Custom Query (2164 matches)

For integration purposes, I would request that a command line parameter be added to specify if a connection is to be active or passive. In this design, if blank, the default settings for the app (Connection/FTP/Transfer Mode) would be used.

This connection string parameter would only be valid in the case that an FTP URL is passed as a parameter.

Thanks for your consideration!

When the local default directory was set to a directory on a non-existent, removable drive, the REMOTE default directory would fail with an error (in red) displaying "no such file or directory". This even failed with a copy and paste from the "Remote Site:" pulldown. No failure was noted for the local directory error. This also occurred in 3.3.4.1 so I immediately upgraded to 3.5.3 with the SAME results.

After modifying the local default to an existing directory, the remote default now works. Switching the local default back to the non-existent directory, to test this again, NOW the problem does NOT occur. (It appears that once the remote directory is accessed as a default, it appears to stay working.)

This occurred 12-15 times as I attempted to figure out what error I was making in the remote directory (It had to be something because there was an error each time. I slowly reduced the number iof subdirectory levels and finally attempted "/" which gave a failure as well. Each attempt was a "Connect" and "Abort" via the Site Manager.

I thought that I had done something wrong on the remote default, somehow, but when I PASTED the result and it still failed, I was perplexed. Yes, I checked for trailing spaces and even added a trailing "/" and removed it, yielding the SAME results. I thought it was a problem with symbolic references to a directory, but only after the LOCAL default directory was made to work did the remote directory work.

All I can say is that it failed several times, but now I know a workaround.

Therefore, I have lowered the priority of this bug, but created the ticket because it HAD happened. There is an easy workaround -- make sure the local directory works properly.

Even if nothing is changed, if someone else has the problem, they might find this workaround -- make sure the local directory exists.

I cannot rename or delete a file that ends in an asterisk (*). Also, the asterisk does not show up on the file listing.

This can be reproduced with the following:

1. log into FTP server
2. Right click test file and chose rename.
3. Add an asterisk to the end of the file name (in my case, test.txt became test.txt*)
4. Important! Refresh the file listing.
5. Notice that the asterisk is no longer in the file listing. (test.txt* now looks like test.txt)
6. Try to rename or delete test.txt

You will get the following error:

Command: DELE test.txt Response: 550 test.txt: A file or directory in the path name does not exist.

If I turn on raw directory listing in debug setting I see the following:

Command: LIST Response: 150 Opening data connection for /bin/ls. Response: 226 Transfer complete. Listing: total 8 Listing: -rw-r----- 1 dlech mis 2346 Feb 03 09:30 test.txt* Status: Directory listing successful

Notice the asterisk is at the end in the debug listing but it's missing in the interface listing.

If I execute the following custom command it will delete: DELE test.txt*

Command: DELE test.txt* Response: 250 DELE command successful.

The server OS in this case is AIX using standard FTP but I can repeat the problem on a linux system using SFTP except it rejects my custom commands when I try to delete manually.

We have an old system that creates temp files then end in asterisk that sometimes need to be cleared out. That is how I discovered the problem. Thank you.

It would be nice if the QuickConnect button would copy the information it uses into the Host/Username/Port boxes.

That way if, for eg, my password changes, or I want to use the same credentials for a new host, I don't need to retype them all - only the bits that are different.

I would like to have you add support for the Unisys OS2200 series mainframes. This OS does not use a hierarchical file structure. The only problem at the moment is that the Site Manager/Advanced tab/remote directory does not accept an OS 2200 file name.

The following is from the Unisys documentation on naming files and elements. Please contact me if you need more information.

3.3. Naming Files and Elements The only part of a file name that is required is the basic file name, as follows:

filename.

All other fields are optional. In some cases, the period is optional; if you do not include a period, the system assumes that it is the name of a file. In other cases, if you do not include a period, the system assumes that it is the name of an element. To avoid confusion, include the period except for file names specified on the @BRKPT statement. The name originally given to a file is called the external file name. A file can also have an internal file name. See Section 7.68 for information on how to use the @USE statement to give a file an internal name. Using Files and Elements

File Name and Element Name Format

[directory-id#qualifier*]filename[(F-cycle)/read-key/write-key.element-name/ version(S-cycle)]

directory-id

Shared files exist exclusively on an OS 2200 mass storage device that is shared by two or more OS 2200 systems. Shared files are different from local files, which are files that are stored on a specific OS 2200 system. The directory-id field is valid when file sharing is configured on the system. It indicates which master file directory contains the file. The directory-ID has two possible values: STD (standard) for local files on any system and SHARED for shared files. The # is a delimiter to separate the directory-ID and qualifier. This field is valid only for the following statements:

@ASG @CAT @FREE @MODE @QUAL @USE

All other statements that have a filename field cannot include the directory-ID in the file name. You can indirectly use a directory-ID for file names on these statements by either giving them internal names with the @USE statement or by specifying directory-IDs with the @QUAL statement, as in the following examples:

@USE P1,STD#QUAL1*FILEX @COPY P1.,X. @QUAL,D SHARED#QUAL1 @COPY FILE1.,FILE2.

In the second example, the default qualifier (QUAL1) and directory-ID (SHARED) that were specified on the @QUAL,D statement are assumed in the file names on the @COPY statement.

For more information on the @USE statement, see Section 7.68. For more information on the @QUAL statement, see Section 7.49. If you omit the directory-ID and # from a file name and do not specify a directory-ID on a previous @QUAL statement, the system uses the director-ID value configured at your site. One exception to this is if a user is cataloguing a shared removable pack file, the system detects it is a shared pack and uses SHARED. If you include the # but not a directory-ID, the system uses the implied directory-ID you specified on a previous @QUAL statement. If you did not specify an implied directory-ID, the system uses the directory-ID value configured at your site. Using Files and Elements

The default directory-ID and default qualifier are the directory-ID and qualifier given on the last @QUAL,D statement performed in a run. The implied directory-ID and implied qualifier are the directory-ID and qualifier given on the last @QUAL with no options performed in a run.

qualifier

The qualifier helps to ensure that file names are unique. Different files can have the same file name, as long as their qualifiers or F-cycles are different. User-IDs and project-IDs are commonly used as qualifiers, but you can create your own. The qualifier can be up to 12 characters long and can include any combination of letters, numbers, hyphens (-), and dollar signs ($). (The @QUAL statement, which also defines the qualifier, is described in Section 7.48.) If you omit the qualifier and asterisk (*) from a file name and do not specify a qualifier on an @QUAL statement, the system uses the project-ID on your @RUN statement as the qualifier. If you include an asterisk (*) but not a qualifier, the system uses the implied qualifier you specified on an @QUAL statement. If you did not specify an implied qualifier, the system uses the project-id on your @RUN statement.

filename

The basic name of a file can be up to 12 characters long and can include any combination of letters, numbers, hyphens (-), and dollar signs ($). Note: Do not use a processor name (ELT or DATA, for example) as a file name. The results will be unpredictable.

F-cycle

The F-cycle number identifies a file in a set of files that have the same qualifier and file name. Normally, you never have to include an F-cycle number with a file name. If you do not supply the F-cycle, you get the highest cataloged file cycle in the set. See Section 3.4 for more information on F-cycles.

read-key/write-key

Keys are like passwords that are required to read a file or write to it. Keys typically apply only to unowned files which are standard in Fundamental Security, but they can be used for owned files if the system is so configured. Keys are an easy way you can protect a file from unauthorized use. When you create a file, you can specify a read key, write key, both, or neither. If you use only a write key, you must precede it with two slashes to indicate that it is a write key, not a read key. Each key can be up to six characters long. You can use any characters but the period (.), comma (,), semicolon (;), slash (/), or blank. To change the read or write key on a file, or to add a read or write key to an existing file, use the @CHG statement as follows: @CHG filename/RK/WK,filename/RK/WK

Using Files and Elements

where filename1 is the file to be changed (including the current read/write keys), and filename2 is the file with its new or changed read or write keys. See Section 7.5 for more information on the @CHG statement.

element-name

Is the name of the element. Element names can be up to 12 characters long and can include any combination of letters, numbers, hyphens (-), and dollar signs ($). version Is used to distinguish between elements that have the same name and type, but contain different information. Using version names is one way you can save different renditions of the same program. Version names can be up to 12 characters long and can include any combination of letters, numbers, hyphens (-), and dollar signs ($). If the element is an omnibus type, some processors can ignore the version. A combination of element name and version need be unique only within a file. Thus, .CH2/VER1 and .CH2/VER2 are two different elements (note the leading period). Then with different files, FUR1*PUR.CH2/VER1, FUR1*PUR.CH2/VER2, FUR2*PUR.CH2/VER1, and FUR2*PUR.CH2/VER2 are four different elements.

S-cycle

Is a number identifying a cycle of a symbolic element. Element cycles are different stages a symbolic element goes through as it is updated by certain processors (U-option). Normally, you do not need to be concerned about element cycling and S-cycle numbers. Some processors, including IPF 1100, do not use element cycling. Processors that accept symbolic element input always default to the most recent cycle of a symbolic element. See Section 3.5 for more information. Only the basic element name is required; the version and cycle numbers are optional. In fact, you cannot use cycle numbers for elements other than symbolic elements. Normally, different elements within the same file have different names. However, elements can have the same name, provided their element types are different (symbolic, relocatable, executable, or omnibus).

File Name Examples

ACCT-MASTER. ACCT-MASTER is the basic file name. The system uses defaults for all other parts of the file name: · It uses the project-ID from the @RUN statement as the qualifier. · It uses the current cycle of the file. · No read key or write key is used.

GL*ACCT-MASTER. GL is the qualifier, and ACCT-MASTER is the basic file name. Using Files and Elements

GL*ACCT-MASTER/SESAME. GL is the qualifier, ACCT-MASTER is the basic file name, and SESAME is the read key.

PR*ACCT-MASTERPOPPY. PR is the qualifier, ACCT-MASTER is the basic file name, and POPPY is the write key. (The two slashes indicate that there is no read key and that POPPY is a write key, not a read key.) The current cycle of the file is used.

GL*ACCT-BAL/SESAME/#%$&!. GL is the qualifier, ACCT-BAL is the basic file name, SESAME is the read key, and #%$&! is the write key.

SHARED#DATA*BASE. SHARED is the directory-id, DATA is the qualifier, and BASE is the basic file name.

Element Name Examples

COLORS.RED,.BLUE,.YELLOW All three elements (RED, BLUE, and YELLOW) are assumed to be in file COLORS.

COLORS.RED,.BLUE,YELLOW The elements RED and BLUE are assumed to be in file COLORS, while element YELLOW is assumed to be in the TPF$ file.

MAIN In this example, only the basic element name is used. Since a cycle number is not specified, the current cycle of the element is used. The version name is blank.

SUB1/APR02

Here, SUB1 is the basic element name, and APR02 identifies the version of the element to be used. Since a cycle number is not specified, the current cycle of element SUB1/APR02 is used.

When you specify an element name without a file name, the system usually assumes that the element specified is in TPF$, a temporary file automatically assigned to your run. An exception occurs when you specify a series of elements on the same control statement. If you include a file name with the first element name and put a period before the subsequent element names, the system assumes that those elements are in the first file named.

Prcedure: In the left pannel right click a file choose OPEN

Expected Results: FileZilla will pass info to the operating system so that the file is opened using the default application for file type.

Acutal results: the file is opened with a different application

Example: Choose a plain text file which I have set at Notepad++ as the default, it is opened with Notepad

Hi, Thanks for the good Job. FileZilla is certainly the REFERENCE. Small bug I noticed in vers 3.5.3. After updating local version of ROBOTS.TXT saved on local disk, (update editor called from FileZilla and saved locally), FileZilla will display message explaining that (local File was updated, will FileZilla delete remote one and upload the new one? I confirm YES, FileZilla reconnects automatically and uploaded new version = OK)

Problem= When I click on the plus sign + to expand the root folder and display the subfolders, it dosn't work. I had to disconnect, reconnect again, then it works.

Thanks for you support. Best Regards TED

Don't know about other platforms, but I notice that on Mac OS X (10.6.8), FileZilla (3.5.3) does not establish itself as the default handler for ftp and sftp links in web browsers. Other FTP clients do this, so it would be nice to have. Also, I notice that when I use the third-party RCDefaultApp preference pane to force an association of ftp and sftp links with FileZilla, FileZilla launches but does not process the url.

I am running a server with the following config

PHP Built On Linux cgmsystems.com 2.6.9-89.35.1.ELsmp #1 SMP Tue Jan 4 22:30:58 EST 2011 i686 Database Version 5.1.52 Database Collation utf8_general_ci PHP Version 5.2.5 Web Server Apache/2.0.52 (Red Hat) WebServer to PHP Interface apache2handler Joomla! Version Joomla! 2.5.1 Stable [ Ember ] 2-Feb-2012 14:00 GMT Joomla! Platform Version Joomla Platform 11.4.0 Stable [ Brian Kernighan ] 03-Jan-2012 00:00 GMT

I am using the following filezilla

FileZilla Client

Version: 3.5.3

Build information:

Compiled for: i586-pc-mingw32msvc Compiled on: x86_64-unknown-linux-gnu Build date: 2012-01-08 Compiled with: i586-mingw32msvc-gcc (GCC) 4.2.1-sjlj (mingw32-2) Compiler flags: -g -O2 -Wall -g -fexceptions

Linked against:

wxWidgets: 2.8.12 GnuTLS: 2.10.4

Operating system:

Name: Windows NT 6.1 (build 7601, Service Pack 1) Version: 6.1 Platform: 64 bit system

under settings, file lists, I have double click action on files - none double click action on directories- enter direcory

I was working on a server folder that had alias folders. When I double clicked on one of these folders, filezilla did not enter the folder but rather downloaded the files in the alias folder.

the os of the server is linux, the os of my desktop is windows 7 64 bit.

OS : Win XP-SP3 FileZilla : 3.5.3 Protocol : sFTP

Uploading large files (1GB) using high-speeds networks, (speeds in excess if 8MB/Sec) fail with below error message

Error: File transfer failed after transferring 762,900,480 bytes in 226 seconds

Applying FileZilla Speed-Limit clipping speed to 3000 KiB/Sec allows to transfer file successfully. Using FileZilla transfer speeds from 4000Kib/Sec --> "unlimited" speed option causes transfers to fail.

We have tested transfering same file(s) using HP-UX sftp client and these upload flawless, speeds seen 6-7MB/Sec.

OpenSSH_4.4p1-hpn12v11, OpenSSL 0.9.7l 28 Sep 2006 HP-UX Secure Shell-A.04.40.011, HP-UX Secure Shell version

On version 3.5 the conenctivity to our MOVEit DMZ server with Filezilla was fine. Now after 3.5.1 and 3.5.3 we get a problem, if i copy a folder from MOVEit to my PC then delelte it from the MOVEit Server filezilla errors. I have a work around which I must enable keep alives and reduce the maximum simultaneous connections to 1.

In the past, keying <F5> to refresh the display after the connection had timed out used to be instant. For the last few versions of FZ, however, it takes at least 30 seconds.

Because the time-out period is so short, refreshing to reconnect so often is becoming very tedious.

Can you make the refresh period instant again, please?

The SFTP permission support is a bit weird. It shows owner/group, setgid/setuid bits but FileZilla doesn't support changing or preserving these settings on a file or directory.

For example chmod 2755 or 4755 or 6755 on a folder/file. Uploading a file with the set(u/g)id permission bit set is not preserved. Or changing the Owner/Group/Other permissions (777-000) will also remove these bits forcing you to launch an external program to set those bits manually.

A switch should be used for SFTP based connections to support Owner/Group changes, as well as the 4th octal permission bit, since FileZilla executes the chmod command it should accept all 4 octal bits, and also support chown/chgrp

Command: chmod x750 "test" Error: chmod: file mode 'x750' contains unrecognised user/group/other specifier 'x'

FileZilla Client
----------------

Version:          3.5.3

Build information:
  Compiled for:   i586-pc-mingw32msvc
  Compiled on:    x86_64-unknown-linux-gnu
  Build date:     2012-01-08
  Compiled with:  i586-mingw32msvc-gcc (GCC) 4.2.1-sjlj (mingw32-2)
  Compiler flags: -g -O2 -Wall -g -fexceptions

Linked against:
  wxWidgets:      2.8.12
  GnuTLS:         2.10.4

Operating system:
  Name:           Windows NT 6.1 (build 7601, Service Pack 1)
  Version:        6.1
  Platform:       64 bit system

When the number of bookmarks is large, it is difficult to navigate through a single large bookmark list. (One needs to scroll down the list)

An option for grouping them in a folder structure would very useful to be able to organise the list. (Similar to web browsers like IE or FF)

I think, even a folder structure of a finite number of levels of depth (eg:2-3) would be fine for most users.

When the list of saved addresses gets big it would be useful to have ability to filter them by starting typing a word.

eg. typing 'exa' will display only addresses having 'exa' in name, eg. ftp.example.com

Many programs is getting search oriented (e.g. Windows 7 menu, Unity, new Ubuntu will even have menu with search ability).

1. Click on a filename to allow renaming
2. Select the filename (or a part thereof)
3. Right-click

Expected result: context menu appears with options to copy/paste/etc. (As in Finder)

Actual result: context menu does not appear; selection is cancelled. Cursor moves to location of click.

I'm running OS X 10.6.8; not sure if this functionality is available in other OSes.

hi,

I want to point out some things that are not quite right:

1. the quick connect history should not record an entry when password or username are wrong (as is the case for virtually all my failed connects) - there is no need to store a record of a mis-typed password.

2. it should also arrange its entries by most used rather than chronologically, or at least have that option available

3. it should also list more than ten previous connections

when I work, on many sites, sometimes concurrently, but often delving back into old projects for updates, the lack of a good history of connects can be awkward in a program I use every day. I would love for the connection process to be simplified :-)

Colin

In Concurrent Transfers default was 4 but Filezilla was trying to send more than 8 simultaneous transfers causing Server to do dozens of 5 second timeouts since it has a limit of 8 concurrent transfers allowed. Verified this was the problem in cPanel, which showed 8 current transfers, with warnings. This occurred when I was uploading about 2200 files. Left about 120 as "failed".

Basically, even though setting was 4 concurrent transfers, program was sending 8 and trying to do more, stalling out my server.

After complete, I changed the settings to 3 simultaneous transfers, put the failed files back into live queue and resent OK.

Conclusion: the amount of files in upload queue caused something to override default simultaneous setting.

Thanks. =============== FileZilla Client

Version: 3.5.3

Build information:

Compiled for: i586-pc-mingw32msvc Compiled on: x86_64-unknown-linux-gnu Build date: 2012-01-08 Compiled with: i586-mingw32msvc-gcc (GCC) 4.2.1-sjlj (mingw32-2) Compiler flags: -g -O2 -Wall -g -fexceptions

Linked against:

wxWidgets: 2.8.12 GnuTLS: 2.10.4

Operating system:

Name: Windows XP (build 2600, Service Pack 3) Version: 5.1 Platform: 32 bit system

When being idle in Filezilla for Mac for about 5-10 minutes, the program tends to "lock" the folder making navigating away from it impossible.

The error msg reads: "Error: Connection lost by timeout Error: Could not fetch folder content list" (translated from Danish)

Only solution to this is establishing a brand new connection to the server, which seems quite time-consuming the fifth time.

I sometimes use the cursor keys to navigate in the remote directory list. Unfortunately, when I e.g. go down three directories, FileZilla loads the file list and when I´m already at the third item, FZ jumps back to the first directory.

To improve the handling of this, I propose that FZ only starts getting the file list after e.g. 100 ms. That way you can navigate with the cursor keys withouth FZ interrupting and you still won´t notice a lag when FZ starts to get the file list.

It´s already possible in the file list to press the first letter of an item you want to go to (instead of clicking with the mouse).

Would be great if this also worked in the directory list.

I a Perl program that checks a folder for incoming files and this program utilizes Perl's file locking mechanism which appears to work universally with other Window's programs. It does not work with your client which either implies that you are not locking the files that you are writing to on the user's local machine or a machine on the LAN. As example would be to fetch a file from a customer's machine via FTP and your client and to write it locally. The Perl program attempts to lock its incoming file which would be your outgoing file and the Perl program is able to lock the incoming file, indicating that your client is not locking the file that it is writing to. I would officially request that this file locking support be added for the local output files.

I won't get environment variables to work in defining local directory on sitemanager.xml:

<LocalDir>$USERPROFILE\My Documents</LocalDir>

Or should they even work? Is the config location only setting where it owrks?

Hi,

on a German 1&1 (1 und 1, 1und1) server, if I download a whole directory incl. subdirs, after a while, I get many 425 errors, can not retrieve directory and can not download (as logged).

But if I go manually to one of that directories, I can see the contents and can download the files.

See attached FTP-Log. Errors begins near line 404.

FileZilla Client

Version: 3.5.3

Build information:

Compiled for: i586-pc-mingw32msvc Compiled on: x86_64-unknown-linux-gnu Build date: 2012-01-08 Compiled with: i586-mingw32msvc-gcc (GCC) 4.2.1-sjlj (mingw32-2) Compiler flags: -g -O2 -Wall -g -fexceptions

Linked against:

wxWidgets: 2.8.12 GnuTLS: 2.10.4

Operating system:

Name: Windows XP (build 2600, Service Pack 3) Version: 5.1 Platform: 32 bit system

Please implement the ability to download list of files.

Example list:

​ftp://user1:password1@ftpserver1/path/filename ​ftp://user2:password2@ftpserver2/path/filename ... ​ftp://userN:passwordN@ftpserverN/path/filename

List of URLs can be either loaded from text file or entered into text box.

Real case: I'm a telecom engineer and very often I have to download list of firmware that contains several files. All of them located at the same server but username and password are different and changed once in several hours. It's very time consuming to enter it manually every time for every single file.

Allow the ability to include more than one site to the filezilla command line so that upon starting the filezilla client there will be will be two or more tabs, each with an already active session.

Thank you.

hey, I'm using filezilla for a long time and i love it.

But there is one feature i'd love to see: git support (or version control support in general) So i don't have to check which files i edited before i upload, for example git can tell me those changes.

See this for an example:

​https://github.com/resmo/git-ftp (Git powered FTP client written as shell script)

I think lot of people would need this.

Sorry for my bad english.

fred

(i'm using filezilla 3.5.3, osx 10.6.8)

Hello,

I think it would be beneficial to include some extra details for ssl/tls connections.

A few that come to mind are the ssl/tls version used example TLSv1, TLSv1.1, etc.

Also, I think it would be helpful to include the signature algorithm used example SHA-1, SHA-256, etc.

Thanks

A small but useful enhancement: When sorting a file list where an entry in that list is already selected, it would be nice if the sorted list could be offset so as to display the selected file after sorting.

Have been using FileZilla client for quite a while without the need to change the layout and sizes of the columns or windows but came across some files with extremely log names and had to change window and column sizes. Wanted to reset the UI back to the original layout but could not find a way to do this with a single click. Seems like this option does not exist, thus this RFE. Please provide a way to reset the layout to the original sizes and proportions.

The ticket repo is full of issues related to view/edit where the wrong file is uploaded, problems with same-name files getting saved back to the wrong place, etc. Please do not think this is fixed... I had a server down 4 hours today because a file was written to the wrong place via FileZilla's view/edit.

The goofy async timing of when FZ notices a change can occur out of order or with unexpected delays. Even if this is not technically a "bug" it is a user interface nightmare and it is no wonder the wrong things occasionally get written to the wrong places. I love the ability to view the contents of a remote file, but the "save" thing really, truly does not always work right.

I would like to be able to set a preference to completely disable the auto-upload/saving of files I am viewing. That is:

• "View/Edit" becomes "View".

• Always view with a fresh copy to be SURE it's the correct remote file.

• Never offer to save changes back to the server.

thanks for listening!

The download queue should be saved periodically. This should be done so that in the event a system or application crash occurs while an upload or download is in progress the data xfer can be resumed and continue where it left off without having to figure out what has completed.

A download was in progress using Filezilla build 7601 on Win 64 NT build which failed to respond due to a timeout. This resulted in loosing all downloads in the queue after Filezilla was killed and restated. If the queue list had been saved, every 2 min's then after the app crash the downloads could have been resumed. Since I did not know the locations I had been ftping from data was lost as a result.

please consider saving items in the queue so when an event does occur it doesn't wipe out the queue and an unexpected system or add crash.

Thank you \

--Steve

Currently there is a way to specify a maximum number of connections to a specific server, however on common configurations the maximum can be very low (3-5) with a large number of these servers also imposing connection based bandwidth restrictions. In the event that the end user is downloading a large quantity of files with a range of sizes the download queue will tend to 'choke' if all of the current download slots are filled with large files.

The proposed functionality is to add the ability to specify a subset of the connection limit to be dedicated to small files. Additionally a configuration on what a 'small' file is should be provided.

Once implemented this functionality will subvert the aforementioned 'choking' by allowing smaller files to complete in parallel then merging the small file connection pool with the main pool, allowing for larger files to use the connections.

After reviewing similar log file tickets (5133, 5169, 2753, and 2422) and observing they didn't mention my concerns, I thought I would offer two suggestions that I think would improve your valuable application.

First, I suggest addition of an option to browse and select a custom folder for log files. That would allow use of a separate data drive which might have more space and/or more frequent backups than the OS drive, and would also be more consistent with the concept that one can already custom locate the ftp store.

Second, I suggest that the date stamp format of each log entry always be MM/DD/YYYY, and the time stamp format always be HH:MM:SS. This format change would use a leading zero when the month, day, or hour is less than 2 digits so that log file text strings can more easily be imported as a columnar or structured data file. Also, the AM and PM entries are sort of redundant or superfluous and could be dropped. The more rigid string format would probably benefit those who try to programmatically automate the review of large log files, but have to first pre-process each line to allow for or rewrite varying time and date strings.

Below, I offer two recent actual log entries and a theoretical third log entry as an example of the current format. My point is to illustrate how current entries don't line up in a columnar fashion because of the varying format of date and time entries:

1: "(001739) 1/14/2012 0:00:02 AM - hbwilliams (192.168.102.34)> QUIT"

2: "(002190) 1/14/2012 16:19:58 PM - hbwilliams (192.168.102.34)> 150 Connection accepted"

3: "(002190) 12/14/2012 16:19:58 PM - hbwilliams (192.168.102.34)> 150 Connection accepted"

Implementing my suggestion would reformat the data as follows where each entry would line up in a columnar fashion, and the data strings would be shorter since AM and PM is dropped:

1: "(001739) 01/14/2012 00:00:02 - hbwilliams (192.168.102.34)> QUIT"

2: "(002190) 01/14/2012 16:19:58 - hbwilliams (192.168.102.34)> 150 Connection accepted"

3: "(002190) 12/14/2012 16:19:58 - hbwilliams (192.168.102.34)> 150 Connection accepted"

In closing, an earlier request to offer shorter time increments for log files would be helpful because a daily log file can grow very large and be slow to directly open and review.

Thanks for such a great product.

Hi there,

I tried to find a solution on the forum as well as on the internet, but it seems no one ever faced this problem. And another user gave me the hint to send you a feature request on this topic.

Regarding proxy traffic at a customer's of mine I face a rather unexpected issue: top 5 of download traffic on the ftp proxy results in local ftp servers, looking at upload traffic it's even top 4. As to say, this is a 3.000 clients plus network so any changes made by hand are no use at all, such as some explanations I found mentioning "switch to IPv6".

Thing is, FileZilla stores some information in XML files, and in theory it surely would be possible to adopt sitemanager.xml for specific sites with <BypassProxy>1</BypassProxy>. However, what I'd really like to do is bypass really every local traffic because there is simply no need at all to weigh the proxy down. Thus, even if a user enters an internal address manually it will not result in proxy requests.

And hardcoding specific addresses in XML files is no use anyhow, as I would have to specify username and password for this machine here as well. I want to assume the companies' compliance managers would rip off my head for doing so

Kind regards, Patrick

I often transfer files through FileZilla that are of varying sizes; and I often find myself prioritizing files by size, since I want to download the smaller ones first, then gradually start downloading larger files as the queue gets completed.

So here's my idea; Say, while you're transferring a 10MB file that takes some time to upload, if the next file is only 2KB, for example, then it will override the transfer limit automatically and upload that 2KB file while it's uploading the 10MB file. if the next file is, say, 3MB, then it will wait until the 10MB file transfer is done.

It's kind of hard to explain but I hope you get the gist of it. This would be nice to have as a option, especially when you are backing up a website via FTP to your computer.

Anyway, I hope the see this feature implemented eventually. Thanks for reading this long request. I have to say that I would not be as happy using FTP if it weren't for FileZilla; it's been the most helpful program for FTP uploads I've ever used. Thank you so much for developing it.

Cheers!

-Camo Yoshi

I'm using a tool, which help me to automate creation of queue for FileZilla.

Recently I've found an issue, why sometimes not all files was uploaded. The reason is, that FileZilla doesn't automatically create directories on the FTP, if that directory is missing. FileZilla tries to switch to that directory, and after unsuccessful try it breaks connection, as it shows on attached screenshots.

Expected behavior: FileZilla shoud automatically create missing directories on the FTP, if they are missing during queue processing.

I really like FileZilla, but today I started to upload some files and I understand that the remote directory is wrong. When overwrite confirmation dialog appears I click "Cancel". But the upload process doesn't stop. And I didn't know how to stop it actually.

We have the "Skip" option in this dialog, so please make "Cancel" button to cancel the upload process.

I've been trying to upload to and download from 2 different servers on the latest version of FileZilla. While the uploading process seems to be working properly, there have been issues when I have tried to close FileZilla.

When I'm uploading/downloading from a single server, I just have to stop the processing of files and close the program. The files will be in the queue the next time I open the program.

With the multiple servers, if I try the same process, after I try to close FileZilla, I get a message telling me that all or part of my queue has been lost. When I open the program again, there will be nothing in the queue. It is rather frustrating to have to re-queue the files that didn't upload.

I am a huge fan of Filezilla and I use it everyday on my windows pc.

One thing that I noticed that is missing from Filezilla is the directory caching feature. I connect to my server several times a day and its a huge waste of bandwidth to retrieve the file listing everytime, not to mention the wastage of time.

I would really like to see Filezilla give an option to cache directory listing.

When a remote folder has 2 subfolders that differ only by capitalization, the Remote Site box content does not change when you switch between these two folders. Example: my site has these two folders: "Manuals" and "manuals" The Remote Site box always show "manuals" regardless of which folder I am in. The folder content are correct though, so it appears to be only a problem with what is displayed in the Remote Site box.

To reproduce:

If I upload a file which is large in size (in this case let's say 500 MB) on one terminal to transfer it to some other terminal and I open filezilla on other terminal while the upload is in progress and see that the file size is 250 MB since it has not been uploaded completely. Now, I wait for the file to upload completely and once I have verified that the upload is complete, on the other terminal (on which the window is already open and the file size show 250 MB since it has not been refreshed), I start the download. The file download progress bar that appears at the bottom takes the size of file to be 250 MB only and shows 100 % completed when only half the file (250 MB) has been downloaded. Though the complete file downloads to the other terminal, for half the time, progress bar says that the download is 100 % completed.

Thanks

would be very useful to have the Site Manager accessible from the Jump List of the FileZilla shortcut in the taskbar, or the option to use the Quickconnect list instead so we can open the program directly connecting to a server.

I routinely start 6-8 long parallel file transfers from my ONE host to my local PC. In general this works very well, including respecting any up/down speed limits that I set. However, there is one specific case where this breaks down.

The transfers seem to "set max their speed" when they are started, some are slow, some are fast. The first ones that start usually run very fast, the others usually end up going much slower.

The problem being that after the fast ones finish first, the slow ones do use the extra bandwidth and continue running as slowly as if there were other fast transfers still present.

This means the slower transfers take a VERY long time to finish even though there is plenty of unused bandwidth available. They never increase their speed up to the speed limit once the other transfers complete.

All of the active transfers should adjust up to use available bandwidth equally, and dynamically increase/decrease speed as transfers come/go.

It was reported that in version 3.5.0 that CVE-2008-5161 and CVE-2008-1657 Cipher-block chaining ciphers were supported. I am currently running 3.5.2 and this is not supported.

Additional information can be found at ​http://web.nvd.nist.gov.

Under menu item "Hilfe", there is an entry "Nach Updates suchen..." (German version); English version should be something like "search for updates". This task fails, as I'm sitting behind a proxy-server. I'm able to set proxies for my ftp connections, but I can't set a proxy for this update feature. This would be a nice feature to add to the FileZilla client. Thanks!

I've created a filter to hide files/folders that start with '.'. This is useful as I usually do not need to access or transfer hidden files. But I occasionally do have to. I was wondering if a quick select (dropdown select box) or a toggle button for any saved filters. Possibly above the 'Quickconnect' bar.

Sometimes FileZilla will - on repeated connections - fail to issue the PASV and LIST -a commands despite geting a "226 Transfer complete." after a LIST command.

This happens on the same server, same remote directory, same conditions as when it works, but when this happens, it almost always continues, each time I abort session and reconnect, or when I restart FileZilla. I have had to reboot the computer, which solved the issue on several occasions. Only once did it work again without a reboot, after 3-5 times reopening and re-trying with FileZilla, then taking a break, before trying again.

It appears to be the client not responding to the LIST command response, as normally the client should issue PASV in response to "226 Transfer complete."

For example:

Status: Resolving address of ########.### Status: Connecting to ##.##.##.##:21 Status: Connection established, waiting for welcome message... Response: 220 Microsoft FTP Service Command: USER ########## Response: 331 Password required for ##########. Command: PASS * Response: 230 User ########## logged in. Command: SYST Response: 215 Windows_NT Command: FEAT Response: 211-FEAT Response: SIZE Response: MDTM Response: 211 END Status: Connected Status: Retrieving directory listing... Command: CWD /###### Response: 250 CWD command successful. Command: PWD Response: 257 "/######" is current directory. Command: TYPE I Response: 200 Type set to I. Command: PASV Response: 227 Entering Passive Mode (##,##,###,##,##,###). Command: LIST Response: 125 Data connection already open; Transfer starting. Response: 226 Transfer complete.

Here it just hangs, instead of the usual:

Command: PASV Response: 227 Entering Passive Mode (##,##,###,##,##,###). Command: LIST -a Response: 125 Data connection already open; Transfer starting. Response: 226 Transfer complete. Status: Directory listing successful

Hello,

I use Windows 7. In august 2011 I experienced that FileZilla Client did not update the content on server side - even after manual refresh. Maybe this bug has been resolved in newer versions of FileZilla Clients?

Thank you.

When selecting multiple files on a server to edit, when it gives you the option to re-download or open local. There should be an option to apply to the queue, like when overwriting files.

An error occurred with TLS renegotiation while downloading file > 1 GB on server using passive FTP with explicit authentication on TLS.

Using latest 3.5.2 Filezilla Client on Win32 platform (XP). Connecting to z/OS v1r9 FTP server, downloading Fixed Block datasets where trailing blanks are critical to retain.

Using Custom Command in Filezilla to send command: site trailingblanks ("site trail" for short)

When the dataset is then downloaded in ASCII format, the blanks have been stripped.

I can perform this command from both a WSFTP LE v12 GUI client and both the standard Windows FTP client and the MoveItFreely FTPS command line client, and it will work and the dataset has blanks properly retained. I cannot attach data since it is sensitive.

When moving multiple entries one for another with the mouse into some folder, that folder always opens.

Now, when you have a long list, you always have to close that folder to get to the position you´ve been before. If a folder the mouse hovers over would only open after a second (and also not when the entry is dropped), this would save some effort.

At the moment, you have to press Alt+D to delete an entry in the site manager. Could you please also add the possibility to use the "del" key for that? I think that is more intuitive.

There are many times when I'm servicing a FTP site when all of my download slots are transferring and I want to edit a site file. I right click on the file and the file is added to the end of what might be a long queue.

A useful feature would be to be able to reserve a download & upload slot for using View/Edit file.

When I upload a file or save a file that gets uploaded to the server. Not the whole file is uploaded. For example, if the page is 54,000 bytes and I add to it then save, it may save at 49,000 bytes and part of the page will be missing. Although, the whole code is still in my editor, it's just not saving to the server. There has been a few circumstances when the whole file would be deleted. A 54,000 byte file would go to 0 bytes when saved.

Recreate the bug: 1- Add couple of files to the queue. 2- Make the simultaneous transfers less than the queued ones and start the transfer. 3- Add another files to the queue.

After finishing the download that already started while adding the new files the queue will start download the latest added files instead the queued ones before.

Hello, while im using sperate connections to different paths, i have to confirm always a connecting - dialog.

May you can set up in preferences, so it will always (like in version 2) connect to the same tab as i want ?

I dont need this dialog. I want disconnect the acutal connection and connect to another without asking me.

This dialog is going to annoying me really :(

I was just wondering if you couldn't change how FileZilla deletes files recursively but also CHMOD as this is pretty much the same thing. If you mark a folder and press delete, FZ scans the folders content and deletes the files one at a time. When using SFTP I am sure that the fastest solution would be "rm -r FOLDER" The same thing goes for CHMOD where Linux incorporates a recursive function into the command too. "chmod -R FOLDER" So everytime I delete a folder I am thinking "Why am I wasting my time on this, when I could do it 100 times quicker on SSH?"

I use putty as a proxy to phpcloud.com, and save the setting in putty. but when I try use FileZilla to connect phpcloud, it will timeout after: Initialised zlib (RFC1950) decompression And when I delete the setting saved in putty, filezilla can connect to phpcloud successful.

I open the debug log, and the things different is: If I save the site setting in putty, FileZilla will doing: Initiating key re-exchange (enabling delay compression) and then will timeout after: Initialised zlib (RFC1950) decompression

It happened at FileZilla 3.5.1.0 and 3.5.2.0

OK, ticket #2173 said the Windows App Data, should be stored in [system drive]:\Documents and Settings\[logged on username]\Application Data\[Filezilla] ...

And there, i suggest it should be point by commond line option, like google chrome's --user-data-dir <path-for-custom> or any else it can do this...

Any body want a portable version for filezilla client on USB strore device really need this feature.

In FileZilla, it is possible to select in the preferences how file sizes are shown. If I select for example the IEC style (KiB = Kibibyte = 1024 Byte) instead the default "Byte" option (=no KB,MB,GB,etc. just Bytes), FileZilla applies this perfectly to everywhere where file size appear, EXCEPT when one file gets transferred. Then, in the transfer queue, the transfer speed of all is still shown in the "KB/s" (for example) format instead of the "KiB/s" format which I selected in the options. All other file size numbers are formatted as selected in the settings. It seems the shown transfer speed completely ignores my settings; it also shows "KB/s" if I selected to show only Bytes.

hi just a simple bug in tranfering tab when use of hebrew lang the numbers and letter is reversed !!!! very annoiny.. thank for your time great app!

image : ​http://img263.imageshack.us/img263/291/16524663.png

I tried to rename MDCT.PROD.C(TTR249) to MDCT.PROD.C(TTR249O2). Below is log with error need to be fixed (output name is incorrect).

Command: PWD Response: 257 "'MDCT.PROD.C'" partitioned data set is working directory. Status: Renaming MDCT.PROD.C(TTR249) to MDCT.PROD.C.TTR249O2 Command: RNFR 'MDCT.PROD.C(TTR249)' Response: 350 RNFR accepted. Please supply new name for RNTO. Command: RNTO 'MDCT.PROD.C.TTR249O2' Response: 550 Cannot rename from partitioned data set MDCT.PROD.C(TTR249) to a different partitioned data set MDCT.PROD.C.TTR249O2.

Hello. I'm using filezilla fondly. However, there is a significant deficiency. Ftp site of dozens of information I registered in the program. If the flash memory is stolen, or disappear, people who run filezilla can reach all the sites. Why does not turn on with a password filezilla? Opening must be a user name and password is. also sites for the encrypted user name and password. I'm sorry, my English is very bad. I use the Google translator. Byee

Saygılarımla... Hakan TOPTAŞ Bilgi İşlem Sorumlusu

I find myself adjusting the "Maximum Simultaneous Transfers" on a fairly regular basis to adjust link usage and favor finishing some longer transfers over more transfers.

Please consider adding two new toolbar buttons to increase / decrease the "Maximum simultaneous transfers" count.

If you shutdown politely (or the system shuts down), the app DOES remember what was in the transfer queue. Unfortunately, it does not resume (reget) the transfers, it restarts them from scratch.

This seems like a bug, but I suppose it could be a feature request.

More specifically, if I shutdown & restart the app, and begin processing the transfer queue, I would like transfers to continue so it does not take many many hours to recover back to where you were.

My php file is associated to editplus. But when I select 'View/Edit' entry in filezilla, it prompts me "the file'test.php' cannot be opened: The associated program (WX_DDE#D:\Program Files\EditPlus 3\EDITPLUS.EXE -nm##System#[open("test.php")]) could not be found. Please check your filetype associations."

I looked in the FAQ but didn't see anything on it. There are 2 little lights at the bottom right corner. One is blinking red and the other is gray. What do these buttons signify?

Thank you in advance. PGelsman

Hi.

FileZilla v3.5.1 does not delete edited local file (manually downloaded *not* edited on-the-fly) when checkbox "Finish editing and delete local file" is checked.

Steps to reproduce:

1. Start FileZilla
2. Connect to FTP
3. Drag-and-drop file from remote panel (or use "download" option from context menu) to local dir, edit it and save changes.
4. A confirm box appears and asks you: "Finish editing and delete local file?" Check this option and click yes / no. Local file is *not* deleted.

Note: edited file is deleted (if checkbox is selected then file is deleted regardless which button I click - and that's ok in my opinion :)) when editing on-the-fly (using context-menu option called: "view / edit".

I don't know if this should help but this problem exists from some previous stable versions of FileZilla (or is it just my imagination?).

Thanks for your kind help and keep it up cause FZ is the best FTP solution ever. :)

While filezilla.xml corrupt, it should be auto repair or something, since it annoying me after open the filezilla, pop out a warning always

FileZilla 3.5.1 Mac OS X 10.6.8

When I change a remote symlink via SSH, it is not being updated in FileZilla's remote window. In fact, the symlink is still leading to a directory that doesn't even exist anymore (and actually displays its contents), so it seems a cache issue or something. Upon restarting the client, the symlink works, making this a very low-priority bug.

I have verified that the symlink is correct in the shell.

When updating filezilla.app the program wont open, the icon bounces & console says theres a permissions problem.

Ive tried:

• Adding System, Staff, Administrators, Wheel, Everyone all as "Read & Write" for the app. Doesn't fix it.

• tried chmoding Filezilla with "chmod -R +rwX FileZilla.app" and "chmod -R +rw FileZilla.app" and nothing works.

i've tried deleting everything associated to the program (application support, the app, xml files, etc) and restarting, clearing trash, resetting pram... and also just installing fresh.

now it seems it's stuck not working.

this has happened on 2 previous updates and it was a month or two ago and i forgot how i fixed it. i want to say the adding system worked last time. but not this time.

ive verified & repaired disk permissions. no dice...

I was downloading a Joomla website with a reasonably large amount of temporary files to my local computer. The total number of files moved was 7,079. However, the folder structure on the target side is not the same as the on the source size and files seem to have ended up in random locations. It looks like that the 'active' folder for the file transfer was set to whatever the 'active' folder was for the thread that was queueing all the files to be processed. For example. I have a folder "templates" on the source side, with subfolders "beez", "ja_purity", etc. But on the target side, the "templates" folder does not have any subfolders, but has just contains a few files "blogger.php", "blogger.xml" and a few more, that should be in "plugins./xmlrpc". The "beez" and "ja_purity" folders have ended up within the "tmp" folder instead of in "templates" on the target side. But then have subfolders themselves that are in the tmp folder on the source side.

I have the max simultaneous transfers set to 1. The source server is a Linux machine, running Apache 2.2.3. The target machine is a Windows 7 one.

I get request for list of files or have to create an html page from a list of files and wondered if that functionality of saving the highlighted list of files into the system clipboard so I can paste into my editor.

Thanks

Hi, I'm always scared that I'll completely remove a failed file transfer because of a problem with the user interface.

Basically, when a file fails, it goes into the failed tab - as you know! You then right click it and there are a list of options. The first option is 'Remove', the second option is 'Reset/Requeue'. Great. Now, it goes to the Queued files area. You right click again - but 'Process Queue' is in the first position - on the other tab, 'Remove' is in the first position. It would be nice if 'remove' was the SECOND right click option on the 'Failed Transfer' tab, as it would create a better User Experience and make it easier to requeue and restart the failed file.

I hope this makes sense! I can include screenshots if not.

I've ranked this as high as new users may expect these options to be in the first position, and may accidentally remove failed files (and it's almost impossible to tell which were removed if there are a lot).

Thanks!! :) Great software.

Whenever i upload images on any server all images get corrupted. I uninstall and reinstall filezilla, but still problem is not solved, so please help me here what's wrong. Thanks.

FileZilla client not able to upload folders containing files to Maverick SSHD version 1.4.32

Folders that do not exist on the Server Maverick sends back: SSH_FXP_STATUS: : File not found reason=2

FileZilla does not appear to be properly interpreting the response and does not issue a SSH_FXP_MKDIR

Program installs. Shortcut pinned to icon menu. But when I click shortcut, it does not display.I see it as a small window above the icon but it does not display large on th screen.

The program works fine on my main computer and on my work computer. But on this second computer at my home I have these problems.

All three computers are running Windows 7 64-bit. The two computers that work are Intel i7 processors. The one that does not is a dual-core AMD processor.

VMS stores several versions of a file.

I am using "Strip VMS revisions" and overwrite if file is newer options.

I would expect to only receive the latest VMS versions on my Windows machine.

It seems like I receive a random version of the file.

It takes appx 3 minutes to get interface loaded if there are network drives present on the system. Further actions take even longer, and cause app to become non-responsive too.

Greetings.

Like many people, I'm now using a widescreen monitor. Currently my FileZilla client is split into four horizontal panels: message log, directory trees, file listings, and transfer queue. As a result these panels are all very short and wide.

It seems the space would be better used if the GUI allowed me to snap the panels to the sides of the screen, so that (for example) I could have the message log along the left, or the transfer queue along the right, or both. Then I would have four columns: message log, local tree/files, remote tree/files, and transfer queue, and there would still be plenty of horizontal room. Alternately, I could leave the log and the queue at the top and bottom, and have the directory trees to the left of their respective file listings rather than above them.

A good example of this behaviour can be found in most IDEs, which tend to have a lot of different panels available, but are flexible in how they can be arranged. You can dock them to an edge of the screen, or split them with another panel already docked to that edge, etc.

Just a suggestion. I love using FileZilla, and an improvement like this would only make me happier.

saltire

Filezilla 3.5.1 cannot connect to sftp servers with either key based or password based methods. Logs with debuging/verbose=3 attached.

I have verified that this is not a function of my system by repeating on an independent box with the same results. As well, I am pretty sure that nothing is wrong with the server because I used "sftp -P 34451 facade@mars" from a shell and connect just fine. As well, using gftp, I am able to connect. The Host OS is Archlinux x86_64 which is up-to-date (Arch is a rolling release).

When I upload a file to the server, the file list on the server directory updates according the new upload. But when I click the button "Update Directory and file listing", the list on the server "goes back" to the previous list.

If I log out and log in again, the list is correct, but if I click that button again, the previous list goes back again.

The listing of the local side is always correct.

MyFileZilla version is 3.5.1 and operation system Windows Professional XP Sp 3 (version 5.1)

There would be very usefull if we could choose the FTP Keep-alive function for each entry in the Site Manager. What about a CheckBox at Advanced tab, on the Site Manager to make it, choose FTP Keep-alive on or off.

I had been testing different scenarios of uploading a bunch of files with different network connections (wlan and eth) and have a good idea for improving upload speed.

• Lots of small files: If transferred in parallel, the whole proccess is completed in less time than transferring them one by one.
• Lots of large files: Transferring them one by one is *faster* than transferring two or more at the same time (the sum of all individual speeds did not equal than the individual upload speed of only one upload).

So with this data I suggest to have a configuration that would enable "automatic" parallel handling that would work as follows:

Configuration needed: 1.-Threshold to consider a file small (called A) 2.-Maximum number of threads to enable if procceeds (called B, can be taken from global config)

Suggested behaviour WITH this new option enabled: 1.-Sort all upload list by size (asc or desc) 2.-Start transfer:

3.-If current filesize is less or equal than threshold A

then

• Upload them in PARALLEL using B connections

else

• IGNORE parallel options and upload files ONE BY ONE

I tested this scenario locally connected directly by ethernet cable, and by wireless lan, and in both cases, speed improvement was really really significant.

So what about having it in config so this can be done automatically?

Upload of files produces zero file size on server. A renamed file on the server produced zero file size upon save. After "Checking for Updates" and uninstall/reinstall, able to upload file as created; however, if server file is deleted and same file is uploaded again, zero file size is result again. Repeated uploads does not resolve. Deleting file from server first and then uploading file does not resolve. Does not happen with just this same file. Closing program and restarting does not resolve. Disconnect/reconnect does not resolve. Only uninstall/reinstall resolve it for one time upload thereafter. Filezilla vers 3.5.1

Is is possible to include a preview pane when clicking on a file either in the upload or download screen. Especially dealing with pictures, I sometimes have to open several different ones just to make sure I am uploading the correct one. I did search tickets for "preview pane" and returned "no results found"

On uninstall, FileZilla doesn't delete settings like sites entries in the Site manager, meaning if someone else installs FileZilla afterwords they can access all servers.

I think it should delete all Entries in Site manager on uninstall.

FileZilla:FileZilla_3.5.1_win32-setup

When trying to upload a file that already exists, I press cancel on the box that asks whether I want to overwrite, and the file is then added to the successful transfers tab, so it makes me think it was uploaded anyway, even though I clicked cancel, I don't think it's a bug since the file isn't uploaded, but it's confusing.

FileZilla allows you to export Settings to an xml file (File | Export Settings). This works. You can export settings on one PC (Windows Server 2003 in my case) and then import (File | Import Settings) the xml file on another PC (Windows 7 Professional 64 bit in my case). This part looks like it works, but it does not work 100%. Let me explain. The import updates your Edit | Setting | Transfers | File Types ascii list and the change works when you do an automatic mode transfer (Transfer | Transfer Type | Auto). However, if you close and reopen the Filezilla Client, your changes to the File Types ascii list are lost. To Test . Export your settings. . Delete one of your ascii File Types, e.g. bat . Close and reopen Filezilla Client . Go to File Types ascii list, bat is not there, great. . Now import your settings that you exported earlier. . Go to File Types ascii list, bat is there now, great. . transfer a bat file to a UNIX box in auto mode, great, that works. . Close and reopen Filezilla Client . Go to File Types ascii list, bat is not there, BUMMER.

The settings that you import are lost on restart.

This is a pain, if you are trying to use the same "standard" File Types definition at more than one location. If this worked, it would make it a lot easier to use automatic mode for a mixed bag of ascii and binary file types. Youp would be able to get repeatable results.

A separate issue is a feature request: to allow the user to choose to remove items from the Files Types List when importing a Setting XML file, if those files are no in the XML that you are importing, but are in your existing File Types List. Not havign this feature, means that you have to empty your Files Types List before importing the XML or manually delete unwanted file types before or after importing the XML.

In the remote tree view..

it would be useful that when a leaf (a folder) is clicked or expanded ('+' icon)..

the subfolders underneath show the '+' icon if they contain at least one subfolder.

That way you can navigate in the folder structure quickly with less clicks. Right now, I have to press the folder with a '?' sign, wait a bit, then I can expand it.

That behavior could be optional and configurable in the settings.

It's interesting to add an advanced option for user can choose not to send the FEAT command to server and therefore not to use the advanced options.

This is useful for servers that advertise support for commands like MLSD but they do it with bug.

on public or web hosting ftp servers which only work for 1 user at a time, filezilla for windows uses 2 connections. please fix.

apparently this does not happen with the unix version, tech support tells me (or he's just on top of the network and gets no delays, and thus no lockout after 5 minutes).

the server is an (possibly an overloaded) ProFTPD as usual.

I have two local directories that I'd like to compare using Filezilla, which I find much more useful than the conventional directory diff utilities. (The synchronized-browsing and directory-comparison tools are especially nice.)

But, unlike with gFTP for example, I can't find a way to load a local directory in the remote pane. Could this be a new feature?

Hi, I saved a site with logon type Interactive. On each logon it asks for user. Then password. I tick remember password for this session. It shows the directory listing and won't ask for pass for navigating anymore. BUT for every single file transfer (in or out), I have to retype my password. Even if I recheck that remember box... Thanks, Max

My ID EQXFF3Q is used for the FTP. Gets message - fails - sometimes restarts.

BPXF024I (EQXFF3Q) Aug 11 11:59:39 ftps 33 : EZZ9819I FTP unable to 821 obtain EXCLUSIVE use of SYS7.G1.UNLDV8.DPVH.DB which is held by: 00A0 EQXFF3Q EXCL on SYSDSN

New version of Filezilla downloaded due to PC refresh. Have been running these FTPs now for several days. Previously was not zippy fast, but the restarts are killing my efforts.

Currently, the My Documents folder is mapped to a network folder via a UNC path:

\\Vickymini\vicky_mini\Users\Tech_Support\Ken_Haberman\My Documents

When selecting "My Documents" in the local directory tree list, the program displays my local folder instead:

C:\Documents and Settings\ken_haberman\My Documents

It takes a few tries of clicking on "My Documents" before it shows the mapped network. Sometimes I have to open the tree to the network folder to get to it.

It would be nice to have FileZilla support a proxy command for sftp. I use the ProxyCommand option in my ~/.ssh/config file to be able to connect to hosts I would otherwise not be able to reach. And while I could try to set up a tunnel some other way outside of FileZilla, that tends to be more of a nuissance.

Ideally, if FileZilla could read my ~/.ssh/config file and take advantage of the hosts and settings I had set up in there, it would be a nice thing to have.

It would be beneficial to download and user interaction response times if FileZilla could reuse the browsing connection (the first connection made and browsed with) to initiate file transfers, and open a second connection to resume browsing until the transfers have completed.

This arose from my use of a ssh+ftp connection which has very quick transfer times but was very slow to establish connections. With this connection policy/option available a simple two-file grab could be much quicker, and by automatically opening a connection before the user browses to the next directory you minimize overall delays experienced by the user.