Crise

With 1.5.5 ApexDC officially moved to using the Plugin API currently in DC++. This essentially does not change much since their plugin API is a cousin to what ApexDC (1.5.x) has had for quite sometime now. One of our personal goals with this was to move the ScriptPlugin (formerly LuaPlugin or Lua Plugin, but hey rose by any other name and all that) to be managed by DC++ team and the people there who actually know more about Lua than yours truly. Unfortunately though, for us, DC++ doesn't provide pre-compiled 64bit binaries at all which means that we will still be providing a download of ScriptPlugin for you for the time being. For end users the change of a vendor for this plugin means two things. There will no longer be LuaJIT flavored builds on offer, which for anyone who ever used any pre-built modules is non-issue, since you wouldn't be using this build variant anyways. However, what does impact end user is that this new flavor of ScriptPlugin uses Lua 5.2 in favor of 5.1. For script migration please see the following links: Language changes in Lua 5.2Bootstrap script (startup.lua) and default scriptsIn comparison, although the Plugin API as it is in DC++ is still not 100% stable (as in stable API in terms of compatibility), it is significantly more likely to remain compatible for longer periods of time since it will no longer see rewrites from version to version (which before submission to DC++ was more of a rule than an exception for me).

I will look into this.

It will be distributed separately due to vendor changing from ApexDC++ to DC++. I'd expect the release announcement to detail this more or less.

If the server does not send a mime type (Apex)DC++ falls back to assuming compressed hublist most likely. Also for what it is worth when working with serverside scripting, and DC++'s limited http support, you might need to explicitly make sure that the mime type and Content-Length headers are sent (also effectively disable chunked transfers, because DC++ hasn't supported them for that long). Also, make sure that serverside gzip compression is not applied to the plain xml files obviously, and check that the files are utf-8 encoded, or compatible, because dcpp's XML classes take some liberties when it comes to encoding.

Sure I can arrange snow here, assuming IPB's skin controls are not horribly broken for me as usual, it is just that I figured it would be better not to... because javascript snow tends to be a reasonably demanding in terms of CPU (which does tell something of the average javascript/rendering performance of browsers). Edit: snow has arrived, tell me when ypu have had enough :P

This is a protocol level limitation. So to answer your question you will have to refine your search. The limits are 10 results per active 5 per passive user.

To make a statement on this, the end point of the requested chunk could in theory be used to detect segmented uploads but to my knowledge that is not done, regardless of settings about segmented downloads. What you ask is not feasible or healthy for the network in my opinion and would also result in impairing your download experience in the end. If you have a specific scenario where all manner of segmented transfers is undesirable for everyone then for those scenarios it could be worth to consider but for now the good of the many outweighs the benefit of the few (tip: if you worry, do not share from a media that is prone to be worn by extreme use, such as SSD, even though for regular use with a good quality SSD it is still a non-issue regardless in my opinion) As for blocking users, here is a link to a statement on that: http://forums.apexdc...ter/#entry30487

The problem with campus networks tends to be with port forwarding... just changing your settings to active mode does not mean you are really in active mode. It just means that the client operates as if direct connections between users were possible, for them to be possible in this scenario ports need to be forwarded. Your latest graph basically outlines a typical scenario where you have bunch of users connecting to a hub in the internet, behind different public facing routers which is (trivially) true for almost any two users from different geographical locations or ISP's. When DC client connects to a hub in the internet the fact that two users in said hub are from the same network does not make any difference to the client, because the client does not know this, as all users typically send their public facing IP to the hub... it attempts connecting using the known public IP for that user, so port forwarding is necessary for the connection to succeed. The only difference to that typical scenario you have yet to state is that with campus networks you typically can not freely forward ports (although same can be true for home networks, f.ex. my ISP blocks most ports < 1000 for incoming connections), so yes if you have more than one user being forced to share the same public IP and port combination because of the fact that you can not forward ports it can't physically ever work because the port and IP the client tries to connect to are the same for two different users.

Lua Plugin closely matches the implementation used by BCDC, only existing documentation for it is in the "startup.lua" file and the few example scripts.

I know, we used to as well... but we decided it is much better to advocate the use of a third party applications designed to filter network traffic globally rather than providing an application specific filter. Most firewalls even allow you to set rules specific to applications, and there are several software solutions available that can do network filtering on different levels. We do not support the idea of blocking transfers from an individual DC user, we believe that if someone really has earned a network block on his IP, it would only make sense for that block to be system wide or at the very least unbiased (ie. not only block yourself from uploading to a user but still being able to download from them). The valid use for a network filter is when you are paying for bandwidth outside your LAN for example, and we wholeheartedly believe there are much better solutions for that already out there that can be set up to better prevent unwanted outbound traffic than an application specific tool for just ApexDC could.

Unfortunately as far as I am aware apple has not updated the interface we use since 09, so if that is broken and there is no new update then there is little we can do (might be possible to find an alternate method though), I can have a look but no promises at this point.

In your case the settings are not saved on the portable device, if you used the installer to set up your copy it will (by default) save all settings tied to the windows user account. Use the slim binary download and you should be fine.

Like I replied elsewhere, this most likely has to do with an issue in 3rd party software called 'PC Tools', please try disabling or uninstalling said software and see if the issue persists. Other than that there is no advice to be given at this time.

I am not completely sure what it is that you are asking here, as to why the temporary file is not moved/renamed when file is downloading that is because while file is downloading the temporary file will be locked (ie. it can not be moved), and once the download compeletes the temp file is no more, so no action is needed.

Unless ApexDC is run as administrator it may not have permission to edit registry (Windows Vista and newer), which would explain the "annoying dialog", despite the option being off. Also, as for magnet links in general, see http://en.wikipedia.org/wiki/Magnet_URI_scheme In other words, magnet links are generic kind of link, what program can use them depends entirely on the hash that is contained. However, since there are several notations of hashes supported by DC (ie. TTH) we can't limit it to one specific kind of magnet link so easily. Also, it should be noted that if you click a magnet link in any hub chat, then ApexDC will always attempt to handle the link regardless of any settings. The options on the dialog depend on whether the magnet link contains a name or not, ie. if we only have a hash all we can do is search, but if we have name hash and size we can directly add the file to queue (of course that does not guarantee availability of any sources). Edit: just looked it up and there is a rather odd requirement for the link to contain a filename before ApexDC does anything with the link automatically. I will look into this better and see if there is an actual basis for this (obviously we can't add nameless files to queue but for searches there should be no reason to have this requirement).

.crt is the public certificate, and .key is the private key.

ApexDC does not associate itself with ed2k links, per se... The subtle difference is that ApexDC associates itself with Magnet links (a generic URI, typically containing file hashes). ApexDC accepts any Magnet URI containing a hash it can use, currently this includes Magnets with Tiger Tree hash (or a bitprint, which is a combination of a tiger tree and sha1 hash, typically used and generated by programs for the Gnutella or Gnutella 2 P2P networks). An application with support for ed2k network can generate Magnet links, but they don't typically contain a TTH, so are thus useless for ApexDC. The more traditional link type for ed2k is using the pseudo protocol identifier ed2k:// which ApexDC obviously does not even attempt to handle.

For the record ApexDC does save settings at certain intervals, I recall it is set to 10 minutes (can be changed by editing the xml file manually). Common issue with saving settings is that ApexDC may be installed in a directory were it does not have write permission by default (eg. under Program Files in recent versions of windows). In such cases moving ApexDC to a non-system partition is preferable, alternatively you may (using the installer) opt to store settings within your windows user account also resolving the problem (however, if multiple installations for the same user account is something that is needed, this solution will not work). If you are having stability issues in regards to hashing, please turn off "fast hashing" in the program settings (under Advanced). Regarding default favorite hubs: in the next major version (2.0) we will introduce a way to opt out of ever connecting to these hubs on first run, however, for 1.x (as well as existing installations) the behavior is unlikely to change. The only way to opt out from these hubs being added with current versions is to add a Favorites file with at least one valid favorite hub in it before application is launched for the first time. Regarding opt-in vs opt-out, the reason for it not being opt-in is because it is intended as a simple demonstration of the concept of hubs (as well as some of the less obvious new features related to them) for new users (unfortunately there is currently no way to tailor this experience based on language or other user data, 2.0 will in due time bring a much changed and improved "new user experience"), if you are already aware of what a hub is and how to find and use one, then the assumption that you already have at least one hub you frequent or know of is valid. Edit: also there is an fairly undocumented keyboard shortcut to suppress most automated actions at program start up, simply hold down shift, this includes the opening of settings dialog on fresh install as well as any automated hub connections.

I did a very basic check, and it works fine here...

Below is some basic usage instructions for the SnCMS, written back in 2007

While back I went through my "development directory" and found some pieces of work that could potentially be useful for other people, if shared, but were either used only by us internally or just to collect dust. So I put in some effort and packaged a few different files for public consumption, if you find them useful then great... if not then it is no-ones loss. I would ask you to comment below if you find something interesting in the stuff listed below, but seeing as our previous blogs haven't been big on comments (which is why there haven't been any more), who am I kidding. Check the full post, after the break for more details. 1. Inno Setup Scripts (partial code dump) This file contains a collection of auxiliary scripts used by our installer, please see the enclosed readme, and the script files themself, for further details. 2. PHP Utilities (Tiger and TTH hashing, Base32 encoding) This file contains two simple classes, again the enclosed readme will shed more light on each (also explains why they are even here, since the original unmodified version of most of the code can be found on google). "This code is namely shared because the current PHP based DC software that make use of the ADC protocol generally uses hard coded hashes, or off loads the task of hashing to a binary dependency, for login. While TTH is not used for that task directly the tiger and tigerfix methods, and the documentation surrounding them, will hopefully be of use to some and lessen the number solutions reliant on static hashes." 3. phpBB based CMS system (aka. SnCMS, as in tin dioxide) This was originally built for something we were planning for here at ApexDC then it was used for something completely different in the end, and after that it was used on the now late ADCPortal. Older version of this is still kind of in use, internally, but it only exists as a spare (though it has been particularly useful in testing the recent server moves). I originally made it based on a concept I had long time ago, that I used on an old AirDC++ web site (yes, way back when it was still running phpBB2, I was somewhat involved in the web side of things there), now with phpBB3 as the backend. I have bundled two styles with it, the original ADCPortal template files and a template'd version of the old layout for the system I made for the old AirDC++ site with some modernizations. I didn't release this before because I couldn't be bothered to create a style that didn't have a branding on it... but seeing as ADCPortal is dead, superseded by www.dcbase.org somewhat, and AirDC++ uses a different style now there is no longer a reason not to release it with the styles unedited. Notably this project in its entierty is licensed under GNU GPL v2, as any phpBB modification should be (even though it doesn't modify phpBB per se). This old topic is also related (as you can see the plans to release it have existed for a while). No real installation instructions, read the comments in config.php, and figure it out from there (if someone actually seriously wants to use this, reply below and I'll give better instructions). Downloads Inno setup scriptsPHP UtilitiesSnCMS - phpBB CMS

Each forum in a phpBB installation has a numeric unique id, same for topics. Start by creating f.ex. a "Site Pages" forum and create topics titled simply Menu and Home in it then set the id's of those topics as instructed in config.php, same goes for forums you want the front page news to fetched from if any. In a deployment environment this forum is naturally only editable by select few, and usually hidden, since every page in it represents one page on the site. Pages would be accessed by either: www.example.com/?id=pagetitle or www.example.com/?id=<topic id> (this is why I prefer to keep topic titles there as one word) Edit: I may have some better instructions written up somewhere, I'll see if I can find them.

First guess, too long path... second guess, the application does not have access to said folder for whatever reason.

Event ID 4226 only applies to SP2 as far as I am aware, not SP3. In any case it is not practical for me to keep anything below XP SP3 supported, because the best I can get in terns of XP these days is the XP mode that ships with windows 7. Besides Microsoft will drop all support for XP in 2014, and it is already on extended support anyways. It would be increasingly troublesome to keep supporting operating systems that have been abandoned in terms of security patches and hotfixes. Because each additional operating system supported adds its own limitations to what can and can't be done with the software, within reasonable effort. Newer is indeed not always better, however, old will always be old and will only get older... that is a fact.

There is a setting that controls the maximum number of simultaneous downloads on the downloaders side (more than one of them in fact) same page where you set your download directory, check the values there.