File copy over HTTP

[Pramod]

Hello All,

I want to copy few files from a server using HTTP URL in C++.

for ex. I have a file located at: http://localhost/myFolder/myFile.txt
which needs to copied at a specific folder on my client machine. The C+
+ program would be running on client machine and has access to this
URL file.

Any pointers in this regards are appreciated.

Thanks,
PSI

 

[Sjouke Burry]

Hello All,

I want to copy few files from a server using HTTP URL in C++.

for ex. I have a file located at: http://localhost/myFolder/myFile.txt
which needs to copied at a specific folder on my client machine. The C+
+ program would be running on client machine and has access to this
URL file.

Any pointers in this regards are appreciated.

Thanks,
PSI

Get hold of "wget.exe" a gnu utility.
then in your program do:
system("wget http://localhost/myFolder/myFile.txt");
which puts it in your current working directory.

 

[Brian Wood]

Hello All,

I want to copy few files from a server using HTTP URL in C++.

for ex. I have a file located at:http://localhost/myFolder/myFile.txt
which needs to copied at a specific folder on my client machine. The C+
+ program would be running on client machine and has access to this
URL file.

Any pointers in this regards are appreciated.

Thanks,
PSI

http://sourceforge.net/projects/urdl/


Brian Wood
http://webEbenezer.net
(651) 251-9384

 

[Sousuke]

Get hold of "wget.exe" a gnu utility.
then in your program do:
        system("wgethttp://localhost/myFolder/myFile.txt");
which puts it in your current working directory.

IMO, using the system function for things like this is rather hackish.
Things like this should be handled in the program or by some library
or API.

If the OP is indeed on Windows, then there's a native API for that:
WinHTTP:

http://msdn.microsoft.com/en-us/library/aa382925(VS.85).aspx

 

[Jorgen Grahn]

On Mar 25, 6:52 am, Sjouke Burry <[email protected]>
wrote:

IMO, using the system function for things like this is rather hackish.
Things like this should be handled in the program or by some library
or API.

The exact code above is hackish, yes: for example he has no direct
control over where the file ends up (if myFile.txt already exists wget
by default saves the data as myFile.txt.1) and it doesn't check if it
succeeded. Might even be a security hole in some cases.

But in many other cases reinventing the wheel or even using a library
would look odd. For example it's fairly normal for Unix programs to
read gzipped files by popen()ing gzip -cd.

/Jorgen

 

[Sjouke Burry]

The exact code above is hackish, yes: for example he has no direct
control over where the file ends up (if myFile.txt already exists wget
by default saves the data as myFile.txt.1) and it doesn't check if it
succeeded. Might even be a security hole in some cases.

But in many other cases reinventing the wheel or even using a library
would look odd. For example it's fairly normal for Unix programs to
read gzipped files by popen()ing gzip -cd.

/Jorgen

??? You dont have control over your current working directory??
Strange...

 

[Jorgen Grahn]

??? You dont have control over your current working directory??
Strange...

I assume you are talking about the myFile.txt versus myFile.txt.1
issue I mentioned above?

No, my /program/ sure doesn't have control over that. And I don't feel
like explaining things like "the program will silently fail if you
have a file named so-and-so in your current directory" in the user
documentation.

So it's not strange to me. It's part of building robust, secure and
usable programs.

(It's not an issue if I'm the only user and will only use the program
for a few days, but that's not normally the case for C++ programs.)

/Jorgen