558789 - Allow to build against system libevent

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Mike Hommey [:glandium]]

Attachment: Patch

Everything in $SUMMARY.

Comment 1

[Mike Hommey [:glandium]]

Comment on attachment 438481 [details] [diff] [review]
Patch

This patch is against trunk, but I also have an equivalent against 1.9.2.3plugin1 if necessary

Comment 2

[Benjamin Smedberg]

Comment on attachment 438481 [details] [diff] [review]
Patch

cjones, did we make any local customizations of libevent, or are there known versioning issues where this would be a terrible idea?

Comment 3

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

We (Mozilla) haven't touched libevent.  Here's what chromium claims to have done
-----
$ cat src/third_party/libevent/README.chromium 
This contains a copy of libevent-1.4.7-stable for use on Linux and Mac

Rather than use libevent's own build system, we just build a Chrome 
static library using SConscript

Changes from upstream:
1) add linux/config.h from a run of configure on a Linux development box.
2) add SConscript and libevent.gyp
3) quotes.patch is applied to allow libevent to be used without
being installed (alternative is adding crazy -I arguments to the commandline)
4) The directories WIN32-Code and WIN32-Prj are not included
5) add mac/config.h from a run of configure on a Mac.
6) sys-queue-macros.patch is applied to build on systems which are missing a
complete sys/queue.h

SConscript does not include a few files that don't build on Linux:
    devpoll.c, evport.c, kqueue.c

TODO: currently we -I the source directory, which exposes event-config.h, which
is dangerous as this file was generated for Linux and is not entirely correct on
other platforms.
-----
We'd have the same kinds of versioning issues with libevent as we would building with system cairo e.g., except that libevent's API should be stabler.  And, chromium isn't using anything exotic.

I don't see any problems with this.

Comment 4

[Mike Hommey [:glandium]]

http://hg.mozilla.org/mozilla-central/rev/cccd01dbb2a8

Comment 5

[LegNeato]

Comment on attachment 438481 [details] [diff] [review]
Patch

a=LegNeato for 1.9.2.5. Please *only* land it on mozilla-1.9.2 default, as we are still working on the 3.6.4 relbranch and do not want this landed there

Comment 6

[Mike Hommey [:glandium]]

http://hg.mozilla.org/releases/mozilla-1.9.2/rev/4b8165a8d602

Comment 7

[Mike Hommey [:glandium]]

For the record, this fixup was needed on m-c:
http://hg.mozilla.org/mozilla-central/rev/5f229488969c
which I obviously should have mentioned here, such that I wouldn't have had the same problem again:
http://hg.mozilla.org/releases/mozilla-1.9.2/rev/a1bffe3814de