerzeichen”.at the end of the character(s).
For example, When I input “zh:中”, no list shows up. After I enter a space “zh:中 “, then the hit list shows up. I need to enter a backspace to remove the space and continue to enter other characters ending with an additional space to get further hit list.
BTW: I installed German and Chinese data files.
Februar 27th, 2008 at 00:39
what do i do with the image document? or is that just to let me know which images to download.
Februar 27th, 2008 at 00:57
Tom,
I LOVE the update!
Searching is now so much faster, and much more pleasurable to use. For example, in v0.50, entering the text “wikipedia” would pause at the 4th character for a LONG time, then process and process again for a LONG time until results were returned. It took literally minutes for results to return. Now, with v0.51, it’s instantaneous.
You once again raised the bar in an EXCELLENT application.
I can’t wait to begin working with images.
Thanks!
-Jim
Februar 27th, 2008 at 02:31
Tom,
Love the update, unicode search now works, but instead of taking directly my Chinese input as the search term, I need to enter a space using the English keyboard for the search to actually take place. (Chinese input on iPhone is done by hooking a customized input method to the Japanese keyboard)
A more important problem in this version: seg fault when using the -t+ flag on some articles:
URL: GET /wiki/zh/%E6%A4%8E%E5%90%8D%E7%A2%A7%E6%B5%81 HTTP/1.1
zsh: segmentation fault ./wikisrvd -p 8888 -v+ -t+
Februar 27th, 2008 at 05:41
Thanks for fixing the problem with the “article not found” bug now animal works =) ,
I more thing i noticed (not really a big problem), if i enter the word “ipod touch” it gives me 3 suggestions, Ipod touch, IPod touch and IPod Touch but only 1 works and the rest say “Redirected from Ipod touch”.
Thanks for the great app! . , Other features to include may be an option to increase the text size of the suggestions and the amount of suggestions =)
Februar 27th, 2008 at 06:20
More information about the search term problem: so the problem is due to the Chinese input method (I am using NativeCN) not firing the onkeyup event. Since NativeCN is a closed source software there’s nothing we can do about it, however adding onchange=”textBoxSearch_Change()” allows user to get search results without switching input method back to English by requiring the user to close the keyboard and focus on other area of the page.
Februar 27th, 2008 at 08:31
Sanford-
how about that: Set up a time in body.onLoad like this (not proper JavaScript!):
_oldValue = textBoxSearch.value;
window.setTimer(CheckInput(), 200);
function CheckInput
{
var newValue = textBoxSearch.value;
if ( _oldValue!=newValue) window.setTimeout(Changed());
_oldValue = newValue;
}
Every 200mms (or so) the value is checked.
-Tom
Februar 27th, 2008 at 09:26
Tom,
That’s even better! But then it’s a rather small problem compared to the sigfault. Have you tried opening the article /wiki/zh/%E6%A4%8E%E5%90%8D%E7%A2%A7%E6%B5%81
with -t+?
The article on zhwiki:
http://zh.wikipedia.org/wiki/%E6%A4%8E%E5%90%8D%E7%A2%A7%E6%B5%81
Perhaps it’s the info box on the right that creates the sigfault?
It uses a template with unicode character as its name:
http://zh.wikipedia.org/wiki/Template:%E9%85%8D%E9%9F%B3%E5%93%A1
I will try it again on 0.50 to see if this problem is introduced in the template fixes in 0.51
Februar 27th, 2008 at 13:25
Hallo Tom,
erstmal vielen dank für deine ganze Arbeit, das ist wirklich ein grandioses Tool geworden! Mach weiter so!
Wäre es evt. möglich in die Suchleiste so ein “x-Button” wie bei Safari zu integrieren, so das beim drücken die komplette Eingabe gelöscht wird?
Das fände ich sehr nützlich.
Mfg Michael
Februar 27th, 2008 at 17:13
Okay… tried it again. so the problem existed in 050b as well as 046a… so it’s not introduced lately.
Also the following works perfectly for chinese text searching:
function setupInput() {
window.setInterval(”CheckInput()”, 200);
}
function CheckInput()
{
var textBoxSearch = document.getElementById(”textBoxSearch”);
var newValue = textBoxSearch.value;
if ( _oldValue!=newValue) {
window.setTimeout(”LoadSuggestions()”, 0);
};
_oldValue = newValue;
}
works perfectly for Chinese input!
Februar 27th, 2008 at 18:57
Hi! I’m having another problem now, when ssh’ing the file articles.bin into the iPhone, it always stops at 2.0Gb and does not allow me to resume, it says “invalid argument”… Anyone know why this might happen? Could it be that cyberduck has a cap on files larger than 2.0 Gb? what other programs do you mac users use to ssh into your iPhone? I tried indexing everything again and even downloading everything again and nothing
Can anyone help me? I really really really want this app!!!
Thanks
Lourenço
Februar 27th, 2008 at 19:05
I have set up a temporary installer source for Wiki2Touch at
http://168weedon.com/i
You should probably submit the following apptapp package plist to one of the trusted sources:
http://168weedon.com/i/plists/Temporary/Wiki2Touch.plist
Februar 27th, 2008 at 20:51
Hi
If someone has an Image download Link please let us know it.
Thanks
Februar 27th, 2008 at 21:48
Sanford-
Wow great! I’ve just tried your installer source and guess what: It worked! Thanks a lot for providing it! Really cool!
Yes, give a time until the weekend. IRL I’m busy tomorow and on Friday. Then I will send the source to one of the trusted source providers. if your server can’t stand the load we can move it to this website.
Anyway, I’m really exited about the the installer package.
-Tom
Februar 27th, 2008 at 21:51
Lourenço-
well hit me but since Leopard I’ve the strange feeling that Cyberduck isn’t reliable as it was before. I’m currently using MacFuse to get connected to the iPhone. The file system of the iPhone is mounted as a Filesystem in User SpacE. Simply a desktop icon to click on. But the setup is not easy.
Otherwise no suggestions. Sorry. Anyone else?
-Tom
Februar 28th, 2008 at 10:03
If you’re on a mac, try it with AFP. It works fine for me though. (doesn’t matter if it is ssh via cyberduck or AFP with finder).
The only negative aspect of AFP is, that finder will create a bunch of folders in the filesystem of youre phone. You could delete them per ssh or mobile finder (ds_store, network trash folder, etc…) cause they’re not needed by the phone or pod.
haven’t had any problems with it.
Chris
Februar 28th, 2008 at 13:42
I use Leopard and last time I had to copy across the articles file I used a program I already owned called ‘iPhoneDrive’. It has recently changed name to MegaPhone. It’s a program that gives you direct access to the main storage area of your iPhone or iPod Touch. It’s payware but I don’t regret buying as it is much more convenient than using Cyberduck all of the time. You still need to use Cyberduck (or a similar program) to copy stuff to the smaller system partition though.
See http://www.ecamm.com/mac/megaphone/
It also copied across much faster than using Cyberduck.
A handy new feature is the ability to copy notes to and from the device.
Also, I should stress I have no association with the makers of the program, just a happy user!
Jason.
Februar 28th, 2008 at 21:59
Another Installer source - I am playing about with distributing the data dumps (Russian 20080220 294MB, Belarusian 20080208 4.3MB, Hebrew 20080225 113MB are up). Image packs later probably.
This will not be up to date to the day (it is primarily for personal use - will get updated when I get around to updating), and will most likely not host the English dump (which is impractical to install via Installer due to size since you need twice the space for the zip to be downloaded and unzipped).
Source: http://www.in7ane.com/iphone.xml (note the lowercase)
Category: Wiki Apps
Just the Russian language.config: http://www.in7ane.com/iphone/wiki/ru/language.config
Tom, let me know if you do not want people mirroring Wiki2Touch and I will take it down.
—
For large mac iPhone file transfers http://code.google.com/p/iphonedisk/ works well (open source, MacFUSE based).
Februar 28th, 2008 at 22:16
Tom, a request to make Image:/Bild: configurable through language.config or otherwise, so that images can be added to non-English and German Wikipedia dumps.
Februar 28th, 2008 at 23:04
Hello,
a 2nd qustion:
for example, when I type in “Mutter Courage und ihre Kinder” (german version) and enter this article. I don’t get the original Wikipedia article.
“Viertes Bild”; “Fünftes Bild”… are not there. Why?
I have followed the instructions
Thanks,
Corben
Februar 29th, 2008 at 10:29
In7ane-
yes for sure, adding the proper prefix for images to the language.config is very important. Further, adding that to the indexer so that the list of images names ist complete. Will do that as soon as spossible.
Mirroring the binaries isn’t usefull I think. Actually there are hosted on Google code and are not very large. For the installer package yes, why not. But finally I would realy love to see it in one of the “official” communitiy sources for the installer.
Having the archives for articles and later images mirrored or at least somewere is grerat for sure. I already found a place the the image pack I’m working on. But the package is for a couple of reason not ready yet.
-Tom
Februar 29th, 2008 at 10:32
Corben,
I can confirm this aswell. These small passages are missing for me too, tried with 0.46 and 0.51. (same article.bin, created with 0.51 indexer - German dump)
maybe it was changed in the meantime, or it is a bug.
Chris
Februar 29th, 2008 at 11:24
Corben, Chris-
when there was no image support (and officialy there is’nt one yet unless you download the images one by one for your own) I’ve not supported embedded images and images galleries when they appear as tags. It simply didn’t made sense.
Now as image support is upcomming I will added support fort these tags of course. But before image support must be established well enough.
-Tom
Februar 29th, 2008 at 12:06
Hey Tom,
it isn’t meant to be a “picture”…. it’s just called “Fünftes Bild” and so on. In Wikipedia online theres just text, no pictures embedded in this article. so that’s what made it so weird.
chris
Februar 29th, 2008 at 15:24
here is the link:
http://de.wikipedia.org/wiki/Mutter_Courage_und_ihre_Kinder
März 1st, 2008 at 00:42
Hi, I have problem with Templates not displaying correctly in czech wikipedia. I´m not sure, but i think Template is Šablona in czech so I put it in language config ( eng wiki article about templates is http://en.wikipedia.org/wiki/Wikipedia:Template_messages and when I click on the czech eqivalent, it´s called Šablona too). But it doesn´t work - it just display a red text ‘ablona:’ (without ‘Š’ - maybe this is the problem) and name of the template. Is there a problem with my translation of Template or its something else ?
And my second problem is that I would like to search articles with diacritic (ŠČĚŘŽÝÁÍÉŤĎŇÓÚ) but I can´t because I don´t have them on my keyboard. Could you change the searching to ignore diacritic so I could search articles with diacritic ?
BTW thanks for excellent app, It´s really great work.
März 1st, 2008 at 00:44
And other thing what I would be happy of is better implementation of mathematical symbols and operations.
März 3rd, 2008 at 22:04
in7ane-
in 0.52 I did the changes you need for image support. May you add the transaltion for “Image:” to the Russian language.config and send it to me again? There is now a key called “imagePrefix”.
Further I like to add it to the next release as one of the languages which are installed by default. Use warz@gmx.net as the email address.
Thanks in advance,
Tom