Hosting site for oxz files

[cim]

The download link contain a random part (I got images/f/f3 now), which we must insert into the manifest.plist within the oxz so an instant update is needed after every upload.

The easiest solution is probably to not put the download_url at all in the copy of the manifest.plist which is inside the OXZ. By the time anyone can read that file, they no longer need the download URL.

[maik]

It is possible to change the wiki configuration to not save uploads in these subdirectories (generated based on an MD5 hash of the file name btw). However, I'd have to move all media that has been uploaded so far into from their subdirectories to a flat file structure. If cim's proposal is ok then I'd rather avoid doing that.

[maik]

Dylan just confirmed the requested changes, uploads to the wiki are now possible up to 150MB.

[Norby]

It is a good news, thank you!

Here is a wiki download counter, can you check it please?
http://www.mediawiki.org/wiki/Extension:DownloadCounter

In this way we should use links like download.php?f=path/some.oxz in wiki pages of oxzs and in the oxz manager also to count all downloads.

[maik]

I already checked it, but since its page says "The code or configuration described here poses a major security risk" I'm not going to install it. Instead, I'd rather take it as an opportunity to understand how to build my own extension.

[maik]

Instead, I'd rather take it as an opportunity to understand how to build my own extension.

My download counter extension is now active on the sandbox wiki.

When you upload a file there, it creates an entry for it in its database. Every[*] download is piped through a script that updates the counter. I created a new parser function {{#downloads: filename}} that shows the download count or "unknown" if the file is not in its downloads database. For an example on how to use it see the source code of my user page there.

Let me know if this makes sense and give it a test please. If all is well I will move it to the proper wiki.

[*] You can still bypass the counter update script by accessing the file URL directly, e.g. http://sandbox.maikschulz.de/images/3/34/Q-Bomb_Detector_1.3.m.oxz. In order to prevent this, I would have to move the upload directory outside the web server's document root. I just don't see the point, it's only a download counter after all.[/size]

[Diziet Sma]

You can still bypass the counter update script by accessing the file URL directly, e.g. http://sandbox.maikschulz.de/images/3/34/Q-Bomb_Detector_1.3.m.oxz. In order to prevent this, I would have to move the upload directory outside the web server's document root. I just don't see the point, it's only a download counter after all.

Unless I'm missing something here, isn't this in fact how the 1.79 OXZ Manager itself accesses the OXZs? By using a direct reference to the file URL, as supplied by the manifest.plist? See, for example, the manifest.plist for the QBD:

Code: Select all

{
	identifier = "oolite.oxp.Diziet.Q-Bomb-Detector";
	version = "1.3";
	title = "Q-Bomb-Detector";
	required_oolite_version = "1.79";
	description = "Detects the launch of a Quirium Cascade Mine and sounds a siren to alert the pilot.";
	file_size = 168938;
	category = "Equipment";
	tags = ("quirium","bomb","mine","detector");
	author = "Diziet Sma";
	license = "CC-BY-NC-SA 3.0";
	download_url = "http://wiki.alioth.net/images/e/ee/Q-Bomb_Detector_1.3.oxz"
	information_url = "http://wiki.alioth.net/index.php/Quirium_Cascade_Mine_Detector";
}

[maik]

You can still bypass the counter update script by accessing the file URL directly, e.g. http://sandbox.maikschulz.de/images/3/34/Q-Bomb_Detector_1.3.m.oxz. In order to prevent this, I would have to move the upload directory outside the web server's document root. I just don't see the point, it's only a download counter after all.

Unless I'm missing something here, isn't this in fact how the 1.79 OXZ Manager itself accesses the OXZs? By using a direct reference to the file URL, as supplied by the manifest.plist? See, for example, the manifest.plist for the QBD:

Code: Select all

{
	identifier = "oolite.oxp.Diziet.Q-Bomb-Detector";
	version = "1.3";
	title = "Q-Bomb-Detector";
	required_oolite_version = "1.79";
	description = "Detects the launch of a Quirium Cascade Mine and sounds a siren to alert the pilot.";
	file_size = 168938;
	category = "Equipment";
	tags = ("quirium","bomb","mine","detector");
	author = "Diziet Sma";
	license = "CC-BY-NC-SA 3.0";
	download_url = "http://wiki.alioth.net/images/e/ee/Q-Bomb_Detector_1.3.oxz"
	information_url = "http://wiki.alioth.net/index.php/Quirium_Cascade_Mine_Detector";
}

The URL would need to be e.g. http://sandbox.maikschulz.de/img_auth.php/3/34/Q-Bomb_Detector_1.3.m.oxz. You see it when you hover over the download link created using the [[Media:Q-Bomb_Detector_1.3.oxz]] download link on the wiki. The difference is in using the php script vs. bypassing php.

[Diziet Sma]

Ok.. thanks for the clarification.

[maik]

@cim: could you please verify that this method still works with your OXZ downloader? The test link would be http://sandbox.maikschulz.de/img_auth.p ... _1.3.m.oxz. The php script streams the file as response.

Cheers,
-Maik

[cim]

Yes, that works fine.

[maik]

Cool, thanks!

[Diziet Sma]

The URL would need to be e.g. http://sandbox.maikschulz.de/img_auth.php/3/34/Q-Bomb_Detector_1.3.m.oxz. You see it when you hover over the download link created using the [[Media:Q-Bomb_Detector_1.3.oxz]] download link on the wiki. The difference is in using the php script vs. bypassing php.

Ok.. as I understand it, at the moment this php script is only installed on your sandbox Wiki, not the main or backup wikis, correct?

Also, reading between the lines of your earlier remarks, I get the impression this will only work correctly on OXZ files uploaded after the script gets installed on the Wiki server.. any files uploaded before then would not get the correct URL applied to them.. is this assumption correct?

(I have a 1.4 version ready for testing, when this goes live)

[maik]

Both correct, Diziet Dma.

[Switeck]

Instead, I'd rather take it as an opportunity to understand how to build my own extension.

My download counter extension is now active on the sandbox wiki.

When you upload a file there, it creates an entry for it in its database. Every[*] download is piped through a script that updates the counter. I created a new parser function {{#downloads: filename}} that shows the download count or "unknown" if the file is not in its downloads database. For an example on how to use it see the source code of my user page there.

Let me know if this makes sense and give it a test please. If all is well I will move it to the proper wiki.

[*] You can still bypass the counter update script by accessing the file URL directly, e.g. http://sandbox.maikschulz.de/images/3/34/Q-Bomb_Detector_1.3.m.oxz. In order to prevent this, I would have to move the upload directory outside the web server's document root. I just don't see the point, it's only a download counter after all.[/size]

ok, I've got a brand new OXZ to upload but am at a loss how to upload the file. I am clueless on PHP.