Improved wget

Improved wget has two patches with reselect to CVS version of wget on 2004.11.11:
1) support of downloading files larger than 2Gb;
2) support of an option which forces wget re-try downloading even if some errors, like connection refusal occurred.

Files to download

Patch with respect to CVS version of 2004.11.13

wget-pet-20041113.tar.bz2 -- tarball of the patched version.

Description:

  1. Large file support. As of 2004.11.13, the latest official version of wget does not allow download files of size more than 2Gb. While development of large file support in the official version of wget is underway, I could not wait and hacked the code myself. I've succeeded in forcing wget to download correctly large files by ftp via network which frequently drops connection. Tested under Linux, Sun and HP-UX.

    Comments:

    • The ftp server from which you are going to download the file should support large files. Many ftp servers do not support it at all, or do not support it by default.

    • Your operating system and your file system should support files larger then 2Gb.

    • Web server Apache 2.0.51 and older does not support large files, for this reason even patched wget is not able to retrieve large files. So far, I know that only cherokee Web server provides limited large file support. This patch was successfully tested for downloading a large file from cherokee Web server, but re-try feature was not tested, since cherokee did not support at the moment of testing restarts.

  2. --persistent option. Without this option, by default, wget considers some errors like failure to connect, failure to authenticate as fatal, i.e. terminates without attempts to re-try. With this option wget will re-try to connect or log in. This option is useful only in conjunction with --wait or --waitretry. This feature is very useful when you download files regularly as a cron job from a remote host which can temporarily shutdown the http or ftp daemon processes for maintenance. Without --persistent option wget will terminate and give up further attempts, since the remote host will refuse connection. In this situation wget with --persistent option will try to re-connect after waiting --wait time.


Last update: 14-NOV-2004 09:24:12
Leonid Petrov