From: zhangxiao <xiao.zhang@windriver.com>
To: <openembedded-devel@lists.openembedded.org>
Cc: "Bruce.Ashfield" <bruce.ashfield@windriver.com>
Subject: How to deal with complex patch including creating new patches?
Date: Fri, 25 Apr 2014 14:25:32 +0800 [thread overview]
Message-ID: <5359FFDC.1000206@windriver.com> (raw)
Hi Experts,
As we know, usually we have to add patches on source tree and in most cases
what we need to do is just add the source of the patch into SRC_URI list
and bitbake will install it automatically in "patch" task.
While for some complex cases, for example the netcat as below:
http://cgit.openembedded.org/cgit.cgi/meta-openembedded/tree/meta-networking/recipes-support/netcat/netcat-openbsd_1.105.bb?h=master
...
8 SRC_URI =
"${DEBIAN_MIRROR}/main/n/netcat-openbsd/netcat-openbsd_${PV}.orig.tar.gz;name=netcat
\
9
${DEBIAN_MIRROR}/main/n/netcat-openbsd/netcat-openbsd_${PV}-7.debian.tar.gz;name=netcat-patch"
...
It has two sources: one source tree as
"netcat-openbsd_${PV}.orig.tar.gz"and the other is a patch
called "netcat-openbsd_${PV}-7.debian.tar.gz".
While the patch is not simple as usual that beside common modifications
on source tree,
it also creates several patches that also need to be installed. To deal
with it, this
bb file make it in do_compile in line 22 as below:
...
20 do_compile() {
21 cd ${S}
22 while read line; do patch -p1 <
${WORKDIR}/debian/patches/$line; done < ${WORKDIR}/debian/patches/series
23 pkgrel=4
24 oe_runmake CFLAGS="$CFLAGS
-DDEBIAN_VERSION=\"\\\"${pkgrel}\\\"\""
25 }
...
Of course it is not good and will cause error when we force re-compile it
since those patches had already been installed in the previous compile task.
The how to fix it?
Currently I can find two methods as below:
First, for line 21, 22 of do_compile that install new created patches,
remove them.
Then:
A) Add a do_patch_append and install those new created patches to this
function;
B) Convert the netcat-openbsd_${PV}-7.debian.tar.gz to a pure patch. Let it
contains all modifications and no need to create more patches any more. Then
place this new created patch locally and modify the SRC_URI to point to it.
Compare with A and B, which one is advised here? Why? Any suggestions
and comments is helpful!
Of course, we can also leave the patch installation in do_compile but just
add a flag to make it. But that is really not a good method. :-)
Thanks
Xiao
next reply other threads:[~2014-04-25 6:25 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-25 6:25 zhangxiao [this message]
2014-04-25 7:46 ` How to deal with complex patch including creating new patches? Koen Kooi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5359FFDC.1000206@windriver.com \
--to=xiao.zhang@windriver.com \
--cc=bruce.ashfield@windriver.com \
--cc=openembedded-devel@lists.openembedded.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.