Updated dependencies (master)#2340
Conversation
|
@cf-bottom Jenkins please :) |
|
Sure, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14160/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14160/ |
|
@cf-bottom Jenkins please :) |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14165/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14165/ |
|
will rebase |
0d4779f to
60f655b
Compare
993418e to
f882a56
Compare
|
@cf-bottom jenkins please |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14174/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14174/ |
|
@cf-bottom jenkins please |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14179/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14179/ |
|
@cf-bottom jenkins please |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14182/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14182/ |
|
PACKAGES_x86_64_mingw successfully built @cf-bottom please jenkins |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14194/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14194/ |
|
|
||
| as_fn_error $? "Could not find pthreads library" "$LINENO" 5 | ||
|
|
||
| +fi |
There was a problem hiding this comment.
we should not be patching the configure script, but rather configure.ac and in this case Makefile.am where I recently added -ldl.
It's a bit of a mess as we add the autotools stuff and then in either patch 0005 or 0009 we add files like this configure which we should really NOT include, preferring to run autoconf/etc instead.
There was a problem hiding this comment.
good catch, moved the -ldl fix into the autotools source. 0012 now patches configure.ac and Makefile.am. 0013 carries the regenerated configure/Makefile.in, same split as 0009/0010.
b4b906a to
679f7b5
Compare
|
@cf-bottom jenkins please |
|
Sure, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14200/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14200/ |
679f7b5 to
160f66d
Compare
|
rebased |
|
@cf-bottom jenkins please |
|
Sure, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14203/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14203/ |
|
@cf-bottom jenkins |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14206/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14206/ |
|
|
||
| lib_LTLIBRARIES = liblmdb.la | ||
| liblmdb_la_SOURCES = mdb.c midl.c module.c | ||
| -liblmdb_la_LDFLAGS=-no-undefined -avoid-version -ldl |
There was a problem hiding this comment.
I don't think you should include this patch. Just revert the previous changes to an existing patch that I made to include -ldl and your changes to configure.ac will handle the situation. Just make sure the patch with configure itself contains the right changes as well, which I think it should at this point.
There was a problem hiding this comment.
Reworked as suggested: dropped patches 0012/0013 and reverted the hard-coded -ldl changes in fb82709 the AC_CHECK_LIB([dl], [dlopen]) detection now lives in 0003 configure.ac and the regenerated check in 0005 configure. With AM_LDFLAGS = -ldl gone the 0003 hunk count is correct again, so the include_HEADERS fix commit is removed entirely.
| From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001 | ||
| From: Ihor Aleksandrychiev <igor.aleksandrychev@northern.tech> | ||
| Date: Tue, 14 Jul 2026 20:53:21 +0300 | ||
| Subject: [PATCH 13/13] Updated generated autotools files |
There was a problem hiding this comment.
I think this patch should also be removed. Just revert my previous changes instead.
There was a problem hiding this comment.
in commit fb82709, remove the changes to deps-packaging/lmdb/0005-Generated-autoconf-files.patch and deps-packaging/lmdb/0009-Added-mdb_dump-and-mdb_load-to-build.patch
| --- | ||
| libraries/liblmdb/Makefile | 117 --------------------------------- | ||
| libraries/liblmdb/Makefile.am | 13 ++++ | ||
| libraries/liblmdb/Makefile.am | 14 ++++ |
There was a problem hiding this comment.
when you fix the earlier patch you can remove this commit entirely.
1bf4320 to
29d07d2
Compare
lmdb 1.0.0's module.c uses dlopen() on Unix, but libdl does not exist on Windows (mingw uses LoadLibrary() instead), so -ldl must not be hard-coded in the Makefile.am autotools source. Instead, configure.ac detects libdl via AC_CHECK_LIB([dl], [dlopen]), which adds -ldl to LIBS only when it actually links: Unix builds still link against libdl, while the mingw cross build simply omits it. The generated configure in 0005 carries the corresponding regenerated check. Ticket: none Changelog: none
Adjusting core code to work with a dynamic max key size in lmdb is a larger effort. Ticket: ENT-14295 Changelog: none
Signed-off-by: Ihor Aleksandrychiev <ihor.aleksandrychiev@northern.tech>
lmdb 1.0.0 added incremental dump/load and MDB_RPAGE_CACHE remapping code that treats the Windows LARGE_INTEGER union as a scalar integer, which fails to compile with x86_64-w64-mingw32-gcc. Add a patch that accesses the .QuadPart member on Windows so offset arithmetic and SetFilePointerEx() use the correct type. POSIX builds are unaffected. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
29d07d2 to
23e513f
Compare
|
@cf-bottom jenkins please |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/14208/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-14208/ |
Automated dependency updates for the master branch
together: cfengine/core#6235