[fox88]

Enig123, on 20 April 2015 - 12:19 AM, said:

Can you guys make the miniupnpc lib version 1.8 available? I failed to get that version compile with VS 2013.

Just replace STATICLIB with MINIUPNP_STATICLIB.

[Ergol]

Tried 0.50b BETA 1 fresh install on Windows 7 and Windows 8.1 so few comments as follows:

1. Max. half open connections are set to 9 by default. Believe the limit could be increased for above OS.
2. No icon for eMule in uninstall programs list.

Hope it helps.

[Eulero]

Ergol, on 27 April 2015 - 09:13 AM, said:

1. Max. half open connections are set to 9 by default. Believe the limit could be increased for above OS.

+1
seems that it has been removed starting from Win Vista SP2: https://technet.micr...6(v=ws.10).aspx

http://deepxw.blogsp...onnections.html

This post has been edited by Eulero: 28 April 2015 - 05:34 PM

[DavidXanatos]

Eulero, on 28 April 2015 - 06:30 PM, said:

Ergol, on 27 April 2015 - 09:13 AM, said:

1. Max. half open connections are set to 9 by default. Believe the limit could be increased for above OS.

+1
seems that it has been removed starting from Win Vista SP2: https://technet.micr...6(v=ws.10).aspx

http://deepxw.blogsp...onnections.html

+2

that would be a awesome step forwards, however somehow i doubt it eMule will dear to do it.

This post has been edited by DavidXanatos: 01 May 2015 - 08:35 AM

[xilolee]

Sorry in advance.
I know I shan't post old bugs (or old possible bugs), but I didn't receive an answer (hence I don't know if it is a bug or if it is intended to work that way): Possible Bug? Delayed Connection To A Server When Safe .met/.dat File Writing Is Set To Always
(When that option is enabled, it slows emule startup and emule shutdown)
Hence my question is: could it be fixed?




EDIT: I've now re-downloaded emule0.50b (because of pc problems) and re-enabled Always, but I haven't noticed any delay:
- Did you already solve it?
- Or do I have too less transfers(26) and shares(900)?

This post has been edited by xilolee: 05 May 2015 - 03:27 AM

[hashhunter]

Could you, please, increase maximum of stored KAD-contacts (nodes) in official release?
Either hard-coded or via an option.

Spoiler

Quote

Read 200 contacts from file.

In 0.50a, when we diconnect from KAD, code from RoutingZone grabs and saves into file only 200 of alive contacts, taken from the "tree of nodes".
Actually, it is *still* possible to find 500 up to 700+ alive nodes,
and there will remain just about a dozen or two of the "dead" ones (tho just disconnected and/or outdated).
There is no need to do a thing about mechanism of routingtable itself, if only expand its size also (?!).

Remark about *still* : looking at situation around public bittorent trackers, one would want to touch/knock wood.

minor bugs:
Kad - when you double-click on a name in "Current Searches", "Search results" (chart) has broken encoding of the name in its title.
"Transfers > On Queue" doesn't seem to work. Some clients/mods have the list updated.

Never split Search results in categories based on "quality", as in some mods (rather looking as someones' odd preferences).

Optional, about the same Search...

Spoiler

When "View shared files" is done on a client, the list in search tab may show the relative path to file, e.g. artist_name\album_name\file .
It is known that in Transfers it's possible to add "categories" (which will become subfolders in Incoming).
Back in search, there are two options in context menu of a result(s): Download and Download (paused).
Is it possible to add some option to make a category, which name would be based on seeder's original folder name , or free-to-manually-edit?
(Otherwise we need to catch paused downloads in "transfers", select and move in categories. Sometimes it becomes too messy.)

This post has been edited by hashhunter: 08 May 2015 - 04:18 PM

[fox88]

Recently for the first time ever I saw messagebox: "An attempt was made to access an unnamed file past its end".
In the Docs Error messages the following explanation could be found:

This message pop up is caused by invalid packets from servers or other clients. This message does not affect eMule's operation and is properly handled in the versions .28a+

If this (or any other) event is known, and handled correctly, and does not require user actions, is it necessary to show message box?
A record in the log - ordinary or even verbose - might be more appropriate; and currently the event is not in the logs at all.

[Some Support]

fox88, on 22 May 2015 - 11:18 PM, said:

Recently for the first time ever I saw messagebox: "An attempt was made to access an unnamed file past its end".
In the Docs Error messages the following explanation could be found:

This message pop up is caused by invalid packets from servers or other clients. This message does not affect eMule's operation and is properly handled in the versions .28a+

If this (or any other) event is known, and handled correctly, and does not require user actions, is it necessary to show message box?
A record in the log - ordinary or even verbose - might be more appropriate; and currently the event is not in the logs at all.

I don't think it was eMule's code creating the messagebox and I think this bug should have been fixed since a decade or so. Did you use a debug build? Overlapped sockets? Other changes you made?

[fox88]

Some Support, on 23 May 2015 - 03:13 AM, said:

I don't think it was eMule's code creating the messagebox and I think this bug should have been fixed since a decade or so.

So this should never happen in eMule now?
Unicode string "An attempt was made to access %1 past its end." could be found in mfc*.dll files in System32.

Some Support, on 23 May 2015 - 03:13 AM, said:

Did you use a debug build? Overlapped sockets? Other changes you made?

I am using debug builds for a long time; overlapped sockets are on.
I will check if my changes could introduce such bugs.

[fox88]

Here is MS article: Targeting your application for Windows.
https://msdn.microso...v=vs.85%29.aspx

Besides reporting incorrect OS version, certain API functions might also return false data.
Fortunately, the fix is simple.

Below is contents of manifest file emule.manifest, that should be created and included into project options.

Spoiler

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">

  <compatibility xmlns="urn:schemas-microsoft-com:compatibility.v1">
    <application>
      <!-- Windows 10 --> 
      <supportedOS Id="{8e0f7a12-bfb3-4fe8-b9a5-48fd50a15a9a}"/>
      <!-- Windows 8.1 -->
      <supportedOS Id="{1f676c76-80e1-4239-95bb-83d0f6d0da78}"/>
      <!-- Windows 8 -->
      <supportedOS Id="{4a2f28e3-53b9-4441-ba9c-d69d4a4a6e38}"/>
      <!-- Windows 7 -->
      <supportedOS Id="{35138b9a-5d96-4fbd-8e2d-a2440225f93a}"/>
      <!-- Windows Vista -->
      <supportedOS Id="{e2011457-1546-43c5-a5fe-008deee3d3f0}"/> 
    </application>
  </compatibility>

  <assemblyIdentity type="win32"
                    name="eMule"
                    version="0.50.1.16"
                    processorArchitecture="x86"
                    publicKeyToken="0000000000000000"
  />
</assembly>

Then following code might be added to eMule, and that code will work correctly:

#define _WINVER_8_1_   0x0603 // 6.3
#define _WINVER_10_    0x0a00 // 10.0

if (osvi.dwMajorVersion == 6 && osvi.dwMinorVersion == 3)
	return _WINVER_8_1_;
if (osvi.dwMajorVersion == 10 && osvi.dwMinorVersion == 0)
	return _WINVER_10_;

[Some Support]

fox88, on 10 June 2015 - 08:35 PM, said:

Here is MS article: Targeting your application for Windows.
https://msdn.microso...v=vs.85%29.aspx

Besides reporting incorrect OS version, certain API functions might also return false data.
Fortunately, the fix is simple.

eMule's code itself is not Win8 / Win10 aware, meaning it makes not special decisions for those OSs. Declaring that they are targeted might have unintentional side effects - as you pointed out, certain API functions return different results based on whether the App is aware of the new OS or not, while the result will always be backwards compatible when not targeting the Win8/10.
So as long as eMule has no reasons for targeting those, it doesn't makes sense to do so, because in this case we would have to add Win8/10 aware handling of any changed API functions (which might not be a bad thing but costs time with little gain)

[fox88]

Some Support, on 11 June 2015 - 02:04 AM, said:

eMule's code itself is not Win8 / Win10 aware, meaning it makes not special decisions for those OSs.

It is possible to use new API functions without manifest; but here manifest is needed for backward compatibility.
Example with DwmIsCompositionEnabled function was not random; that API call is used in Preferences.cpp.
According to description, the function should return false starting from Win 8.1 for old applications which were not properly "targeted".

Some Support, on 11 June 2015 - 02:04 AM, said:

Declaring that they are targeted might have unintentional side effects - as you pointed out, certain API functions return different results based on whether the App is aware of the new OS or not, while the result will always be backwards compatible when not targeting the Win8/10.

Until somebody decides to rewrite that part of code (possibly using SystemParametersInfo function as suggested by MS), manifest should be good enough.

This post has been edited by fox88: 18 July 2015 - 10:38 PM

[awdrifter]

I'm glad there's finally a new version, I've been using it for a few days. It's pretty good. Hopefully there will be an IPv6 compatible version soon.

[Mymlok]

Hi!
The year is 2015 and further eMule has a minimum upload limit of 10 kB / s after which the download is already without restrictions. I propose to raise this limit to 100kB / s and even 200kB / s.

[Zangune]

This wouldn't be a good idea, in example here in Italy it is very very common to have a bandwidth of about 35 kB/s.

[fox88]

I think eMule could be overoptimistic in setting slot speed for a new downloader when there is a lot of unused bandwidth.
It might be seen qutie often when eMUle is just launched, and first clients appear in upload list.
Commonly upload speed in the list could be about 50KB/s or more, and several hundreds of KB swiftly go into uploaded amount.
Then speed goes down, and amount stays constant (sometimes another bunch of data goes up later).
Finally either speed settles at the real capability of downloader, or connection times out and disconnects.
Simple calculation: if downloader can accept 3KB/s, then initially given (for example) 120KB will be received in 40 seconds, which is close to timeout.

[startsteel]

Good software

[arturx]

But it takes a lot for the release of the final version?

[Some Support]

Indeed. Although aside from the minaor smaller things mentioned in here there doesn't seem to be any major problem with the beta.

[fox88]

Should any moves forward be expected in the near future?