FlashFXP 4.1 RC 2 build 1629

bigstar · 08-12-2011, 08:35 PM

FlashFXP 4.1 RC 2 build 1629

Fixed UI graphic related defects when running FlashFXP on Windows 2000.
Fixed caching defect when moving files/folders on a server while navigating to another cached folder.

FlashFXP 4.1 RC 2 build 1628

The previous build contains a defect that could cause FlashFXP get stuck in an infinite loop. Please update ASAP.

FlashFXP 4.1 RC 2 build 1627

Fixed installer > In some situations the installer would report insufficient free disk space and fail to install, this has been resolved.
Added some additional checking around the FTP command RMD <folder>, some FTP servers do not like absolute paths, now depending on the failure message FlashFXP will try again using the relative path. Such as X2 WS_FTP Server. IpSwitch has confirmed the issue when the "Show Home Folder as root" option is selected for the specific user account. They do not have an ETA for the release of this fix.

Please download the update via LiveUpdate

JoeDirt · 08-13-2011, 04:51 AM

Not sure what the problem is but this last version hangs constantly on my Win7/64

It's actually impossible to use now. Just get those R and then it fades out and lockup.

JoeDirt · 08-13-2011, 07:23 AM

Actually the problem seems to be with my queue file somehow.
I have a long queue and as soon as I try to transfer it this happens (it does not even try to log in on any of the two sites I plan to transfer between) and I noticed the queue was not visible with old version of FlashFXP (3.6).

bigstar · 08-13-2011, 07:27 AM

Thank you for your bug report, Please update to the new release to resolve this issue.

JoeDirt · 08-13-2011, 09:25 AM

Thanx!

Whatever you did, new version solved the problem.

msg7086 · 08-14-2011, 02:40 PM

Hello

Not sure if i should post here, but i just found a small bug.

When moving lots of files, user could have a chance to browse to other directories (more possible if directory panel is displayed). As you know, the moving action would reset the cache status of the working directory. When user switches to some other cached directory and the moving action is still being processed, the cache status of the displayed directory will be cleared. (though I hadn't test if switching directories before moving the first file could cause the cache status of working directory not cleared, but i guess it could happen either)

Not a big problem, but i think i should let you know.

regards
msg7086

msg7086 · 08-14-2011, 02:47 PM

well, just found another bug related with this.

after switching to new directory, the item.delete in the listview / internal array would apply to the switched directory, leaving the working directory un-synchronized with the real world.

say i move file a1, a2, ..., a10, and when it's a3's time i switch to other directory, and when a6 i switch back (before moving process is finished, the cache flag is cleared but i can still switch back and see the cache data), a3~a5 would just stay in the list. you may have a check.

bigstar · 08-14-2011, 06:22 PM

Thank you for your bug report, We have released a new build that should address this issue.

msg7086 · 08-15-2011, 01:28 AM

and another improvement (which could be a core level enhancement)

transfer with a server with high network lantancy (300ms+ ping) is far slower than filezilla.

say i'm in china and the server is located near new york city, and i'll have to suffer a delay of 300ms+ or even more. but both sides have good network, plenty of bandwidth. with filezilla i can reach a speed of 3MB/s+ (or even 8MB/s+) easily, but with flashfxp i can only have a peak of 50KB/s (or even less).

the clue could be the tcpwindowsize, or say, transfer as much data as possible before the next ACK package. ymmv.

bigstar · 08-15-2011, 07:20 AM

Have you tried adjusting the TCP/IP Buffer size setting within FlashFXP?
Globally (located in Options > Preferences > Connection) or Per-site via the Site Manager.

msg7086 · 08-15-2011, 08:32 AM

[shame] how could i miss that.

it can reach up to 4MB/s under 240ms latancy network after i set the buffer size to the maxium value (1024KB per 240ms). but it still need to run the command below to enable the tcp window tuning every some minutes or it will go back to xxKB/s again

Code:

netsh interface tcp set global autotuninglevel=disabled
netsh interface tcp set global autotuninglevel=normal

that should be a bug in the os.

anyway thank you for your reply

Blubi · 08-15-2011, 11:41 AM

1629 crashes sometimes in ws2_32.dll on Live Update when user is pressing on Help and then fast on Check for Updates when the program is still loading.

bigstar · 08-15-2011, 12:54 PM

Quote:

Originally Posted by msg7086

[shame] how could i miss that.

it can reach up to 4MB/s under 240ms latancy network after i set the buffer size to the maxium value (1024KB per 240ms). but it still need to run the command below to enable the tcp window tuning every some minutes or it will go back to xxKB/s again

Code:

netsh interface tcp set global autotuninglevel=disabled
netsh interface tcp set global autotuninglevel=normal

that should be a bug in the os.

anyway thank you for your reply

Did you try values in between default and the max value?

Can you please type netsh interface tcp dump global and show me the result.

The only thing I can think of off hand would be to try
netsh interface tcp set global autotuninglevel=highlyrestricted

If you can think of a public ftp server with high latency I would be more than happy to dig deeper into this issue, perhaps there's more that can be done client side.

bigstar · 08-15-2011, 01:41 PM

Quote:

Originally Posted by Blubi

1629 crashes sometimes in ws2_32.dll on Live Update when user is pressing on Help and then fast on Check for Updates when the program is still loading.

I attempted to reproduce this on Windows XP SP3 but was unsuccessful. The crash dump indicates a buffer stack overflow.

ExceptionAddress: 71ab4f27 (ws2_32!NSPROVIDER::Initialize+0x00000163)
ExceptionCode: c0000409 (Stack buffer overflow)

Perhaps a 3rd party winsock LSP (Layered Socket Provider) is generating the stack buffer overflow? This could be something like a firewall to a vpn client driver, depending on which winsock LSPs are registered on your computer.

For more info: http://en.wikipedia.org/wiki/Layered_Service_Provider

How to determine and to recover from Winsock2 corruption in Windows Server 2003, in Windows XP, and in Windows Vista
http://support.microsoft.com/kb/811259

The last link might not apply, but I found it my research and thought you may want to at least check it out.

Blubi · 08-15-2011, 03:52 PM

I just uninstalled "HMA! Pro VPN", my only proxy software and the same.