How to obtain and compile wine, and problems that may arise...
If you are going to compile Wine, either to use the most recent code possible or to improve it, then the first thing to do is to obtain a copy of the source code. We'll cover how to retrieve and compile the official source releases from the FTP archives, and also how to get the cutting edge up-to-the-minute fresh Wine source code from CVS (Concurrent Versions System). Both processes of source code installation are similar, and once you master one, you should have no trouble dealing with the other one.
You may also need to know how to apply a source code patch to your version of Wine. Perhaps you've uncovered a bug in Wine, reported it to the Wine mailing list, and received a patch from a developer to hopefully fix the bug. We will show you how to safely apply the patch and revert it if the patch doesn't work.
The safest way to grab the source is from one of the official FTP archives. An up to date listing is in the ANNOUNCE file in the Wine distribution (which you would have if you already downloaded it). Here is a list of FTP servers carrying Wine:
ftp://ftp.infomagic.com/pub/mirrors/linux/sunsite/ALPHA/wine/development/
ftp://ftp.fu-berlin.de/unix/linux/mirrors/sunsite.unc.edu/ALPHA/wine/development/
The official releases are tagged by date with the format "Wine-YYYYMMDD.tar.gz". Your best bet is to grab the latest one.
Once you have downloaded this, you must first compile wine, and then install it. This is not very hard to do. First switch to the directory containing the file you just downloaded. Then extract the source with (e.g.):
# tar xzvf wine-20021031.tar.gz |
Then, switch to the directory that was created and compile it by typing (e.g.):
# ./tools/wineinstall |
NOTE: You must make sure that you are not the superuser (root) when doing this, and that you have write permission to the directory that was created by the tar command as well as all of its subdirectories and files..
The official web page for Wine CVS is http://www.winehq.com/development/.
First, you need to get a copy of the latest Wine sources using CVS. You can tell it where to find the source tree by setting the CVSROOT environment variable. You also have to log in anonymously to the wine CVS server. In bash, it might look something like this:
$ export CVSROOT=:pserver:cvs@cvs.winehq.com:/home/wine $ cvs login Password: $ cvs checkout wine |
That'll pull down the entire Wine source tree from winehq.com and place it in the current directory (actually in the 'wine' subdirectory). CVS has a million command line parameters, so there are many ways to pull down files, from anywhere in the revision history. Later, you can grab just the updates:
$ cvs -PAd update |
cvs update works from inside the source tree. You don't need the CVSROOT environment variable to run it either. You just have to be inside the source tree. The -P, -A and -d options make sure your local Wine tree directory structure stays in sync with the remote repository.
After you've made changes, you can create a patch with cvs diff -u, which sends output to stdout (the -u controls the format of the patch). So, to create an my_patch.diff file, you would do this:
$ cvs diff -u >my_patch.diff |
You can call cvs diff from anywhere in the tree (just like cvs update), and it will always grab recursively from that point. You can also specify single files or subdirectories:
$ cvs diff -u dlls/winaspi >my_aspi_patch.diff |
Experiment around a little. It's fairly intuitive.
If you have the Wine source code, as opposed to a binary distribution, you have the option of applying patches to the source tree to fix bugs and add experimental features. Perhaps you've found a bug, reported it to the Wine mailing list, and received a patch file to fix the bug. You can apply the patch with the patch command, which takes a streamed patch from stdin:
$ cd wine $ patch -p0 <../patch_to_apply.diff |
To remove the patch, use the -R option:
$ patch -p0 -R <../patch_to_apply.diff |
If you want to do a test run to see if the patch will apply successfully (e.g., if the patch was created from an older or newer version of the tree), you can use the --dry-run parameter to run the patch without writing to any files:
$ patch -p0 --dry-run <../patch_to_apply.diff |
patch is pretty smart about extracting patches from the middle of a file, so if you save an email with an inlined patch to a file on your hard drive, you can invoke patch on it without stripping out the email headers and other text. patch ignores everything that doesn't look like a patch.
The -p0 option to patch tells it to keep the full file name from the patch file. For example, if the file name in the patch file was wine/programs/clock/main.c. Setting the -p0 option would apply the patch to the file of the same name i.e. wine/programs/clock/main.c . Setting the -p1 option would strip off the first part of the file name and apply the patch instead to programs/clock/main.c . The -p1 option would be useful if you named your top level wine directory differently to the person who sent you the patch. For the -p1 option patch should be run from the top level wine directory.