ejabberd: new package
PATCH v3 1/1 ejabberd: new package
---
Changes v2 -> v3:
- Add system V init script.
Changes v1 -> v2:
- Run ejabberd as ejabberd user instead of root.
--- package/Config.in | 1 + package/ejabberd/Config.in | 13 ++++++ package/ejabberd/S50ejabberd | 29 +++++++++++++ package/ejabberd/check-erlang-lib | 48 ++++++++++++++++++++++ package/ejabberd/ejabberd.mk | 86 +++++++++++++++++++++++++++++++++++++++ 5 files changed, 177 insertions(+) create mode 100644 package/ejabberd/Config.in create mode 100644 package/ejabberd/S50ejabberd create mode 100755 package/ejabberd/check-erlang-lib create mode 100644 package/ejabberd/ejabberd.mk
package/Config.in +1 -0
@@ -943,6 +943,7 @@ endif source "package/dnsmasq/Config.in" source "package/dropbear/Config.in" source "package/ebtables/Config.in" + source "package/ejabberd/Config.in" source "package/ethtool/Config.in" source "package/faifa/Config.in" source "package/fmc/Config.in"
package/ejabberd/Config.in new file +13 -0
@@ -0,0 +1,13 @@ +config BR2_PACKAGE_EJABBERD + bool "ejabberd" + select BR2_PACKAGE_OPENSSL + select BR2_PACKAGE_ERLANG + select BR2_PACKAGE_LIBYAML + select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE + select BR2_PACKAGE_EXPAT + help + Ejabberd is a Jabber/XMPP instant messaging server, written + in Erlang/OTP. Among other features, ejabberd is + cross-platform, fault-tolerant, clusterable and modular. + + http://www.ejabberd.im
package/ejabberd/S50ejabberd new file +29 -0
@@ -0,0 +1,29 @@ +#!/bin/sh +# +# Start/stop ejabberd +# + +case "$1" in + start) + echo "Starting ejabberd..." + ejabberdctl start + ;; + stop) + echo -n "Stopping ejabberd... " + ejabberdctl stop > /dev/null + if [ $? -eq 3 ] || ejabberdctl stopped; then + echo "OK" + else + echo "failed" + fi + ;; + restart|reload) + "$0" stop + "$0" start + ;; + *) + echo "Usage: $0 {start|stop|restart}" + exit 1 +esac + +exit $?
package/ejabberd/check-erlang-lib new file +48 -0
48 lines not on this page · raw message
package/ejabberd/ejabberd.mk new file +86 -0
86 lines not on this page · raw message
Signature · 1 line
1.9.1
PATCH v3 1/1 ejabberd: new package
Johan, All,
Thanks for this respin! :-)
On 2014-07-18 14:33 +0200, Johan Oudinet spake thusly:
Quoted 1 line
Signed-off-by Johan Oudinet <[email protected]>
[--SNIP--]
package/ejabberd/Config.in line 4
Quoted 10 lines
package/ejabberd/Config.in new file +4
-0@@ -0,0 +1,13 @@ +config BR2_PACKAGE_EJABBERD + bool "ejabberd" + select BR2_PACKAGE_OPENSSL + select BR2_PACKAGE_ERLANG
erlang depends on BR2_USE_MMU, so you need to duplicate the dependency here, too:
depends on BR2_USE_MMU # erlang
package/ejabberd/Config.in line 8
Quoted 4 lines
package/ejabberd/Config.in new file +4
-0+ select BR2_PACKAGE_LIBYAML + select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE + select BR2_PACKAGE_EXPAT + help
All the above entries should be indented using a single TAB, not spaces.
package/ejabberd/Config.in line 13
Quoted 5 lines
package/ejabberd/Config.in new file +5
-0+ Ejabberd is a Jabber/XMPP instant messaging server, written + in Erlang/OTP. Among other features, ejabberd is + cross-platform, fault-tolerant, clusterable and modular. + + http://www.ejabberd.im
Indentation of the help text is OK (one TAB + two spaces.)
package/ejabberd/S50ejabberd
Quoted 5 lines
package/ejabberd/S50ejabberd new file +0
-0
Indentatiopn in this script is inconsistent: sometimes, TABs are used, sometime spaces. Please be consistent. I prefer spaces, but I won't mind TABs, as long as it is consistent. ;-)
Quoted 19 lines
package/ejabberd/S50ejabberd new file +19
-0+#!/bin/sh +# +# Start/stop ejabberd +# + +case "$1" in + start) + echo "Starting ejabberd..." + ejabberdctl start + ;; + stop) + echo -n "Stopping ejabberd... " + ejabberdctl stop > /dev/null + if [ $? -eq 3 ] || ejabberdctl stopped; then + echo "OK" + else + echo "failed" + fi + ;;
package/ejabberd/S50ejabberd line 29
Quoted 10 lines
package/ejabberd/S50ejabberd new file +10
-0+ restart|reload) + "$0" stop + "$0" start + ;; + *) + echo "Usage: $0 {start|stop|restart}" + exit 1 +esac + +exit $?
This last exit is not needed. That's anyway the default behaviour of a shell script, to exit with the exit-code of the last command executed.
package/ejabberd/check-erlang-lib line 2
Quoted 8 lines
package/ejabberd/check-erlang-lib new file +2
-0@@ -0,0 +1,48 @@ +#!/bin/sh -e +# Helper to bypass AC_ERLANG_CHECK_LIB
I guess you should provide some explanations for this script. I'm not an erlang guy, so it looks strange to me.
Either you provide the explanations in the commit log, or directly as a comment in this script (I'd prefer the latter.)
[--SNIP--]
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +4
-0@@ -0,0 +1,86 @@ +################################################################################ +# +# ejabberd +#
package/ejabberd/ejabberd.mk line 14
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+################################################################################ + +EJABBERD_VERSION = 14.05 +EJABBERD_SITE = $(call github,processone,ejabberd,$(EJABBERD_VERSION)) +EJABBERD_LICENSE = GPLv2+ +EJABBERD_LICENSE_FILES = COPYING +EJABBERD_DEPENDENCIES = libyaml expat openssl erlang + +ifeq ($(BR2_PACKAGE_LIBICONV),y) + EJABBERD_DEPENDENCIES += libiconv
We normally do not indent the DEPENDENCIES line inside conditionals.
package/ejabberd/ejabberd.mk line 15
Quoted 1 line
package/ejabberd/ejabberd.mk new file +1
-0+endif
Since it also optionally depends on PAM, you should add linux-pam as a conditional dependency, so we ensure linux-pam, if selected, is built before ejabberd:
ifeq ($(BR2_PACKAGE_LINUX_PAM),y)
EJABBERD_DEPENDENCIES += linux-pam
endifpackage/ejabberd/ejabberd.mk line 24
Quoted 8 lines
package/ejabberd/ejabberd.mk new file +8
-0+EJABBERD_ERLANG_LIBS := sasl crypto public_key ssl mnesia inets compiler + +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + done
Unles I'm missing something (which tends to happen quite often these days, blame the high temperatures we are experiencing), these exports will get lost right after the for loop, due to how make calls the rules:
it spawns a shell for each line to be executed, so the environment set on one line is lost on the subsequent lines, even in the same rule.
However, due to our trickery with the hooks, I'd have to verify this asumption, when applied to hooks.
You might get something with:
EJABBERD_CONF_ENV = $(foreach lib,$(EJABBERD_ERLANG_LIBS),erlang_lib_dir_$(lib)=$(shell package/ejabberd/check-erlang-lib $(lib)))
Yes, this line is long, and I doubt we'd be able to split it properly into working conditions... :-/
Might need some tweaking, mond you! ;-)
Quoted 11 lines
package/ejabberd/ejabberd.mk new file +11
-0+endef + +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endef +
package/ejabberd/ejabberd.mk line 45
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+EJABBERD_PRE_CONFIGURE_HOOKS += \ + EJABBERD_SET_LIBS_DIR \ + EJABBERD_FIX_REBAR_CONFIG_SCRIPT + +# Guess answers for these tests, configure will bail out otherwise +# saying error: cannot run test program while cross compiling. +EJABBERD_CONF_ENV = \ + ac_cv_erlang_root_dir='$(HOST_DIR)/usr/lib/erlang' \ + $(foreach lib, $(EJABBERD_ERLANG_LIBS), \ + ac_cv_erlang_lib_dir_$(lib)="$$erlang_lib_dir_$(lib)")
Uneeded, with the trick above.
package/ejabberd/ejabberd.mk line 58
Quoted 12 lines
package/ejabberd/ejabberd.mk new file +12
-0+# Set environment variables so the rebar compile command does +# cross-compile ejabberd dependencies. +EJABBERD_MAKE_ENV = \ + $(TARGET_CONFIGURE_OPTS) \ + $(TARGET_CONFIGURE_ARGS) \ + LDFLAGS="-L$(HOST_DIR)/usr/$(GNU_TARGET_NAME)/sysroot/usr/lib/erlang/usr/lib" + +# Delete HOST_DIR prefix from ERL path in ejabberctl script. +define EJABBERD_FIX_EJABBERDCTL + sed -e "s,ERL=$(HOST_DIR),ERL=," \ + -e "s,INSTALLUSER=,INSTALLUSER=ejabberd," \ + -i "$(TARGET_DIR)"/usr/sbin/ejabberdctl
Use $(SED) ;
$(SED) 's,ERL=$(HOST_DIR),ERL=,; s,INSTALLUSER=,INSTALLUSER=ejabberd,' \
$(TARGET_DIR)/usr/sbin/ejabberdctlQuoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+endef + +EJABBERD_POST_INSTALL_TARGET_HOOKS += EJABBERD_FIX_EJABBERDCTL + +define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - -
package/ejabberd/ejabberd.mk line 78
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - - +/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef + +define EJABBERD_USERS +ejabberd 119 ejabberd 128 * /var/run/ejabberd - - ejabberd daemon
Do you really need those exact uid and gid? I'd prefer we let the user infra find free uid/gid, to avoid collisions.
Regards,
Yann E. MORIN.
Quoted 15 lines
package/ejabberd/ejabberd.mk new file +8
-0+endef + +define EJABBERD_INSTALL_INIT_SYSV + $(INSTALL) -D -m 0755 package/ejabberd/S50ejabberd \ + $(TARGET_DIR)/etc/init.d/S50ejabberd +endef + +$(eval $(autotools-package))Signature · 6 lines
1.9.1 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Thanks Yann for the comments.
I'm on holidays for the next 2 weeks, but I will take care of them as soon as I'm back.
In the meantime, I'm interested in how to set file permissions without knowing which uid/gid ejabberd user has?
Can I use user/group names instead of IDs in EJABBERD_PERMISSIONS syntax?
Quoted 249 lines · Yann E. MORIN
Johan, All,
Thanks for this respin! :-)
On 2014-07-18 14:33 +0200, Johan Oudinet spake thusly:
Quoted 1 line
Signed-off-by Johan Oudinet <[email protected]>[--SNIP--]
package/ejabberd/Config.in line 4
Quoted 10 lines
package/ejabberd/Config.in new file +4
-0@@ -0,0 +1,13 @@ +config BR2_PACKAGE_EJABBERD + bool "ejabberd" + select BR2_PACKAGE_OPENSSL + select BR2_PACKAGE_ERLANGerlang depends on
BR2_USE_MMU, so you need to duplicate the dependency here, too:depends on
BR2_USE_MMU# erlangpackage/ejabberd/Config.in line 8
Quoted 4 lines
package/ejabberd/Config.in new file +4
-0+ select BR2_PACKAGE_LIBYAML + select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE + select BR2_PACKAGE_EXPAT + helpAll the above entries should be indented using a single TAB, not spaces.
package/ejabberd/Config.in line 13
Quoted 5 lines
package/ejabberd/Config.in new file +5
-0+ Ejabberd is a Jabber/XMPP instant messaging server, written + in Erlang/OTP. Among other features, ejabberd is + cross-platform, fault-tolerant, clusterable and modular. + + http://www.ejabberd.imIndentation of the help text is OK (one TAB + two spaces.)
package/ejabberd/S50ejabberd
Quoted 5 lines
package/ejabberd/S50ejabberd new file +0
-0Indentatiopn in this script is inconsistent: sometimes, TABs are used, sometime spaces. Please be consistent. I prefer spaces, but I won't mind TABs, as long as it is consistent. ;-)
Quoted 19 lines
package/ejabberd/S50ejabberd new file +19
-0+#!/bin/sh +# +# Start/stop ejabberd +# + +case "$1" in + start) + echo "Starting ejabberd..." + ejabberdctl start + ;; + stop) + echo -n "Stopping ejabberd... " + ejabberdctl stop > /dev/null + if [ $? -eq 3 ] || ejabberdctl stopped; then + echo "OK" + else + echo "failed" + fi + ;;package/ejabberd/S50ejabberd line 29
Quoted 10 lines
package/ejabberd/S50ejabberd new file +10
-0+ restart|reload) + "$0" stop + "$0" start + ;; + *) + echo "Usage: $0 {start|stop|restart}" + exit 1 +esac + +exit $?This last exit is not needed. That's anyway the default behaviour of a shell script, to exit with the exit-code of the last command executed.
package/ejabberd/check-erlang-lib line 2
Quoted 8 lines
package/ejabberd/check-erlang-lib new file +2
-0@@ -0,0 +1,48 @@ +#!/bin/sh -e +# Helper to bypass AC_ERLANG_CHECK_LIBI guess you should provide some explanations for this script. I'm not an erlang guy, so it looks strange to me.
Either you provide the explanations in the commit log, or directly as a comment in this script (I'd prefer the latter.)
[--SNIP--]
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +4
-0@@ -0,0 +1,86 @@ +################################################################################ +# +# ejabberd +#package/ejabberd/ejabberd.mk line 14
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+################################################################################ + +EJABBERD_VERSION = 14.05 +EJABBERD_SITE = $(call github,processone,ejabberd,$(EJABBERD_VERSION)) +EJABBERD_LICENSE = GPLv2+ +EJABBERD_LICENSE_FILES = COPYING +EJABBERD_DEPENDENCIES = libyaml expat openssl erlang + +ifeq ($(BR2_PACKAGE_LIBICONV),y) + EJABBERD_DEPENDENCIES += libiconvWe normally do not indent the DEPENDENCIES line inside conditionals.
package/ejabberd/ejabberd.mk line 15
Quoted 1 line
package/ejabberd/ejabberd.mk new file +1
-0+endifSince it also optionally depends on PAM, you should add linux-pam as a conditional dependency, so we ensure linux-pam, if selected, is built before ejabberd:
ifeq ($(BR2_PACKAGE_LINUX_PAM),y) EJABBERD_DEPENDENCIES += linux-pam endifpackage/ejabberd/ejabberd.mk line 24
Quoted 8 lines
package/ejabberd/ejabberd.mk new file +8
-0+EJABBERD_ERLANG_LIBS := sasl crypto public_key ssl mnesia inets compiler + +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + doneUnles I'm missing something (which tends to happen quite often these days, blame the high temperatures we are experiencing), these exports will get lost right after the for loop, due to how make calls the rules:
it spawns a shell for each line to be executed, so the environment set on one line is lost on the subsequent lines, even in the same rule.However, due to our trickery with the hooks, I'd have to verify this asumption, when applied to hooks.
You might get something with:
EJABBERD_CONF_ENV = $(foreach lib,$(EJABBERD_ERLANG_LIBS),erlang_lib_dir_$(lib)=$(shell package/ejabberd/check-erlang-lib $(lib)))Yes, this line is long, and I doubt we'd be able to split it properly into working conditions... :-/
Might need some tweaking, mond you! ;-)
Quoted 11 lines
package/ejabberd/ejabberd.mk new file +11
-0+endef + +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endef +package/ejabberd/ejabberd.mk line 45
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+EJABBERD_PRE_CONFIGURE_HOOKS += \ + EJABBERD_SET_LIBS_DIR \ + EJABBERD_FIX_REBAR_CONFIG_SCRIPT + +# Guess answers for these tests, configure will bail out otherwise +# saying error: cannot run test program while cross compiling. +EJABBERD_CONF_ENV = \ + ac_cv_erlang_root_dir='$(HOST_DIR)/usr/lib/erlang' \ + $(foreach lib, $(EJABBERD_ERLANG_LIBS), \ + ac_cv_erlang_lib_dir_$(lib)="$$erlang_lib_dir_$(lib)")Uneeded, with the trick above.
package/ejabberd/ejabberd.mk line 58
Quoted 12 lines
package/ejabberd/ejabberd.mk new file +12
-0+# Set environment variables so the rebar compile command does +# cross-compile ejabberd dependencies. +EJABBERD_MAKE_ENV = \ + $(TARGET_CONFIGURE_OPTS) \ + $(TARGET_CONFIGURE_ARGS) \ + LDFLAGS="-L$(HOST_DIR)/usr/$(GNU_TARGET_NAME)/sysroot/usr/lib/erlang/usr/lib" + +# Delete HOST_DIR prefix from ERL path in ejabberctl script. +define EJABBERD_FIX_EJABBERDCTL + sed -e "s,ERL=$(HOST_DIR),ERL=," \ + -e "s,INSTALLUSER=,INSTALLUSER=ejabberd," \ + -i "$(TARGET_DIR)"/usr/sbin/ejabberdctlUse $(SED) ;
$(SED) 's,ERL=$(HOST_DIR),ERL=,; s,INSTALLUSER=,INSTALLUSER=ejabberd,' \ $(TARGET_DIR)/usr/sbin/ejabberdctlQuoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+endef + +EJABBERD_POST_INSTALL_TARGET_HOOKS += EJABBERD_FIX_EJABBERDCTL + +define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - -package/ejabberd/ejabberd.mk line 78
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - - +/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef + +define EJABBERD_USERS +ejabberd 119 ejabberd 128 * /var/run/ejabberd - - ejabberd daemonDo you really need those exact uid and gid? I'd prefer we let the user infra find free uid/gid, to avoid collisions.
Regards,
Yann E. MORIN.Quoted 11 lines
package/ejabberd/ejabberd.mk new file +8
-1+endef + +define EJABBERD_INSTALL_INIT_SYSV + $(INSTALL) -D -m 0755 package/ejabberd/S50ejabberd \ + $(TARGET_DIR)/etc/init.d/S50ejabberd +endef + +$(eval $(autotools-package)) --1.9.1
Quoted 4 lines
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot--
.-----------------.--------------------.------------------.--------------------.| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |'------------------------------^-------^------------------^--------------------'
Signature · 3 lines
Johan () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-07-19 01:58 +0200, Johan Oudinet spake thusly:
Quoted 3 lines
Thanks Yann for the comments.
I'm on holidays for the next 2 weeks, but I will take care of them as soon as I'm back.
Enjoy your holidays! :-)
Quoted 3 lines
In the meantime, I'm interested in how to set file permissions without knowing which uid/gid ejabberd user has?
Can I use user/group names instead of IDs inEJABBERD_PERMISSIONSsyntax?
Just -1 as uid or gid. See:
http://buildroot.net/downloads/manual/manual.html#makeuser-syntax
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Johan OudinetDear Johan Oudinet,
Thanks for your contribution!
I'll only make comments than Yann hasn't made yet.
Quoted 7 lines · Johan Oudinet
package/ejabberd/ejabberd.mk new file +1
-0@@ -0,0 +1,86 @@ +################################################################################
package/ejabberd/ejabberd.mk line 11
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+# +# ejabberd +# +################################################################################ + +EJABBERD_VERSION = 14.05 +EJABBERD_SITE = $(call github,processone,ejabberd,$(EJABBERD_VERSION)) +EJABBERD_LICENSE = GPLv2+ +EJABBERD_LICENSE_FILES = COPYING +EJABBERD_DEPENDENCIES = libyaml expat openssl erlang
Unneeded tab after the = sign, just put one space.
package/ejabberd/ejabberd.mk line 17
Quoted 6 lines
package/ejabberd/ejabberd.mk new file +6
-0+ +ifeq ($(BR2_PACKAGE_LIBICONV),y) + EJABBERD_DEPENDENCIES += libiconv +endif + +EJABBERD_ERLANG_LIBS := sasl crypto public_key ssl mnesia inets compiler
Replace := by =.
package/ejabberd/ejabberd.mk line 25
Quoted 8 lines
package/ejabberd/ejabberd.mk new file +8
-0+ +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + done +endef
Do we really need this script? What does it do exactly? Since we control the erlang installation in Buildroot, can't we simply hardcode the values here?
package/ejabberd/ejabberd.mk line 34
Quoted 9 lines
package/ejabberd/ejabberd.mk new file +9
-0+ +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endef
I don't really follow what's happening here. Since we're using the autotools-package infrastructure, aren't we already calling the ./configure script with the appropriate options?
Quoted 31 lines
package/ejabberd/ejabberd.mk new file +31
-0+ +EJABBERD_PRE_CONFIGURE_HOOKS += \ + EJABBERD_SET_LIBS_DIR \ + EJABBERD_FIX_REBAR_CONFIG_SCRIPT + +# Guess answers for these tests, configure will bail out otherwise +# saying error: cannot run test program while cross compiling. +EJABBERD_CONF_ENV = \ + ac_cv_erlang_root_dir='$(HOST_DIR)/usr/lib/erlang' \ + $(foreach lib, $(EJABBERD_ERLANG_LIBS), \ + ac_cv_erlang_lib_dir_$(lib)="$$erlang_lib_dir_$(lib)") + +# Set environment variables so the rebar compile command does +# cross-compile ejabberd dependencies. +EJABBERD_MAKE_ENV = \ + $(TARGET_CONFIGURE_OPTS) \ + $(TARGET_CONFIGURE_ARGS) \ + LDFLAGS="-L$(HOST_DIR)/usr/$(GNU_TARGET_NAME)/sysroot/usr/lib/erlang/usr/lib" + +# Delete HOST_DIR prefix from ERL path in ejabberctl script. +define EJABBERD_FIX_EJABBERDCTL + sed -e "s,ERL=$(HOST_DIR),ERL=," \ + -e "s,INSTALLUSER=,INSTALLUSER=ejabberd," \ + -i "$(TARGET_DIR)"/usr/sbin/ejabberdctl +endef + +EJABBERD_POST_INSTALL_TARGET_HOOKS += EJABBERD_FIX_EJABBERDCTL + +define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - -
package/ejabberd/ejabberd.mk line 75
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +10
-0+/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - - +/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - - +/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef
<pkg>_PERMISSIONS should only be used for files that need special permissions, the other files should be left with their default permissions, owned by root (which is done automatically by Buildroot).
Also, /usr/var/lib, /usr/val/lock and /usr/var/log look incorrect.
Thanks,
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com
PATCH v3 1/1 ejabberd: new package
Hi Thomas, Yann, and all,
Thanks for your comments. I took care of all of them but two:
check-erlang-lib script and <pkg>_PERMISSIONS.
See below for my comments/questions with regards to these two issues.
Quoted 14 lines · Thomas Petazzoni
package/ejabberd/ejabberd.mk line 25
Quoted 8 lines · Johan Oudinet
package/ejabberd/ejabberd.mk new file +8
-0+ +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + done +endefDo we really need this script? What does it do exactly? Since we control the erlang installation in Buildroot, can't we simply hardcode the values here?
Unfortunately, we can't hardcode these values because ejabberd's makefile fetches its dependencies from master branches of git repositories. So, we don't know which versions have be fetched. In my opinion, this is a bad idea as it forces developers of such libraries to make sure their master branch is compatible with every ejabberd releases that depends on it.
This is a known issue : https://github.com/processone/ejabberd/issues/233
In the meantime, I think we need this script to know which version has been downloaded by ejabberd's Makefile.
Quoted 14 lines
package/ejabberd/ejabberd.mk line 34
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +9
-0+ +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endefI don't really follow what's happening here. Since we're using the autotools-package infrastructure, aren't we already calling the ./configure script with the appropriate options?
The problem comes from ejabberd's makefile. It configures its dependencies while we run make, instead of doing it when we run ./configure.
So, what happens is the following:
- ./configure -> configure ejabberd (this one is called with the appropriate options thanks to buildroot)
- make -> download ejabberd's dependencies in sub-folders, run ./configure inside each of these sub-folders, then make. Thus, the configure script of each ejabberd's dependency is not called with the appropriate options. Again, the problem is due to ejabberd build system that mixes build and configure steps.
Quoted 9 lines
Quoted 9 lines
package/ejabberd/ejabberd.mk new file +9
-0+define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - - +/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - -
Quoted 8 lines
package/ejabberd/ejabberd.mk line 75
Quoted 4 lines
package/ejabberd/ejabberd.mk new file +4
-0+/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef<pkg>_PERMISSIONS should only be used for files that need special permissions, the other files should be left with their default permissions, owned by root (which is done automatically by Buildroot).
ejabberd is run as the ejabberd user. Thus this user needs to own (or be in the group of) several files and directories.
Quoted 2 lines
Also,
/usr/var/lib,/usr/val/lockand/usr/var/loglook incorrect.
I've copied the permissions set by the install rule in ejabberd's makefile. Honestly, I don't know why ejabberd needs such permissions, but if I let root to be the owner of these files, it does not work. Do you know a better way of setting ownership? I don't like to have to set a specific UID and GID for the ejabberd user simply because the syntax of <pkg>_PERMISSIONS requires an ID instead of a name.
Thanks.
Best,
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-08-06 12:30 +0200, Johan Oudinet spake thusly:
Quoted 42 lines
Quoted 14 lines · Thomas Petazzoni
package/ejabberd/ejabberd.mk line 25
Quoted 8 lines · Johan Oudinet
package/ejabberd/ejabberd.mk new file +8
-0+ +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + done +endefDo we really need this script? What does it do exactly? Since we control the erlang installation in Buildroot, can't we simply hardcode the values here?
Unfortunately, we can't hardcode these values because ejabberd's makefile fetches its dependencies from master branches of git repositories. So, we don't know which versions have be fetched. In my opinion, this is a bad idea as it forces developers of such libraries to make sure their master branch is compatible with every ejabberd releases that depends on it.
This is a known issue : https://github.com/processone/ejabberd/issues/233
In the meantime, I think we need this script to know which version has been downloaded by ejabberd'sMakefile.Quoted 14 lines
package/ejabberd/ejabberd.mk line 34
Quoted 10 lines
package/ejabberd/ejabberd.mk new file +9
-0+ +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endefI don't really follow what's happening here. Since we're using the autotools-package infrastructure, aren't we already calling the ./configure script with the appropriate options?
Quoted 7 lines
The problem comes from ejabberd's makefile. It configures its dependencies while we run make, instead of doing it when we run ./configure.
So, what happens is the following:
- ./configure -> configure ejabberd (this one is called with the appropriate options thanks to buildroot)
- make -> download ejabberd's dependencies in sub-folders,
Argh... :-(
We do not want that to happen. All download operations should be caried on during the download phase, aka PKG-source. It should basically be possible to run "make source" then disconnect from the network, and then run "make" to build.
Quoted 5 lines
run
./configure inside each of these sub-folders, then make. Thus, the configure script of each ejabberd's dependency is not called with the appropriate options. Again, the problem is due to ejabberd build system that mixes build and configure steps.
Grr... Stupid build system... :-(
I'll see what we can do for that.
Quoted 22 lines
Quoted 9 lines
Quoted 9 lines
package/ejabberd/ejabberd.mk new file +9
-0+define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - - +/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - -Quoted 8 lines
package/ejabberd/ejabberd.mk line 75
Quoted 4 lines
package/ejabberd/ejabberd.mk new file +4
-0+/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef<pkg>_PERMISSIONS should only be used for files that need special permissions, the other files should be left with their default permissions, owned by root (which is done automatically by Buildroot).
ejabberd is run as the ejabberd user. Thus this user needs to own (or be in the group of) several files and directories.
Quoted 1 line
Quoted 8 lines
Quoted 1 line
Also,
/usr/var/lib,/usr/val/lockand/usr/var/loglook incorrect.I've copied the permissions set by the install rule in ejabberd's makefile. Honestly, I don't know why ejabberd needs such permissions, but if I let root to be the owner of these files, it does not work. Do you know a better way of setting ownership? I don't like to have to set a specific UID and GID for the ejabberd user simply because the syntax of <pkg>_PERMISSIONS requires an ID instead of a name.
Agreed, I would not want ejabberd to run as root either. I'll see what we can do for those premissions...
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Johan OudinetJohan, All,
On 2014-07-18 14:33 +0200, Johan Oudinet spake thusly:
Quoted 6 lines
Signed-off-by Johan Oudinet <[email protected]>---
Changes v2 -> v3:
- Add system V init script.
Changes v1 -> v2:
- Run ejabberd as ejabberd user instead of root.
I tried building ejabberd in a minmal config, with:
BR2_x86_i686=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_INIT_NONE=y
# BR2_PACKAGE_BUSYBOX is not set
BR2_PACKAGE_EJABBERD=y
# BR2_TARGET_ROOTFS_TAR is not set
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error in rebar_core: {'EXIT',
{badarg,
[{re,replace,
[[40,41,32,123,32,32,108,111,99,97,108,32,
82,69,84,61,34,36,123,63,125,34,59,10,32,
[--SNIP a lot of numbers--]
115,116,111,114,121,32,45,97,59,10,32,112,
114,111,109,112,116,95,102,105,114,115,
116,61,48,10,125],
[92,36,40,"doPrompt",40,92,115,124,36,41,
124,123,"doPrompt",125,41],
[[],"\\2"],
[global,{return,list}]],
[{file,"re.erl"},{line,355}]},
{rebar_port_compiler,merge_each_var,2,
[{file,"src/rebar_port_compiler.erl"},
{line,380}]},
{rebar_port_compiler,setup_env,2,
[{file,"src/rebar_port_compiler.erl"},
{line,167}]},
{rebar_core,'-setup_envs/2-fun-0-',2,
[{file,"src/rebar_core.erl"},{line,436}]},
{lists,foldl,3,
[{file,"lists.erl"},{line,1261}]},
{rebar_core,process_dir1,6,
[{file,"src/rebar_core.erl"},{line,190}]},
{rebar_core,process_commands,2,
[{file,"src/rebar_core.erl"},{line,61}]},
{rebar,main,1,
[{file,"src/rebar.erl"},{line,58}]}]}}
make[2]: *** [deps/.got] Error 1Regards,
Yann E. MORIN.
Quoted 231 lines
--- package/Config.in | 1 + package/ejabberd/Config.in | 13 ++++++ package/ejabberd/S50ejabberd | 29 +++++++++++++ package/ejabberd/check-erlang-lib | 48 ++++++++++++++++++++++ package/ejabberd/ejabberd.mk | 86 +++++++++++++++++++++++++++++++++++++++ 5 files changed, 177 insertions(+) create mode 100644 package/ejabberd/Config.in create mode 100644 package/ejabberd/S50ejabberd create mode 100755 package/ejabberd/check-erlang-lib create mode 100644 package/ejabberd/ejabberd.mkpackage/Config.in +1
-0@@ -943,6 +943,7 @@ endif source "package/dnsmasq/Config.in" source "package/dropbear/Config.in" source "package/ebtables/Config.in" + source "package/ejabberd/Config.in" source "package/ethtool/Config.in" source "package/faifa/Config.in" source "package/fmc/Config.in"package/ejabberd/Config.in new file +13
-0@@ -0,0 +1,13 @@ +config BR2_PACKAGE_EJABBERD + bool "ejabberd" + select BR2_PACKAGE_OPENSSL + select BR2_PACKAGE_ERLANG + select BR2_PACKAGE_LIBYAML + select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE + select BR2_PACKAGE_EXPAT + help + Ejabberd is a Jabber/XMPP instant messaging server, written + in Erlang/OTP. Among other features, ejabberd is + cross-platform, fault-tolerant, clusterable and modular. + + http://www.ejabberd.impackage/ejabberd/S50ejabberd new file +29
-0@@ -0,0 +1,29 @@ +#!/bin/sh +# +# Start/stop ejabberd +# + +case "$1" in + start) + echo "Starting ejabberd..." + ejabberdctl start + ;; + stop) + echo -n "Stopping ejabberd... " + ejabberdctl stop > /dev/null + if [ $? -eq 3 ] || ejabberdctl stopped; then + echo "OK" + else + echo "failed" + fi + ;; + restart|reload) + "$0" stop + "$0" start + ;; + *) + echo "Usage: $0 {start|stop|restart}" + exit 1 +esac + +exit $?package/ejabberd/check-erlang-lib new file +48
-0@@ -0,0 +1,48 @@ +#!/bin/sh -e +# Helper to bypass AC_ERLANG_CHECK_LIB + +usage() { + cat <<EOF +Usage: + $0 library +Look for Erlang's library in TARGET_DIR/usr/lib/erlang/lib. + +If the library is found, it returns the path to the latest version, +relative to TARGET_DIR. Otherwise, it returns "not found". + +If there are several versions, it returns an error because it does not +know which one Erlang uses. + +EOF +} + +die () { + echo "$@" >&2 + exit 1 +} + +if [ $# -ne 1 ]; then + usage + exit 0 +else + library="$1" +fi + +target_dir="${TARGET_DIR:-../../output/target}" + +[ -d "$target_dir" ] || die "TARGET_DIR is not a directory. Please \ +specify the TARGET_DIR environment variable." + +case "$(ls -1d -- "$target_dir/usr/lib/erlang/lib/$library-"* | wc -l)" in + 0) + echo "not found" + ;; + 1) + echo "$target_dir/usr/lib/erlang/lib/$library-"* \ + | sed -e "s,^$target_dir,," + ;; + *) + die "Several versions of $library have been found. Please \ + remove the unused ones." + ;; +esacpackage/ejabberd/ejabberd.mk new file +86
-0@@ -0,0 +1,86 @@ +################################################################################ +# +# ejabberd +# +################################################################################ + +EJABBERD_VERSION = 14.05 +EJABBERD_SITE = $(call github,processone,ejabberd,$(EJABBERD_VERSION)) +EJABBERD_LICENSE = GPLv2+ +EJABBERD_LICENSE_FILES = COPYING +EJABBERD_DEPENDENCIES = libyaml expat openssl erlang + +ifeq ($(BR2_PACKAGE_LIBICONV),y) + EJABBERD_DEPENDENCIES += libiconv +endif + +EJABBERD_ERLANG_LIBS := sasl crypto public_key ssl mnesia inets compiler + +# Do AC_ERLANG_CHECK_LIB job without erlang. +define EJABBERD_SET_LIBS_DIR + for lib in $(EJABBERD_ERLANG_LIBS); do \ + export \ + erlang_lib_dir_$$lib="`package/ejabberd/check-erlang-lib $$lib`"; \ + done +endef + +# Set cross-compilation options to the configure called by rebar. +define EJABBERD_FIX_REBAR_CONFIG_SCRIPT + sed -e "s,./configure,./configure \\ \ + --target=$(GNU_TARGET_NAME) \\ \ + --host=$(GNU_TARGET_NAME) \\ \ + --build=$(GNU_HOST_NAME)," \ + -i "$(@D)"/rebar.config.script +endef + +EJABBERD_PRE_CONFIGURE_HOOKS += \ + EJABBERD_SET_LIBS_DIR \ + EJABBERD_FIX_REBAR_CONFIG_SCRIPT + +# Guess answers for these tests, configure will bail out otherwise +# saying error: cannot run test program while cross compiling. +EJABBERD_CONF_ENV = \ + ac_cv_erlang_root_dir='$(HOST_DIR)/usr/lib/erlang' \ + $(foreach lib, $(EJABBERD_ERLANG_LIBS), \ + ac_cv_erlang_lib_dir_$(lib)="$$erlang_lib_dir_$(lib)") + +# Set environment variables so the rebar compile command does +# cross-compile ejabberd dependencies. +EJABBERD_MAKE_ENV = \ + $(TARGET_CONFIGURE_OPTS) \ + $(TARGET_CONFIGURE_ARGS) \ + LDFLAGS="-L$(HOST_DIR)/usr/$(GNU_TARGET_NAME)/sysroot/usr/lib/erlang/usr/lib" + +# Delete HOST_DIR prefix from ERL path in ejabberctl script. +define EJABBERD_FIX_EJABBERDCTL + sed -e "s,ERL=$(HOST_DIR),ERL=," \ + -e "s,INSTALLUSER=,INSTALLUSER=ejabberd," \ + -i "$(TARGET_DIR)"/usr/sbin/ejabberdctl +endef + +EJABBERD_POST_INSTALL_TARGET_HOOKS += EJABBERD_FIX_EJABBERDCTL + +define EJABBERD_PERMISSIONS +/etc/ejabberd d 750 0 128 - - - - - +/etc/ejabberd/ejabberd.yml-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberd.yml f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg-new f 640 0 128 - - - - - +/etc/ejabberd/ejabberdctl.cfg f 640 0 128 - - - - - +/etc/ejabberd/inetrc f 644 0 128 - - - - - +/usr/sbin/ejabberdctl f 550 0 128 - - - - - +/usr/lib/ejabberd/priv/bin/captcha.sh f 750 119 0 - - - - - +/usr/var/lib/ejabberd d 750 119 0 - - - - - +/usr/var/lock/ejabberdctl d 750 119 0 - - - - - +/usr/var/log/ejabberd d 750 119 0 - - - - - +endef + +define EJABBERD_USERS +ejabberd 119 ejabberd 128 * /var/run/ejabberd - - ejabberd daemon +endef + +define EJABBERD_INSTALL_INIT_SYSV + $(INSTALL) -D -m 0755 package/ejabberd/S50ejabberd \ + $(TARGET_DIR)/etc/init.d/S50ejabberd +endef + +$(eval $(autotools-package))Signature · 6 lines
1.9.1 _______________________________________________ buildroot mailing list buildroot at busybox.net http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Yann, all,
Quoted 9 lines · Yann E. MORIN
I tried building ejabberd in a minmal config, with:
BR2_x86_i686=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_INIT_NONE=y
#BR2_PACKAGE_BUSYBOXis not set
BR2_PACKAGE_EJABBERD=y
#BR2_TARGET_ROOTFS_TARis not set
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.
$ cat defconfig BR2_arm=y BR2_arm1176jzf_s=y BR2_KERNEL_HEADERS_VERSION=y BR2_DEFAULT_KERNEL_VERSION="3.12.20" BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_3_12=y BR2_TOOLCHAIN_BUILDROOT_LARGEFILE=y BR2_TOOLCHAIN_BUILDROOT_LOCALE=y BR2_TOOLCHAIN_BUILDROOT_CXX=y BR2_TARGET_GENERIC_GETTY_PORT="ttyAMA0" BR2_LINUX_KERNEL=y BR2_LINUX_KERNEL_CUSTOM_GIT=y BR2_LINUX_KERNEL_CUSTOM_REPO_URL="git://github.com/raspberrypi/linux.git" BR2_LINUX_KERNEL_CUSTOM_REPO_VERSION="a25152fdfbfd3b4330da16fc3aeb6b1ef8843805" BR2_LINUX_KERNEL_DEFCONFIG="bcmrpi_quick" BR2_LINUX_KERNEL_ZIMAGE=y BR2_PACKAGE_RPI_FIRMWARE=y BR2_PACKAGE_EJABBERD=y
In the meantime, I'll try with yours.
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Yann, All,
Quoted 1 line · Johan Oudinet
In the meantime, I'll try with yours.
I've just tried with your minimal defconfig and the compilation finished without an error :-/
I did a make clean after modifying the defconfig. Should I do something else?
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-08-11 12:07 +0200, Johan Oudinet spake thusly:
Quoted 6 lines · Johan Oudinet
Quoted 1 line
In the meantime, I'll try with yours.
I've just tried with your minimal defconfig and the compilation finished without an error :-/
I did a make clean after modifying the defconfig. Should I do something else?
Yes, please: could you de-install any erlang on your host system, and be sure that the one in use is undeed the one built by Buildroot.
I do not have erlang installed on my machine, for the records.
Also, I'm building your configuration right now, I'll let you know how that goes...
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Johan OudinetJohan, All,
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 33 lines · Yann E. MORIN
Quoted 9 lines
I tried building ejabberd in a minmal config, with:
BR2_x86_i686=y
BR2_TOOLCHAIN_EXTERNAL=y
BR2_INIT_NONE=y
#BR2_PACKAGE_BUSYBOXis not set
BR2_PACKAGE_EJABBERD=y
#BR2_TARGET_ROOTFS_TARis not setQuoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.$ cat defconfig BR2_arm=y BR2_arm1176jzf_s=y BR2_KERNEL_HEADERS_VERSION=y BR2_DEFAULT_KERNEL_VERSION="3.12.20" BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_3_12=y BR2_TOOLCHAIN_BUILDROOT_LARGEFILE=y BR2_TOOLCHAIN_BUILDROOT_LOCALE=y BR2_TOOLCHAIN_BUILDROOT_CXX=y BR2_TARGET_GENERIC_GETTY_PORT="ttyAMA0"
Quoted 8 lines
BR2_LINUX_KERNEL=y BR2_LINUX_KERNEL_CUSTOM_GIT=y BR2_LINUX_KERNEL_CUSTOM_REPO_URL="git://github.com/raspberrypi/linux.git" BR2_LINUX_KERNEL_CUSTOM_REPO_VERSION="a25152fdfbfd3b4330da16fc3aeb6b1ef8843805" BR2_LINUX_KERNEL_DEFCONFIG="bcmrpi_quick" BR2_LINUX_KERNEL_ZIMAGE=y BR2_PACKAGE_RPI_FIRMWARE=y BR2_PACKAGE_EJABBERD=y
I tried this defconfig (except I use an external toolchain to avoid rebuilding it again and again) and I got the exact same error I had previously (I did not check the numbers where the same, though.)
My build machine is an x86_64, running an up-to-date Ubuntu 14.04. I'll check in a 32-bit chroot to see if the error is the same.
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-08-11 12:33 +0200, Yann E. MORIN spake thusly:
Quoted 2 lines
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 1 line · Yann E. MORIN
[--SNIP--]
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,
Quoted 4 lines
Quoted 4 lines
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.
[--SNIP--]
Quoted 1 line
I'll check in a 32-bit chroot to see if the error is the same.
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<---
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got
Uncaught error in rebar_core: {'EXIT',
{undef,
[{crypto,start,[],[]},
{rebar,run_aux,2,
[{file,"src/rebar.erl"},{line,143}]},
{rebar,main,1,
[{file,"src/rebar.erl"},{line,58}]},
{escript,run,2,
[{file,"escript.erl"},{line,752}]},
{escript,start,1,
[{file,"escript.erl"},{line,276}]},
{init,start_it,1,[]},
{init,start_em,1,[]}]}}
make[2]: *** [deps/.got] Error 1
---8<---So, no better luck... :-(
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Hi Yann and all,
Quoted 33 lines · Yann E. MORIN
Johan, All,
On 2014-08-11 12:33 +0200, Yann E. MORIN spake thusly:
Quoted 2 lines
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 1 line · Yann E. MORIN
[--SNIP--]
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,Quoted 4 lines
Quoted 4 lines
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.[--SNIP--]
Quoted 1 line
I'll check in a 32-bit chroot to see if the error is the same.
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<--- /home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got Uncaught error in rebar_core: {'EXIT', {undef, [{crypto,start,[],[]}, {rebar,run_aux,2, [{file,"src/rebar.erl"},{line,143}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]},
Quoted 8 lines
{escript,run,2,
[{file,"escript.erl"},{line,752}]},
{escript,start,1,
[{file,"escript.erl"},{line,276}]},
{init,start_it,1,[]},
{init,start_em,1,[]}]}}
make[2]: *** [deps/.got] Error 1
---8<---I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
Frank
PATCH v3 1/1 ejabberd: new package
Frank, All,
On 2014-08-13 14:48 -0400, Frank Hunleth spake thusly:
Quoted 43 lines · Yann E. MORIN
Quoted 33 lines
Johan, All,
On 2014-08-11 12:33 +0200, Yann E. MORIN spake thusly:
Quoted 2 lines
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 1 line · Yann E. MORIN
[--SNIP--]
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,Quoted 4 lines
Quoted 4 lines
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.[--SNIP--]
Quoted 1 line
I'll check in a 32-bit chroot to see if the error is the same.
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<--- /home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got Uncaught error in rebar_core: {'EXIT', {undef, [{crypto,start,[],[]}, {rebar,run_aux,2, [{file,"src/rebar.erl"},{line,143}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]},Quoted 8 lines
{escript,run,2, [{file,"escript.erl"},{line,752}]}, {escript,start,1, [{file,"escript.erl"},{line,276}]}, {init,start_it,1,[]}, {init,start_em,1,[]}]}} make[2]: *** [deps/.got] Error 1 ---8<---
Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
Ah, good idea.
I currently have them installed, but I think I installed them recently. Let me retry a build, once with them installed, once with them uninstalled.
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Frank HunlethFrank, All,
On 2014-08-13 14:48 -0400, Frank Hunleth spake thusly:
Quoted 43 lines · Yann E. MORIN
Quoted 33 lines
Johan, All,
On 2014-08-11 12:33 +0200, Yann E. MORIN spake thusly:
Quoted 2 lines
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 1 line · Yann E. MORIN
[--SNIP--]
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,Quoted 4 lines
Quoted 4 lines
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.[--SNIP--]
Quoted 1 line
I'll check in a 32-bit chroot to see if the error is the same.
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<--- /home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got Uncaught error in rebar_core: {'EXIT', {undef, [{crypto,start,[],[]}, {rebar,run_aux,2, [{file,"src/rebar.erl"},{line,143}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]},Quoted 8 lines
{escript,run,2, [{file,"escript.erl"},{line,752}]}, {escript,start,1, [{file,"escript.erl"},{line,276}]}, {init,start_it,1,[]}, {init,start_em,1,[]}]}} make[2]: *** [deps/.got] Error 1 ---8<---
Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
Alas, no better luck.
I did de-install my libssl-dev package, so I only get the runtime from openssl, and not the development files.
Then, I did three tests, all starting afresh from a clean tree, with the ejabberd patch applied on top of master (e3bb4ac: fbv: fix typo in pkg-config variable):
- with just the patch;
- then with a modified erlang.mk to compile our host erlang without ssl;
- then with a modified erlang.mk to compile our host erlang with ssl.
All three build failed in the exact same way.
To disable ssl, I did:
HOST_ERLANG_CONF_OPT += --without-ssl
To enable ssl, I did:
HOST_ERLANG_CONF_OPT += --with-ssl
HOST_ERLANG_DEPENDENCIES += host-opensslRegards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Yann, All,
Quoted 60 lines · Yann E. MORIN
Frank, All,
On 2014-08-13 14:48 -0400, Frank Hunleth spake thusly:
Quoted 43 lines · Yann E. MORIN
Quoted 33 lines
Johan, All,
On 2014-08-11 12:33 +0200, Yann E. MORIN spake thusly:
Quoted 2 lines
On 2014-08-11 11:36 +0200, Johan Oudinet spake thusly:
Quoted 1 line · Yann E. MORIN
[--SNIP--]
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
And it fails at build time, with a lengthy error message, that I do not completely grasp:
/home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escriptrebar
get-deps && :> deps/.got
==> rel (get-deps)
Uncaught error inrebar_core: {'EXIT',
{badarg,Quoted 4 lines
Quoted 4 lines
Very strange. It looks like the rebar script didn't accept the get-deps argument.
Could you try with the same defconfig as mine to see if the problem comes from there.[--SNIP--]
Quoted 1 line
I'll check in a 32-bit chroot to see if the error is the same.
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<--- /home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got Uncaught error in rebar_core: {'EXIT', {undef, [{crypto,start,[],[]}, {rebar,run_aux,2, [{file,"src/rebar.erl"},{line,143}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]},Quoted 8 lines
{escript,run,2, [{file,"escript.erl"},{line,752}]}, {escript,start,1, [{file,"escript.erl"},{line,276}]}, {init,start_it,1,[]}, {init,start_em,1,[]}]}} make[2]: *** [deps/.got] Error 1 ---8<---Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
Alas, no better luck.
I did de-install my libssl-dev package, so I only get the runtime from openssl, and not the development files.
Then, I did three tests, all starting afresh from a clean tree, with the ejabberd patch applied on top of master (e3bb4ac: fbv: fix typo in pkg-config variable):
Quoted 5 lines
- with just the patch;
- then with a modified erlang.mk to compile our host erlang without ssl;
- then with a modified erlang.mk to compile our host erlang with ssl.All three build failed in the exact same way.
Do you have an output/host/usr/lib/erlang/lib/crypto-3.3/priv/lib directory with a couple shared libraries in it? I assume not based on the failure, but wanted to check.
Also, does the log from Erlang's ./configure provide any clues?
FWIW, everything that you're doing sounds good to me. I've only seen the "crypto" undef error on targets when I've forgotten to enable openssl. I've never seen it on the host, but my host has a lot of header files, so it's probably not a good test.
Quoted 18 lines
To disable ssl, I did:
HOST_ERLANG_CONF_OPT += --without-sslTo enable ssl, I did:
HOST_ERLANG_CONF_OPT += --with-ssl HOST_ERLANG_DEPENDENCIES += host-opensslRegards,
Yann E. MORIN.--
.-----------------.--------------------.------------------.--------------------.| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |'------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Frank, All,
On 2014-08-13 17:23 -0400, Frank Hunleth spake thusly:
[--SNIP--]
Quoted 38 lines
Quoted 38 lines
Quoted 21 lines
Quoted 12 lines
OK, here's what it has to say when run in a 32-bit sueeze chroot:
---8<--- /home/ymorin/dev/buildroot/O/host/usr/lib/erlang/bin/escript rebar get-deps && :> deps/.got Uncaught error in rebar_core: {'EXIT', {undef, [{crypto,start,[],[]}, {rebar,run_aux,2, [{file,"src/rebar.erl"},{line,143}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]},Quoted 8 lines
{escript,run,2, [{file,"escript.erl"},{line,752}]}, {escript,start,1, [{file,"escript.erl"},{line,276}]}, {init,start_it,1,[]}, {init,start_em,1,[]}]}} make[2]: *** [deps/.got] Error 1 ---8<---Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
Alas, no better luck.
I did de-install my libssl-dev package, so I only get the runtime from openssl, and not the development files.
Then, I did three tests, all starting afresh from a clean tree, with the ejabberd patch applied on top of master (e3bb4ac: fbv: fix typo in pkg-config variable):
- with just the patch;
- then with a modified erlang.mk to compile our host erlang without ssl;
- then with a modified erlang.mk to compile our host erlang with ssl.
Quoted 5 lines
Quoted 1 line
All three build failed in the exact same way.
Do you have an output/host/usr/lib/erlang/lib/crypto-3.3/priv/lib directory with a couple shared libraries in it? I assume not based on the failure, but wanted to check.
I re-ran the test with erlang.mk patched to depend on host-openssl and added back the --with-ssl flag when building the host erlang.
Indeed, I do not have this directory.
Quoted 1 line
Also, does the log from Erlang's ./configure provide any clues?
Aha!
checking for static ZLib to be used by SSL in standard locations... no checking for OpenSSL >= 0.9.7 in standard locations... no
configure: WARNING: No (usable) OpenSSL found, skipping ssl, ssh and crypto applications
Dang... It seems we have to tell it where to look for our own openssl. Lemme see... Yes, we can say --with-ssl=PATH . I'll try that, and will report later...
Thanks for the hint! :-)
Regards,
Yann E. MORIN.
Quoted 23 lines
FWIW, everything that you're doing sounds good to me. I've only seen the "crypto" undef error on targets when I've forgotten to enable openssl. I've never seen it on the host, but my host has a lot of header files, so it's probably not a good test.
Quoted 18 lines
To disable ssl, I did:
HOST_ERLANG_CONF_OPT += --without-sslTo enable ssl, I did:
HOST_ERLANG_CONF_OPT += --with-ssl HOST_ERLANG_DEPENDENCIES += host-opensslRegards,
Yann E. MORIN.--
.-----------------.--------------------.------------------.--------------------.| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |'------------------------------^-------^------------------^--------------------'
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Frank, All,
On 2014-08-13 23:49 +0200, Yann E. MORIN spake thusly:
Quoted 4 lines
On 2014-08-13 17:23 -0400, Frank Hunleth spake thusly:
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
[--SNIP--]
Quoted 4 lines
I re-ran the test with erlang.mk patched to depend on host-openssl and added back the --with-ssl flag when building the host erlang.
Indeed, I do not have this directory.
OK, now I have it. My host-erlang is compiled with ssl support. But...
Quoted 11 lines
Quoted 1 line
Also, does the log from Erlang's ./configure provide any clues?
Aha!
checking for static ZLib to be used by SSL in standard locations... no checking for OpenSSL >= 0.9.7 in standard locations... no
configure: WARNING: No (usable) OpenSSL found, skipping ssl, ssh and crypto applicationsDang... It seems we have to tell it where to look for our own openssl. Lemme see... Yes, we can say --with-ssl=PATH . I'll try that, and will report later...
Now, my error is back to this sequence of numbers:
Uncaught error in rebar_core: {'EXIT',
{badarg,
[{re,replace,
[[40,41,32,123,32,32,108,111,99,97,108,32,
82,69,84,61,34,36,123,63,125,34,59,10,32,
108,111,99,97,108,32,105,115,95,110,101,
116,95,109,110,116,112,116,61,48,59,10,32,
108,111,99,97,108,32,72,71,80,76,65,73,78,
59,10,32,108,111,99,97,108,32,76,79,65,68,
...
66,76,85,125,36,123,80,83,49,125,36,123,
65,95,78,79,82,125,34,59,10,32,104,105,
115,116,111,114,121,32,45,97,59,10,32,112,
114,111,109,112,116,95,102,105,114,115,
116,61,48,10,125],
[92,36,40,"doPrompt",40,92,115,124,36,41,
124,123,"doPrompt",125,41],
[[],"\\2"],
[global,{return,list}]],
[{file,"re.erl"},{line,355}]},
{rebar_port_compiler,merge_each_var,2,
[{file,"src/rebar_port_compiler.erl"},
{line,380}]},
{rebar_port_compiler,setup_env,2,
[{file,"src/rebar_port_compiler.erl"},
{line,167}]},
{rebar_core,'-setup_envs/2-fun-0-',2,
[{file,"src/rebar_core.erl"},{line,436}]},
{lists,foldl,3,
[{file,"lists.erl"},{line,1261}]},
{rebar_core,process_dir1,6,
[{file,"src/rebar_core.erl"},{line,190}]},
{rebar_core,process_commands,2,
[{file,"src/rebar_core.erl"},{line,61}]},
{rebar,main,1,
[{file,"src/rebar.erl"},{line,58}]}]}}make[1]: *** [deps/.got] Error 1
OK, I'm lost... :-(
One thing I just noticed, are those two lines (visible above, too):
[92,36,40,"doPrompt",40,92,115,124,36,41,
124,123,"doPrompt",125,41],doPrompt is the content of my $PROMPT_COMMAND environment variable, in case that rings a bell somewhere...
I'll let you guys investigate further. I can run tests if you need to.
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Yann,
Quoted 66 lines · Yann E. MORIN
Frank, All,
On 2014-08-13 23:49 +0200, Yann E. MORIN spake thusly:
Quoted 4 lines
On 2014-08-13 17:23 -0400, Frank Hunleth spake thusly:
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
I think that you can get this error if host erlang is compiled without openssl. Maybe you don't have the openssl headers installed on your system and Johan does?
[--SNIP--]
Quoted 4 lines
I re-ran the test with erlang.mk patched to depend on host-openssl and added back the --with-ssl flag when building the host erlang.
Indeed, I do not have this directory.
OK, now I have it. My host-erlang is compiled with ssl support. But...
Quoted 11 lines
Quoted 1 line
Also, does the log from Erlang's ./configure provide any clues?
Aha!
checking for static ZLib to be used by SSL in standard locations... no checking for OpenSSL >= 0.9.7 in standard locations... no
configure: WARNING: No (usable) OpenSSL found, skipping ssl, ssh and crypto applicationsDang... It seems we have to tell it where to look for our own openssl. Lemme see... Yes, we can say --with-ssl=PATH . I'll try that, and will report later...
Now, my error is back to this sequence of numbers:
Uncaught error in
rebar_core: {'EXIT',{badarg, [{re,replace, [[40,41,32,123,32,32,108,111,99,97,108,32, 82,69,84,61,34,36,123,63,125,34,59,10,32, 108,111,99,97,108,32,105,115,95,110,101, 116,95,109,110,116,112,116,61,48,59,10,32, 108,111,99,97,108,32,72,71,80,76,65,73,78, 59,10,32,108,111,99,97,108,32,76,79,65,68, ... 66,76,85,125,36,123,80,83,49,125,36,123, 65,95,78,79,82,125,34,59,10,32,104,105, 115,116,111,114,121,32,45,97,59,10,32,112, 114,111,109,112,116,95,102,105,114,115, 116,61,48,10,125], [92,36,40,"doPrompt",40,92,115,124,36,41, 124,123,"doPrompt",125,41], [[],"\\2"], [global,{return,list}]], [{file,"re.erl"},{line,355}]}, {rebar_port_compiler,merge_each_var,2, [{file,"src/rebar_port_compiler.erl"}, {line,380}]}, {rebar_port_compiler,setup_env,2, [{file,"src/rebar_port_compiler.erl"}, {line,167}]}, {rebar_core,'-setup_envs/2-fun-0-',2, [{file,"src/rebar_core.erl"},{line,436}]}, {lists,foldl,3, [{file,"lists.erl"},{line,1261}]}, {rebar_core,process_dir1,6, [{file,"src/rebar_core.erl"},{line,190}]}, {rebar_core,process_commands,2, [{file,"src/rebar_core.erl"},{line,61}]}, {rebar,main,1, [{file,"src/rebar.erl"},{line,58}]}]}}
Quoted 9 lines
make[1]: *** [deps/.got] Error 1
OK, I'm lost... :-(
One thing I just noticed, are those two lines (visible above, too):
[92,36,40,"doPrompt",40,92,115,124,36,41, 124,123,"doPrompt",125,41],
The lists of numbers are Erlang's way of printing strings when it gets confused. Presumably your $PROMPT_COMMAND is "\\$(doPrompt(\\s|$)|{doPrompt})". It looks like rebar is trying to do a regular expression search and replace in a string that starts with "() { local RET=\"${?}\";\n local is_net_mntpt=0;\n local HGPLAIN;\n local LOAD".
I looked through the rebar source code to see why it would do this, but ejabberd must have an old version of rebar since it didn't match. Just as an aside, in the Erlang world, people often copy the "compiled" version of rebar to their projects so that any one who wants to build them doesn't have to download and build the build tool. Rebar has some issues, so many projects have switched back to Makefiles, but clearly not ejabberd.
I have two thoughts: first is to try to build with a minimal environment just in case rebar is getting confused on something in yours (like your prompt settings). The second is to replace the version of rebar in ejabberd with the latest and greatest in case they fixed it. See https://github.com/rebar/rebar/releases.
To be honest, this feels like it's getting above and beyond the level of debug that you should have to do. I have some fondness for Erlang, so I might try to put in some more time this weekend to figure out what's going on.
Frank
Quoted 15 lines
doPrompt is the content of my $PROMPT_COMMAND environment variable, in case that rings a bell somewhere...
I'll let you guys investigate further. I can run tests if you need to.
Regards,
Yann E. MORIN.--
.-----------------.--------------------.------------------.--------------------.| Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |'------------------------------^-------^------------------^--------------------'
Signature · 4 lines
Frank Hunleth Troodon Software LLC Embedded Software Development http://troodon-software.com/
PATCH v3 1/1 ejabberd: new package
Frank, All,
Thanks for all the help so far! :-)
On 2014-08-14 08:40 -0400, Frank Hunleth spake thusly:
Quoted 2 lines · Yann E. MORIN
Quoted 1 line
On 2014-08-13 23:49 +0200, Yann E. MORIN spake thusly:
[--SNIP--]
Quoted 13 lines
Quoted 13 lines
Now, my error is back to this sequence of numbers:
Uncaught error in
rebar_core: {'EXIT',{badarg, [{re,replace, [[40,41,32,123,32,32,108,111,99,97,108,32, 82,69,84,61,34,36,123,63,125,34,59,10,32, 108,111,99,97,108,32,105,115,95,110,101, ... 115,116,111,114,121,32,45,97,59,10,32,112, 114,111,109,112,116,95,102,105,114,115, 116,61,48,10,125], [92,36,40,"doPrompt",40,92,115,124,36,41,
[--SNIP--]
Quoted 6 lines
The lists of numbers are Erlang's way of printing strings when it gets confused. Presumably your $PROMPT_COMMAND is "\\$(doPrompt(\\s|$)|{doPrompt})". It looks like rebar is trying to do a regular expression search and replace in a string that starts with "() { local RET=\"${?}\";\n local is_net_mntpt=0;\n local HGPLAIN;\n local LOAD".
$ echo $PROMPT_COMMAND
doPromptdoPrompt is a shell function that prepares my prompt with a lot of information (git status, time, battery left...)
I've run a build where PROMPT_COMMAND was missing in the envirnment (but doPrompt still exists), to no avail. I still get all those numbers... I did decode all those numbers, and they happen to be exactly the content of my doPrompt function.
I really wonder what is so strange that rebar chokes on it. Why doe it even have to look at those, to begin with? :-(
Oh, could that be that doPrompt is exported?
$ env |grep doPrompt
PROMPT_COMMAND=doPrompt
doPrompt=() { local RET="${?}";I have this in my .bashrc :
export -f doPrompt
Hm... I'll try to find some time to run without that in my env... Yet, it should not choke on it.
Quoted 5 lines
I have two thoughts: first is to try to build with a minimal environment just in case rebar is getting confused on something in yours (like your prompt settings). The second is to replace the version of rebar in ejabberd with the latest and greatest in case they fixed it. See https://github.com/rebar/rebar/releases.
I was thinking about the exact same solution, to build our own rebar.
Quoted 2 lines
To be honest, this feels like it's getting above and beyond the level of debug that you should have to do.
Well, to be honest, I don't think ejabberd can go in Buildroot at all for now. It has to do downloads during build time, which is a big show-stopper IMHO. We currently have no other package that downloads stuff during the build phase, and for those that usualy do (eg. tvheadend), we fixed them not to (e.g. to rely on another package, like dtv-scan-tables.)
And to be further honest, I see Erlang and its ecosystem to be just completely broken, even more broken than Perl+CPAN (which I do not like much to start with.)
Yet, I would like to see a Jabber server in Buildroot. I would *love* that. Hence my persistence in trying to find a solution. But I'm here at a loss, with no further idea... :-(
Quoted 3 lines
I have some fondness for Erlang,
so I might try to put in some more time this weekend to figure out what's going on.
That'd be great, thanks! :-)
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Yann E. MORINYann, all,
Quoted 3 lines · Yann E. MORIN
My build machine is an
x86_64, running an up-to-date Ubuntu 14.04. I'll check in a 32-bit chroot to see if the error is the same.
I removed ejabberd and all erlang packages installed on the host, even purged them to be sure buildroot does not use their configuration files. It is also an up-to-date Ubuntu 14.04 64bits.
After that, make clean all (with your minimal defconfig) still work :-/ Is there anything more I can do?
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-08-11 15:13 +0200, Johan Oudinet spake thusly:
Quoted 10 lines · Yann E. MORIN
Quoted 3 lines
My build machine is an
x86_64, running an up-to-date Ubuntu 14.04. I'll check in a 32-bit chroot to see if the error is the same.I removed ejabberd and all erlang packages installed on the host, even purged them to be sure buildroot does not use their configuration files. It is also an up-to-date Ubuntu 14.04 64bits.
After that, make clean all (with your minimal defconfig) still work :-/ Is there anything more I can do?
No, sorry. I am clueless now...
If I get a bit of time, I'll run some tests tonight...
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Johan OudinetJohan, All,
On 2014-07-18 14:33 +0200, Johan Oudinet spake thusly:
package/Config.in endif · line 949
Quoted 12 lines
package/Config.in +1
-0@@ -943,6 +943,7 @@ endif source "package/dnsmasq/Config.in" source "package/dropbear/Config.in" source "package/ebtables/Config.in" + source "package/ejabberd/Config.in" source "package/ethtool/Config.in" source "package/faifa/Config.in" source "package/fmc/Config.in"
[--SNIP--]
We've been discussing this ejabberd patch during the current Developpers Day in D?sseldorf.
Given the current issues that ewre noticed with building ejabberd, we decided to mark the patch as Rejected in patchwork.
This means that we rejected that very specific patch, *not* that we decided not to have ejabberd. But before it gets included, all the issues must be fixed first, especially the build-time downloading of dependencies, as well as all the remarks that were made by Thomas and I.
Of course, we are happy to help solve them, if you need help/advice on it.
Thank you for working on this!
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Hi Yann, all,
Quoted 11 lines · Yann E. MORIN
We've been discussing this ejabberd patch during the current Developpers Day in D?sseldorf.
Given the current issues that ewre noticed with building ejabberd, we decided to mark the patch as Rejected in patchwork.
This means that we rejected that very specific patch, *not* that we decided not to have ejabberd. But before it gets included, all the issues must be fixed first, especially the build-time downloading of dependencies, as well as all the remarks that were made by Thomas and I.
Sure, it makes sense to me as well. As I said in a previous mail, I did fix my patch to include your remarks, but two:
1. Fix ejabberd build system to not download dependencies at build time.
2. Avoid to use a fix UID and GID for ejabberd user.
I don't know if I should republish the patch before fixing these two issues. I'd like your help on both of them.
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager, p1_cache_tab, p1_iconv, p1_stringprep, p1_stun, p1_tls, p1_utils, p1_xml, p1_yaml, p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from the all' makefile rule and deps/.build from the src' rule dependency, it might work.
Do you have a better idea?
** Avoid to use a fix UID and GID for ejabberd user ** ===================================
I still don't know how I could set the correct permissions to ejabberd files in <PKG>_PERMISSIONS if I set -1 for UID and GID in <PKG>_USERS
Best regards,
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Johan, All,
On 2014-10-16 14:38 +0200, Johan Oudinet spake thusly:
Quoted 32 lines · Yann E. MORIN
Quoted 11 lines
We've been discussing this ejabberd patch during the current Developpers Day in D?sseldorf.
Given the current issues that ewre noticed with building ejabberd, we decided to mark the patch as Rejected in patchwork.
This means that we rejected that very specific patch, *not* that we decided not to have ejabberd. But before it gets included, all the issues must be fixed first, especially the build-time downloading of dependencies, as well as all the remarks that were made by Thomas and I.
Sure, it makes sense to me as well. As I said in a previous mail, I did fix my patch to include your remarks, but two:
1. Fix ejabberd build system to not download dependencies at build time.
2. Avoid to use a fix UID and GID for ejabberd user.I don't know if I should republish the patch before fixing these two issues. I'd like your help on both of them.
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager,p1_cache_tab,p1_iconv,p1_stringprep,p1_stun,p1_tls,p1_utils,p1_xml,p1_yaml,p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from theall' makefile rule and deps/.build from thesrc' rule dependency, it might work.
Quoted 1 line
Do you have a better idea?
No, I have no better idea. I don't know how rebar works, but we indeed need to have one package for each dependency. Having ejabberd (rebar, in fact) do the download at configure/build time is not acceptable.
Howeever, I can see where that brings us:
- there are so far 14 identified such dependencies;
- maybe (probably?) each of those dependencies mya in turn have more
dependencies, which add up to the list;
- I have absolutely no idea how rebar (or the whole erlang ecosystem) behaves for corss-compilation (hint: probably badly), which would make for very tricky package recipes;
- thus, we'd highly benefit from a specialised package infrastructure, like we have for perl, python or lua packages, though I'm afraid that would not be a piece of cake to add such an infra.
So, I really don;t know where to go from there... :-(
If you were inclined into looking into this, then I think the best course of actions for you would be somelthing along those lines:
(1) create a new package for the simplest, leaf-most dependency; this will allow you to assess how complex it is to cross-build a single, simple erlang package;
(2) add a new package that depends directly on, and only on the one above; this will allow you to assess how complex it is to express erlang dependencies in cross-compilation;
(3) progress with a few more packages (say, up to 3-5 packages); this will allow you to have a global understanding about the comonalities and specifities of erlang packages;
(4) see how to turn that into a specific package infrastructure;
(5) post your results to the list, so we can all agree on that new infra; this will allow all of us to get to solve issues in a clean and sane way;
(6) introduce that new infra, and convert the existing packages over
to that new infra;
(7) add every missing ejabberd dependencies one after the other, adapting the new infra in case it is needed, until you eventually reach to packaging ejabberd;
(8) rejoice! ;-)
Note that this will probably be a lot of work, and I do not expect it to be easy at all. I'm willing to participate in this effort, if you want.
Quoted 5 lines
** Avoid to use a fix UID and GID for ejabberd user ** ===================================
I still don't know how I could set the correct permissions to ejabberd files [...]
A quick solution to that would be to have all of ejabberd's files in a specific directory, like /var/ejabberd (find a better location1), add the necessary symlinks to point to that location, and use that location in EJABBERD_USERS:
ejabberd -1 ejabberd -1 * /var/ejabberd - - ejabberd daemon
Quoted 1 line
[...] in <PKG>_PERMISSIONS if I set -1 for UID and GID in <PKG>_USERS
Indeed, there is no provision for using UIDs/GIDs from the _USERS variable, from the _PERMISSIONS.
I'll look into that...
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
In reply to Johan OudinetQuoted 11 lines · Johan Oudinet
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager,p1_cache_tab,p1_iconv,p1_stringprep,p1_stun,p1_tls,p1_utils,p1_xml,p1_yaml,p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from theall' makefile rule and deps/.build from thesrc' rule dependency, it might work.
Quoted 1 line
Do you have a better idea?
Perhaps it's an option to run rebar in download-only mode in the download step?
We'd probably need to add a rebar-specific download option, and make sure that it downloads to the download directory, and make sure that it doesn't do any network access if the files are already there. This may be complicated...
If you need to create separate packages, it's probably worthwhile to create a
script to automate that, like Francois Perrad did for the perl modules.
Regards,
Arnout
Signature · 6 lines
Arnout Vandecappelle arnout at mind be Senior Embedded Software Architect +32-16-286500 Essensium/Mind http://www.mind.be G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle GPG fingerprint: 7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
PATCH v3 1/1 ejabberd: new package
Arnout, All,
On 2014-10-18 01:06 +0200, Arnout Vandecappelle spake thusly:
Quoted 12 lines · Johan Oudinet
Quoted 11 lines
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager,p1_cache_tab,p1_iconv,p1_stringprep,p1_stun,p1_tls,p1_utils,p1_xml,p1_yaml,p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from theall' makefile rule and deps/.build from thesrc' rule dependency, it might work.
Quoted 3 lines
Quoted 1 line
Do you have a better idea?
Perhaps it's an option to run rebar in download-only mode in the download step?
The problem is that rebar is an executable bundled in the ejabberd archive, so, we can't run it before we extract ejabberd... :-(
Unless we can switch to using an external rebar. Johan?
Regards,
Yann E. MORIN.
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------. | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: | | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ | | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no | | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. | '------------------------------^-------^------------------^--------------------'
PATCH v3 1/1 ejabberd: new package
Yann, Arnout, All,
Le 18 oct. 2014 12:00, "Yann E. MORIN" <[email protected]> a ?crit :
Quoted 16 lines
On 2014-10-18 01:06 +0200, Arnout Vandecappelle spake thusly:
Quoted 15 lines · Johan Oudinet
Quoted 11 lines
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager,p1_cache_tab,p1_iconv,p1_stringprep,p1_stun,p1_tls,p1_utils,p1_xml,p1_yaml,p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from theall' makefile rule and deps/.build from thesrc' rule dependency, it might work.Quoted 1 line
Do you have a better idea?
Perhaps it's an option to run rebar in download-only mode in the
download step?
Quoted 5 lines
The problem is that rebar is an executable bundled in the ejabberd archive, so, we can't run it before we extract ejabberd... :-(
Unless we can switch to using an external rebar. Johan?
Well, we could package the rebar project and use it instead of the binary included in ejabberd but I don't think it's a good idea.
Actually, I've looked at how Debian has packaged ejabberd and they did
patch it to get rid of the get-deps rule and packages every dependency separately.
I'm going to have a look how perl is integrated in buildroot.
Thanks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20141018/e2704eb2/attachment.html>
PATCH v3 1/1 ejabberd: new package
Quoted 23 lines · Johan Oudinet
Yann, Arnout, All,
Le 18 oct. 2014 12:00, "Yann E. MORIN" <[email protected]> a ?crit :Quoted 21 lines
On 2014-10-18 01:06 +0200, Arnout Vandecappelle spake thusly:
Quoted 12 lines · Johan Oudinet
Quoted 11 lines
** Fix ejabberd build system
** ===================
The only solution i see is to create a buildroot package for every ejabberd dependency, listed in rebar.config.script :
esip, goldrush, lager,p1_cache_tab,p1_iconv,p1_stringprep,p1_stun,p1_tls,p1_utils,p1_xml,p1_yaml,p1_zlib, xmlrpc
Then, modify ejabberd package to not use rebar at all, or at least to not call rebar get-deps. Thus, if I remove deps from theall' makefile rule and deps/.build from thesrc' rule dependency, it might work.Quoted 4 lines
Quoted 1 line
Do you have a better idea?
Perhaps it's an option to run rebar in download-only mode in the
download step?The problem is that rebar is an executable bundled in the ejabberd archive, so, we can't run it before we extract ejabberd... :-(
Quoted 8 lines
Quoted 1 line
Unless we can switch to using an external rebar. Johan?
Well, we could package the rebar project and use it instead of the binary included in ejabberd but I don't think it's a good idea. Actually, I've looked at how Debian has packaged ejabberd and they did patch it to get rid of the get-deps rule and packages every dependency separately. I'm going to have a look how perl is integrated in buildroot. Thanks.
I (with the help of a colleague) manage to package every ejabberd dependency separately. We also had to package rebar because some erlang packages rely on it but do not provide it. Therefore, if an erlang package includes a rebar script, we use it. Otherwise, we add a dependency on host-erlang-rebar and we use this one instead. To simplify the task (12 erlang packages), we created a pkg-rebar.mk file that provides {host-}rebar-package, similar to pkg-autotools.mk. In doing so, we had to factorize the hooks in pkg-autotools.mk.
I guess the minor modifications of pkg-autotools.mk should be pushed in a separate patch.
Then, should I push the rest of my commits as a big new revision of this patch, or should I push each erlang package as a separate patch ?
Signature · 1 line
Johan
PATCH v3 1/1 ejabberd: new package
Dear Johan Oudinet,
Quoted 9 lines · Johan Oudinet
I (with the help of a colleague) manage to package every ejabberd dependency separately. We also had to package rebar because some erlang packages rely on it but do not provide it. Therefore, if an erlang package includes a rebar script, we use it. Otherwise, we add a dependency on host-erlang-rebar and we use this one instead. To simplify the task (12 erlang packages), we created a pkg-rebar.mk file that provides {host-}rebar-package, similar to pkg-autotools.mk. In doing so, we had to factorize the hooks in pkg-autotools.mk.
Quoted 4 lines
I guess the minor modifications of pkg-autotools.mk should be pushed in a separate patch.
Then, should I push the rest of my commits as a big new revision of this patch, or should I push each erlang package as a separate patch ?
Each should be a separate patch. Basically, you should send a patch series, which contains separate patches for the various functionalities (new package infrastructure, changes to existing package infrastructure, new packages, etc.). Do not forget to document what you are doing in the commit logs, especially for the modification to existing package infrastructures and the new package infra.
Thanks,
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com