summaryrefslogtreecommitdiff
path: root/HACKING
diff options
context:
space:
mode:
authorThomas Vander Stichele <thomas@apestaart.org>2004-02-13 13:04:54 +0000
committerThomas Vander Stichele <thomas@apestaart.org>2004-02-13 13:04:54 +0000
commit8ba3ae3970c2fa6a0912ecfe2b102f1d2007ea9b (patch)
tree20b425329805de9a51ed113653ce63f408b05cfe /HACKING
parentc46ba9cf8deb3bd85b090d762f775041c06e8ac7 (diff)
fixing up module
Original commit message from CVS: fixing up module
Diffstat (limited to 'HACKING')
-rw-r--r--HACKING36
1 files changed, 36 insertions, 0 deletions
diff --git a/HACKING b/HACKING
new file mode 100644
index 0000000..ba064f6
--- /dev/null
+++ b/HACKING
@@ -0,0 +1,36 @@
+What we are trying to achieve:
+
+satisfy:
+ patching of CVS checkout using our patch files placed in our CVS
+
+ passing of
+ make
+ make distcheck
+ non-srcdir build (ie, mkdir build; cd build; ../configure; make)
+
+How it works:
+
+* configure checks whether or not it should update ffmpeg from CVS by looking
+ at the nano version number
+ - if it's 1, we're in cvs mode, and it should check it out
+ - if it's not 1, we're in prerel or rel mode, and the code should already
+ be on disk
+ FIXME: we could change this to really check out the source code if some
+ required files aren't there just in case someone checks out from CVS
+ but CVS is not at nano 1
+
+* patching of the checked-out copy happens at
+
+Axioms under which we work:
+- the dist tarball needs to include either
+ - the pristine ffmpeg checkout + our patches + a patch mechanism on make
+ or
+ - the ffmpeg checkout with patches already applied
+
+- configure/make is not allowed to touch files that already live in the source
+ tree; if they need to then they need to be copied first and cleaned
+ afterward
+
+- it would be very nice if, on update of either the Tag file or the patch set,
+ make would know exactly what to do with it.
+