Michael Kremser wrote:
Neil wrote:
make -j2 (use larger numbers if you have several processors) makes compiling a new tree faster.
:-( I only have one processor in my machine.
So do I, but -j2 is still faster.
make -s makes make faster by reducing the amount of output. Or you could run make from a hardware terminal instead of gnome-terminal.
Hm, I'm gonna try this out. What do you mean by "hardware terminal"?
I think if you press Ctrl+Alt+F2-7 you get alternate console where you can log in and run text commands. Ctrl+Alt+F1 should give you the system console and Ctrl+Alt+F7 should give you your X terminal.
make chrome only makes the XUL and JS. You can optimise this further by running make realchrome from the folder containing the parent jar.mn of the file(s) you are changing (or its objdir equivalent, if you are using one).
Doesn't work. If I command make chrome, the following appears:
[EMAIL PROTECTED] mozilla]$ make -f client.mk chrome make: F�r das Ziel �chrome� ist nichts zu tun.
That means "Nothing to do for destination �chrome�". I also tried "editor" instead of "chrom", but no success at all.
No -f client.mk, just make chrome. -j2 doesn't work with -f client.mk either, although -s does.
You can configure your build for flat chrome or both. This speeds up make by only copying changed files.
Have I understood that right... flat means, Mozilla does not extract the XUL from the JAR archive but takes the corresponding file from the file system, so I can edit it on the fly and don't have to rebuild Mozilla or at least rebuild the archive?
It means that the files are copied from the CVS versions to the dist/bin/chrome subfolders. Then Mozilla reads them from the folders, rather than from the archives.
What does "both" mean?
This means that the folders are there, even though Mozilla uses the archives. Normally to refresh an archive all of the files are copied from the CVS versions to the dist/bin/chrome subfolders each time you make chrome. This way, although Mozilla continues to use the archives, only update files need to be copied.
As you're on Linux, you can configure your build to use symlink chrome format. This means that in most cases you don't have to remake for XUL/JS changes to take effect. If you also disable the XUL cache you don't even need to restart Mozilla, just close and reopen the window.
How can I disable the XUL cache? Where takes Mozilla the files from, then?
If you enable the XUL cache then Mozilla takes the files from XUL.mfasl, if you disable it then it takes it from the archives/files. Note that only when you restart Mozilla does it checks the timestamps to see if the XUL cache needs to be cleared. You disable the cache in Preferences Debug Networking.
How to change chrome format in an existing tree:
First, edit your .mozconfig with ac_add_options --enable-chrome-format=[jar|flat|both|symlink] and run configure, or run configure with the --enable-chrome-format=[jar|flat|both|symlink] option.
configure is not recognized by bash.
Your path probably doesn't include . so use ./configure instead.
My .mozconfig.mk file has currently this content:
As the file suggests, do not edit it. Create .mozconfig or ~/.mozconfig instead.
If you are changing to or from symlink chrome, I recommend you delete your entire dist/bin/chrome, although deleting dist/bin/chrome/installed-chrome.txt may suffice, or for changing between flat and symlink chrome you may find that just make chrome suffices.
If I delete dist/bin/chrome, where from does Mozilla take the files?
make chrome recopies them from the CVS versions.
-- Warning: May contain traces of nuts. _______________________________________________ mozilla-editor mailing list [EMAIL PROTECTED] http://mail.mozilla.org/listinfo/mozilla-editor
