This question already has answers here:
Don't lock files of application while running
(4 answers)
Closed 9 years ago.
I created a small application which users launch through a desktop shortcut pointing to a network share. Whenever I recompile the app and want to move it to the network share to make it available, it is always locked by many users who are using it, understandably. What I don't get is that I can close all the handles (net file 1234 /close) and the users are unaffected, i.e. they can still work on the app. I then copy the new file and ask them to restart.
Is there a way to "cut off" programmatically the users from the network exe file once they have launched it, so that I don't have to manually close all of the handles every the time?
They'll be affected but the way the jitter works indeed makes it a bit unlikely that their program will crash. Crashes are likely when you use Reflection in your code or the user's machine is under memory pressure from other processes and the program executes code that has not been jitted. YMMV.
The proper deployment choice here is ClickOnce. That creates a local copy of the program on the user's machine. And it automatically sees your update when they restart.
A possible band-aid is renaming the executable. Which works because Windows puts a lock on the file content, not the directory entry. Which lets you copy the update with the same name. Any user that restarts the app will now get the new version. You'll eventually be able to delete the renamed file. But do favor ClickOnce first.
Related
Good morning.
I'm in the process of writing a vb.NET forms application that will read in a selection of .xlsX files and import their contents into a SQL2012 database. Because the files are in different folders and are all formatted differently, I'm having to write it so that each folder's contents is handled by its own dedicated module. However, one thing that's common across each folder is the process that I need to go through, which is to open each file, read its contents into a DataTable, carry out any manipulation required (i.e. removing empty rows) and then run a SqlBulkCopy to load the data into SQL before moving the original file into an archive location.
So far, so good. I've written and successfully run three of these modules, but the fourth one is giving me the error that's detailed in the Title of this post - the exception is thrown at the point where I'm trying to open the connection string to the Excel object. Again I stress that I've done this three times before, and each time has been successful.
Also, I've noticed that the exception only occurs when running the code in Debug mode. If I run it in Release mode it works without any complaints.
I'm developing this application in a 64-bit environment (VS2010 on Windows 8.1), but targeting the application to x86. I'm happy to continue writing in Release rather than Debug mode, but I'm curious as to why it works in one but not the other and I'd like to be able to code for both modes if at all possible.
TIA
Imagine there are 3 or more independent locations where a file can be modified. These locations communicate to each other through email or mail (direct flash drive restoration). Though there is a big room for flow - to make simultaneous editing to the file and screw up things, this client won't change too much. He rather call everyone that he is working on the last update or tell the other guys that he is waiting for third guy's last update. Anyway, at some point after several exchanges, due to one of participants unintentional error THE LAST VERSION of the file eventually gets mixed up. From this point everyone searches for the last version BY LOOKING THE CONTENT of the file.
This client wants to have a central location (he has actually, that is his PC's some location) and let everybody (including himself) copy any new or suspected new file to this location but prevent file's last version being copied. From this location he has to easily copy, send or open the file and work.
So, here is my concept (2 steps):
step 1: I made an ad to the main application where this file is created or edited. This ad prompts the user to give a version number to the file with every invoked save command from the editing application. In fact the file can be re-saved multiple times but not considered modified (file attributes creation, save etc. do not have great meaning here). This said the user can cancel my ad-in but have saved the file, not saving a new file version.
step 2: multiple solutions:
solution A: I'm thinking to have a folder/file watch and prevent the last version of the file being overwritten. As you know, FileSystemWatcher will fire the change/delete etc., events AFTER FACT so, I have to back copy overwritten file after the fact (w/ some tricks).
solution B: have a database to store all version of files and built-in some shell extension to extract/view files from the database. Move all copied/pasted files to the database (my program folder) and restore latest file in working folder after watcher fires change/delete event.
solution 3: find out built-in windows tools (API etc.) to greatly rely on it with some programming.
Any ideas?
Thanks in advance.
I came by this page and was thinking about the best method to distribute my VbaProject.OTM file (located into %appdata%\Microsoft\Outlook\) to a bunch of ~30 users at my office. Is it better to simply copy/paste the OTM file onto the network and then copy/paste it back to all users' computers (manually or with a .bat) OR would it be better to use the method described in the link above to generate a OPS file and import it back with Proflwiz.exe? What's the difference?
We are all on Microsoft Office Outlook 2003 actually, we might upgrade to 2007 one day but still years from now.
Finally came up with some elements to deploy a Outlook VBA Project. There are a lot of ways to do this, but the easiest way to do so without installing anything and keeping the same methodology would be to run a OTM file directly from a server. I found out that the process outlook.exe has a parameter altvba that allows to specify another path to run the OTM file from. Here is en example:
outlook.exe /altvba "\\myServer\myFolder\myFile.otm"
This allows me to update only one file to get all computers updated. Obviously, if the file is big and the server's ping is on the high side, it may delay the launch of Outlook. The other problem with this method is that everybody will have to shut down Office if you want to update the OTM file on the server (and if you do work in an office where everyone uses Outlook, you do know that it is impossible to get everyone to shut it down at the same time, except if you code a macro to do so eventually). To prevent both those problems, I could setup a batch file to copy the server OTM file clientside everytime there is a new version (just have to check the NTFS last-modify attribute). This way, Outlook will boot with a local file, the batch file take 2-3 seconds to copy the file if needed (or will launch Outlook instantaneously) and there will be no problem updating the OTM file on the server. Users will have to start Outlook with the batch file (or with the slightly different outlook.exe path with the altvba parameter, so either way they need a different shortcut/file to start off the first time). One other advantage of the altvba is that it's still easy for the user to run Outlook without it (to see if the VBA is problematic or not in case Outlook is sluggish) and the file will remain unchanged after a Outlook reinitialization.
Others solutions include a COM complement that can be developed in a lot on languages including VB6 (no conversion needed from VBA). There is also a bunch of tools included into Microsoft Office XP Developer that could help getting the job done (not free however, especially if you need the most up-to-date version).
I am developing an application in .NET Compact Framework 3.5, for Windows CE 6. I need to automate some data synchronization between the mobile application and a PC. So what I do is that, in the .NETCF application, I have a while loop that checks every few seconds, using System.IO.File.Exists(fullFilePath), whether the file I am expecting has been copied from the PC to the mobile device, through an ActiveSync connection.
Once the mobile app finds the file, it will open and read it using a StreamReader.
My problem is that, some times, File.Exists(fullFilePath) would return true, yet the subsequent use of StreamReader may fail with a FileNotFoundException.
Ruling out the possibility of the file getting deleted between the call to File.Exists() and the instantiation of the StreamReader, would anybody know how the exception may occur? Is it because file copy through ActiveSync is not atomic, such that the file is not actually ready for reading even though File.Exists() returns true?
A file copy is most certainly not going to be atomic. When you start the copy, the file is created, but has no data in it. As the file data gets written, the file "exists", but is not yet fully complete. You'd want to know not only that it exists, but that it's not currently in use (meaning the writer is done). Trying to open it exclusively would verify that.
I've got an Access 2007 database that I developed which connects to SQL Server for the actual data storage. I used the Package Solution Wizard to create a distributable installer which included access runtime (with an ACCDE file) which I went around and installed on 15 or so PCs. Anyway, my question is, what is the best way to distribute updates to this database? Right now I'd need to go around and remove and reinstall. That's not a problem... I was just wondering if there was another way.
I've tried leaving the front end on a network share but it seems that most people suggest storing the front-end on the local machine, which makes sense. The problems I've run into when I leave it on a network share (at least with Access 2003 mdbs) is that I find myself needing to compact and repair often and I also have to kill the open sessions (user's who have the file open) when upgrading. I would imagine it could also hypothetically create an unnecessary bottleneck if the user was not on the local network.
Automating front-end distribution is trivial. It's a problem that has been solved repeatedly. Tony Toews's http://autofeupdater.com is one such solution that is extremely easy to implement and completely transparent to the end user.
We developed a vbscript 'launcher' for our access apps. That is what is linked to on the start menu of user's pcs and it does the following.
It checks a version.txt file located on a network share to see whether it contains different text to a locally stored copy
If the text is different it copies the access mdb and the new version.txt to the user's hard drive.
Finally it runs the mdb in access
In order to distribute an update to the user's pc all that is required is to change the text in version.txt on the network share.
Perhaps you can implement something similar to this
Make a batch file on the server (network drive).
Create a shortcut link to that batch file.
Copy the shortcut to User's Desktop.
When user double-clicks on shortcut, it will copy a fresh copy from network to local.
Replace old database.adp on the server drive when you update a new version.
Each user gets a copy of database.adp on their machine.
Remove Security warning when opening file from network share is here.
Batch File
#ECHO OFF
REM copy from network drive to local
xcopy "Your_Network_Drive\database.adp" "C:\User\database.adp" /Y /R /F
REM call your database file - Access 2007
"C:\Program Files\Microsoft Office\Office12\MSAccess.EXE" "C:\User\database.adp"
This is a very old post and I used the autofeupdater until it stopped working so I wrote one of my own and it has evolved over the last few years into something that I have used with many clients. It's so simple to use and there is no interface. Just an EXE and a very simple config file.
Please check it out here. I can also help with custom solutions if none of the configurations work for your needs. http://www.dafran.ca/MS-Access-Front-End-Loader.aspx
After trying all of the solutions above (not exactly these solutions but these are the common suggestions in the Access community), I developed a system entirely within Access using VBA that allows an admin DB to create and publish objects to client DBs without the need for user intervention or management of multiple DB files.
This approach has several benefits:
1. It simplifies the development process by having a dedicated environment (admin DB) for development and testing totally separate from the client DBs.
2. It simplifies the update/distribution process by allowing a developer to push out updates in real time that client DBs can implement in the background, without involving users. Can also allow devs to roll back to previous versions if desired.
3. It could be used as a kind of change management system within Access for developers who want to commit multiple changes to objects and modules and retain past changes.
4. It allows for easier user access control by allowing an admin to easily assign certain objects to specific users/roles without needing to maintain multiple versions of the DB.
I will hopefully post the code to GitHub soon, I just have to get clearance from my workplace to release it. I will edit this post to include the link when I have.
We have usually kept the Access front ends on network drives, and just put up with the need to compact and repair on a regular basis. You will probably find you need to do that even when they are installed locally, anyway.
If you must have it installed locally, there are various tools which will enable you to "push out" software updates, and the guys over on ServerFault would have more information on those. Assuming such tools aren't available, the only other option I can think of is to write a small loader program that checks the local .MDB against a master copy on the server, and re-copies it across if they are different, before then launching the MDB.