Hello.
What will be the result of pkg check -dan when such a condition occurs?
That might not be a problem, but the only way to know for sure is to ask the
person who can make it happen :)
Regards.
I have progressed to the point where I just have to rebuild chromium.
During the building of Chromium, a tool linked with libxml2 was created, and in
order to use it, a libmap for that was also needed.
Make sure that the libmap is applied to WRKDIR directory.
[/usr/ports/]
libxml2.so.16 libxml2.
Hello.
If this is executed twice or more times with sleep and sync in between, will
there be any change in the results?
Additionally, check whether the relevant files exist in an appropriate state
before checking.
There is some strange behavior, isn't there? :)
Regards.
On 2025/06/30 21:35, Tomoaki AOKI wrote:
On Mon, 30 Jun 2025 14:59:02 +0300
Gleb Popov wrote:
I think it'd be enough to
# cp libxml2.so /usr/local/lib/compat/
instead of dancing with libmap.conf
Yes. It is usually enough.
This is the recommendation.
cp -L /usr/local/lib/libxml2.so.2 /usr/l
On 2025/06/30 18:25, Baptiste Daroussin wrote:
On Mon 30 Jun 02:10, Jack L. wrote:
Yes, it seems the libxml2 upgrade has broken all my ports that depend on it
since those ports didn't get a bumped portrevision
portrevision has been bumped now.
Best regards,
Bapt
Thank you very much.
It is l
Um.
I don't know :), but it might just be a specific version issue with the pkg
side, so it could possibly be fixed with something like poudriere bulk -C
textproc/php81-xml.
Regards.
On 2025/06/30 16:24, void wrote:
On Mon, Jun 30, 2025 at 04:01:03PM +0900, Tatsuki Makino wrote:
Hello.
This time, it seems that the portrevision bump for the consumer port due to the
version upgrade of libxml2 has not been carried out, so poudriere-3.4.2 may not
have rebuilt the new package.
pkg check notifies that many ports have missing libraries :)
Regards.
Hello.
On 2025/06/23 1:40, A FreeBSD User wrote:
> Is there a clean way - apart from changing rc.d-script - to provide
environment variables at
> boottime?
How about defining a variable named kea_env in /etc/rc.conf?
Regards.
Hello.
On 5/30/25 16:50, Robert Clausecker wrote:
Am Wed, May 28, 2025 at 09:21:06PM -0700 schrieb Mark Millard:
Looking . . .
# file /usr/local/lib/qt5/libQt5WebEngineCore*
/usr/local/lib/qt5/libQt5WebEngineCore.prl:ASCII text, with very long
lines (770)
/usr/local/lib/qt5/libQt5WebE
Hello.
On 5/29/25 05:29, Mark Millard wrote:
First, for reference, showing libQt5WebEngine.so.5 provided
when I tried installing an official one (14.2-RELEASE context):
. . .
Yet my build experiments get a missing
libQt5WebEngineCore.so.5 dependency
(note the "Core" vs. not):
[00:06:41] [01]
On 5/20/25 01:51, Thierry Thomas wrote:
I sent an email directly to the committer who made this change, but it
seems it did not pass the SPF check, so I will post a similar message
here :)
Could you please send me (privately) the error message of this mail?
It might just be an issue with th
Hello.
I haven't used it at all, so it's all just speculation
After looking at devd.cc for about 5 seconds, it doesn't seem like the conf for
devd is being reloaded within the event_loop.
It seems that if a new conf for devd is installed, devd must be restarted.
I still do not know if that
This is probably it.
devel/py-anyio
It is interesting to see the BUILD_DEPENDS of it.
make -C /usr/ports/devel/py-anyio/ -V BUILD_DEPENDS
The setuptools that it requires will end up installing wheel044, which will
prevent the installation of the wheel(0.45) that it needs :)
After installing t
Hello.
Even if that's swapped out for devel/py-wheel044 , something else might require
devel/py-wheel, so it's not possible to install/update everything at once with
something like portmaster :)
Regards.
Hello.
I will mention something completely unrelated... :)
To achieve the same behavior as daemon of FreeBSD 10.x, the current daemon
requires argument -m 0.
However, it is not written that 0 can be used for option -m.
(I stated the same thing in bug 236117.)
pcscd (port of devel/pcsc-lite) lea
Hello.
PHP 8.2 has an extension called mhash.
It is enabled by --with-mhash.
It seems that this has not been ported yet.
If this is ported, the build will fail in environments where
xxhash(devel/xxhash) is installed.
The cause is the usual one due to the order of the include directories
specif
Hello.
It's unrelated, but since I was just touching on something similar, I'll write
a little something.
In the case of clang, we can display which directory is targeted with a command
like the following.
clang -E -x c -Wp,-v /dev/null
/usr/src/contrib/llvm-project/clang/lib/Driver/ToolChai
Hello.
I think that being hard-coded refers more to this part.
https: //svn.apache.org
/viewvc/httpd/httpd/branches/2.4.x/os/unix/unixd.h?revision=1797368&view=markup#l68
The current patch concerns whether to create users and groups with predefined
UIDs and GIDs, so it seems reasonable to prov
Hello.
The following two lines exist in INDEX-*.
tdb-1.4.9,1|/usr/ports/databases/tdb|/usr/local|Trivial
Database|/usr/ports/dat...
tdb-1.4.10,1|/usr/ports/databases/tdb1410|/usr/local|Trivial
Database|/usr/port...
I haven't checked which one takes precedence, but I guess it's bad that
packa
Hello.
My environment is that :), but I have llvm19 installed, so I will write about
what I did a little research.
In llvm19, a file that seems to have 32-bit libc++ linked.
The command looks like this.
find /usr/local/ -exec sh -c 'readelf -d -- ${1} 2> /dev/null | grep -e
libc++\\.so\\.1' a
define DEFAULT_COMPRESSION TXZ
+ define DEFAULT_COMPRESSION TZS
}
cc-check-tools ar ranlib strip
On 2025/01/26 0:18, Piotr Smyrak wrote:
On Sat, 25 Jan 2025 11:49:40 +0100
Piotr Smyrak wrote:
On Sat, 25 Jan 2025 15:08:24 +0900
Tatsuki Makino wrote:
PKG_COMPRESS_FORMAT of FreeBSD-14
Hello.
PKG_COMPRESS_FORMAT of FreeBSD-14 is tzst, so it doesn't seem to cause any
problems.
On FreeBSD-12.4-STABLE, poudriere cannot succeed pkg-repo because
archive_write_add_filter_zstd returns ARCHIVE_WARN.
So which result does FreeBSD 13.x get?
If you see a log like below, then there is a
I digress from the original.
On 2024/12/15 7:01, Kevin Oberman wrote:
Thanks, but -I has the same issue as -V. -P went away and I killed it
after 2 minutes.
Recent pkg-version takes an unusually long time.
My 12.4-STABLE needs about 5 minutes and 30 seconds for it.
My 14.2-STABLE also start
Hello.
I'm going to omit a lot of things, but about this part.
On 2024/12/15 2:51, Kevin Oberman wrote:
# pkg version -vL=
drm-61-kmod-6.1.92.1500026_3 < needs updating (index has
6.1.92.1500029_3)
gpu-firmware-intel-kmod-alderlake-20230625.1500026 < needs updating
(index has 20230625
Hello.
This may be due to the fact that the ability to build+install some ports at the
same time as buildkernel+installkernel is still incomplete.
Any fixes for the next report of Bugzilla are expected to improve flavor
support and parallelization performance.
https://bugs.freebsd.org/bugzilla
Hello.
Poudriere's log says e17ba537c, so it seems that porttree is still
pre-0ca0be4ae.
e17ba537c is also just yuri's commit.
Regards.
On 2024/12/12 6:26, Ronald Klop wrote:
https://cgit.freebsd.org/ports/commit/?h=2024Q4&id=0ca0be4ae9db2a693ceae90519cc722d93090b8e
Van: Yuri
Datum: 11 dece
Hello.
On 2024/12/11 17:28, Nuno Teixeira wrote:
I'm trying to find why there is no /etc/localtime in poudriere jails and I
need to create manually a symlink for some R-cran tests inside interactive
jail.
Any hints on how to automatize this task?
Hook is considered to be the most suitable for
Hello.
On 2024/12/01 19:59, Daniel Engberg wrote:
On 2024-11-30T23:26:38.000+01:00, Ronald Klop
wrote:
PATCH_WRKSRC+= ${WRKSRC}/src/third_party/mozjs-60/extract:mozpatch
Without affecting where other patches are applied.
This was a long time ago.
After www/seamonkey was removed, there w
Hello.
(I don't subscribe to the current mailing list, so I don't send it there.)
If changing the CPUTYPE changes the occurrence of the problem, is there
-Wl,--as-needed involved?
Not so long ago, Firefox used libraries loaded by other binaries, but that
stopped working.
As an example, recently
Hello.
On 2024/11/15 18:40, Dimitry Andric wrote:
When you must use an old version, you could add this as a patch for groups.c. It
could be conditionalized on COMPILER_VERSION < 17.
Thank you for the patch.
In my case, it was dealt with by rewriting make.conf.
FETCH_DEPENDS+= llvm-config17:
Hello.
On 2024/11/11 13:32, Gleb Popov wrote:
I'm surprised that someone is using a desktop-related package on such
an old system.
Why don't you upgrade?
I'll answer about it, but it's a work in progress.
I was going to start using it from 14.1-STABLE, but it doesn't finish when it
gets to 14
Hello.
It seems that it is mandatory to use llvm17 or higher to build gnutls 3.8.8.
If not, the following error cannot be resolved. Changing -std=c11 is powerless.
groups.c:93:2: error: initializer element is not a compile-time constant
group_x25519,
^~~~
If you are stil
Hello.
Since option PORT_LLVM is not turned on in the official version, it seems that
there is no problem with using of the bundle llvm.
However, if we want to use llvm instead of a bundle on an architecture where
the compiler causes problems, that version is no longer supported and we need
to
Hello.
bsdisks-0.36 is now something that can't be built for people like me who are
still using environments with incomplete c20 headers.
If you want to force it to build, apply a patch like the one attached below and
build it :)
Regards.
diff --git a/sysutils/bsdisks/files/patch-block.cpp
b
Hello.
It has nothing to do with ports, but I'm replying here because the status of my
subscription can only reply here.
On 2024/10/16 22:25, Tomoaki AOKI wrote:
+1.
And I cannot understand why ee is not in /rescue.
On rescue, man pages are NOT promised to be read on another vty.
Having modal
Hello again.
On 2024/10/17 18:45, Zsolt Udvari wrote:
On Thu, Oct 17, 2024 at 04:59:48PM +0900, Tatsuki Makino wrote:
And USE_TMPFS="all" seems to be one of the workarounds.
The `testport' fails with USE_TMPFS too.
This seems to be restricted by markfs in /usr/local/
Hello.
Does it mean that poudriere fails when testport or bulk -t?
And USE_TMPFS="all" seems to be one of the workarounds.
It's a guess because I'm not in a situation where I can test it right now :)
Regards.
Hello.
I feel like it would be a good to do the following...
if ($makevar{USE_LDCONFIG} ne "") {
if ($makevar{USE_LDCONFIG} eq "$makevar{PREFIX}/lib" && !$found_prefix_so) {
&perror("WARN", ... objects into \${PREFIX}/lib.");
} elsif (!$found_so) {
&perror("WARN", ... objects.");
}
Hello.
On 2024/10/09 0:16, Ronald Klop wrote:
Ok, I took a look. Does it mean just replacing %%JAVAJARDIR%% with
${JAVAJARDIR}?
It wants to use the make variable in the outer file, so it replaces it with a
%%FOO%% placemark.
Where a make variable can be used directly, use it directly.
I thi
if we want to use % as a separator, like pkg query %n%%%v, we can't
do it.
But I still don't understand how to submit pull request :)
Regards.
On 2024/09/19 15:45, Dag-Erling Smørgrav wrote:
Tatsuki Makino writes:
While looking at libpkg/pkg_config.c, I noticed that the conditio
Hello.
On 2024/09/24 20:56, Matthias Fechner wrote:
you are aware, that dhcpd is EOL since end of 2022:
https://www.isc.org/dhcp/
and was replaced by kea?
Not sure if it makes sense to put much effort into it.
I was just trying to set it up there, so it interested me.
net/kea is the port on
Hello.
# This is also a test for sending email :)
As seen from the output of set -x, I think it has been trimmed too much.
https: //cgit.freebsd.org
/ports/tree/net/isc-dhcp44-server/files/isc-dhcpd.in#n505
Is there a workaround?
dhcpd_flags="-q -4"
or
dhcpd_flags="-4 -q "
Regards.
Hello.
While looking at libpkg/pkg_config.c, I noticed that the condition of the
conditional branch in function pkg_close_devnull may not be correct.
It will not be able to close descriptor only when fd becomes 1.
I don't think it will ever be assigned 1, but I'm a little concerned about it :)
R
Hello.
rust-1.79.0 is the last version we can build with 12.4-STABLE (and also 13.2?).
Because the bootstrap compiler cannot be used due to the following output.
===> Building for rust-1.80.1
Building bootstrap
running: /wrkdirs/usr/ports/lang/rust/work/bootstrap/bin/cargo build
--manifest-path
Additional research on the changes that occur when qt6-svg is installed.
Since the target Qt::Svg will exist, svgtoqml subdirectory will be added.
https: //code.qt.io /cgit/qt/qtdeclarative.git/tree/tools/CMakeLists.txt#n70
Running the qt_internal_add_tool adds several modules to
QT_KNOWN_MODULE
Hello.
It's a subject that doesn't make sense :), but it's as follows.
When qt6-declarative-6.7.2 (x11-toolkits/qt6-declarative) is built (make
configure) in an environment where qt6-svg-6.7.2 (graphics/qt6-svg) is
installed, the following differences occur
---
usr/ports/x11-toolkits/qt6-decla
Hello.
As always, the inconsequential story provided by me :)
poudriere does a shallow clone when git porttree is created by poudriere ports
-c.
In addition, a full clone of git porttree is created by adding -D.
One more thing, -B specifies the branch, but for example,
poudriere ports -c -p git
Hello.
Piotr Smyrak wrote on 2024/08/12 06:20:
> It has been documented in MOVED file:
>
> devel/py-setuptools_scm|devel/py-setuptools-scm|2024-02-29|Package name
> changed
>
> Still if one does not know or forgets about it's existence such a move
> is quite opaque. I was also surprised by this
Thank you very much.
In this case, I think the problem with residual unnecessary packages can be
solved by using poudriere pkgclean -j jail -a.
However, using -a takes a significant amount of time.
So, we can filter the origin of the port to only what we need and run it...
Matthias Fechner wrote
If you want to read the important parts of this email first, scroll to the
bottom and read from the bottom :)
Tomoaki AOKI wrote on 2024/06/10 20:58:
> I think poudriere would be better ignoring last 3 (at least 2) digits
> of __FreeBSD_version for stable, releng and release branches, as
> API/KP
Hmmm,
I have never used poudriere to create all the packages.
However, since I was using the STABLE version for my jail, I have lost all the
packages I have built on several occasions.
It seems to happen when the value returned by the uname -KU of the jail changes.
It seems that poudriere-pkgcl
Hello.
I feel that there is a slight possibility that this upgrade may include the
following issues.
( echo py39-foo ; echo py311-foo ) | sort
However, I have only thought about the possibility and have not looked into it
yet :)
Regards.
Hello.
Matthias Fechner wrote on 2024/06/07 15:20:
> BBB_svcj_options=${BBB_svcj_options:-"net_basic"}
> But I cannot find anything in the porters handbook about that new parameter.
For these variables, it is better to search in /etc/rc.subr as well.
In /usr/src, it seems to be in /usr/src/libexe
Hello.
If it's still reproducible in an environment, does setting the automatic flag
on the pkg itself change the behavior?
.. like this.
pkg set --automatic 1 pkg
Regards.
Hello.
What I say may turn out to be bullshit :)
In short, it doesn't matter what the license is, it's how it controls
Can the FreeBSD installer include built packages?
Can packages be included on FreeBSD installer disks sold on the FreeBSD Mall (
https://www.freebsdmall.com/ ) ?
When the distfi
Hello, especially those who add some kind of tweaks to poudriere :)
In my case, I try to make it a hook script first, so it is not affected by
errexit of shell.
However, when modifying poudriere to incorporate the codes, some tmpfs are not
mounted depending on USE_TMPFS value, so poudriere exits
Hello.
Mark Millard wrote on 2024/05/05 18:15:
> [Part of my hack's notation does not generalize well across
> various USE_TMPFS= alternatives.]
>
> "<" below is what is new, ">" is what was original, in
> /usr/local/share/poudriere/common.sh :
>
> 5928,5934d5927
> < tmpfs_at_
Hello.
Mark Millard wrote on 2024/05/06 04:36:
> I've been seeing the likes of:
>
> [main-CA76-default] [2024-05-05_01h48m07s] [committing] Queued: 269 Built:
> 269 Failed: 0 Skipped: 0 Ignored: 0 Fetched: 0 Tobuild: 0Time:
> 04:48:47
> [04:48:50] Logs:
> /usr/local/poudriere/data/
Hello.
Brooks Davis wrote on 2024/05/03 04:08:
> On Thu, May 02, 2024 at 08:39:24PM +0200, Peter wrote:
>> fetch: transfer timed out
> Hmm, testing locally there's only a single commit to that path (moving
> it from Move devel/py-dateutil) so git has to read the whole commit
> history to discover
Hello.
Mark Millard wrote on 2024/04/29 10:48:
> I've modified my local poudriere-devel to have Success and Failure lines also
> report the tmpfs size at that point. Using, say, script to log the output to
I am a bit curious because I am using tmpfs to reduce HDD IO from stage phase
to package p
Baptiste Daroussin wrote on 2024/05/01 05:21:
> Le 30 avril 2024 13:18:41 GMT+02:00, Alexander Leidinger
> a écrit :
>> In my case it complains about missing shared libs for exactly those ports
>> which it wants to deinstall, but they are present in the system...
>> ---snip---
>> # pkg check -da
Hello.
Baptiste Daroussin wrote on 2024/04/30 08:50:
> Le 29 avril 2024 22:33:09 GMT+02:00, Miroslav Lachman <000.f...@quip.cz> a
> écrit :
>>
>> My notes from 2024-04-23:
>>
>> # pkg -r /tmp/be_mount.bN62 check -Ba
>> Checking all packages:
>> (0ad-0.0.26_23) /usr/local/bin/ActorEditor - require
Yasuhiro Kimura wrote on 2024/04/22 15:38:
> But the line hasn't changed since 2012. So it doesn't seems to be the
> cause of the issue.
It seems to be around here that pkg and pkg-devel make it impossible to delete
without setting the force flag.
https: //github.com
/freebsd/pkg/blob/e482b6668b
Good morning.
I think the reason why the automatic flag in pkg is not up is because there is
an .undef here.
https: //cgit.freebsd.org /ports/tree/ports-mgmt/pkg/Makefile#n72
The package pkg is guessed to have been installed by a manual operation by the
bootstrap pkg. (This has not been confirm
Hello.
Nuno Teixeira wrote on 2024/04/19 17:11:
> (...)
>
> If anyone wants to test, I've included patch in commit:
> https://cgit.freebsd.org/ports/commit/?id=89e4a1f2c1d1997b8414e2e648c7b4b7c829b63e
I don't know how to write the language R, but it seems to me that it matches
1.0e-6 even when
Hello.
bob prohaska wrote on 2024/04/15 08:02:
> The typo I had in mind was in the line
> FreeBSD 14.0-RELEASE-p5October 29, 2014FreeBSD
> ^
Does that mean this year is 2024?
This manpage is just brought over from NetBSD (or OpenBSD), w
Mathieu Arnold wrote on 2024/04/12 15:53:
> No, no, you should not do that,
:)
It is difficult to prioritize when there are two methods, WITH_*_PORTS and
WITHOUT_*_PORTS.
Perhaps we have experienced even something like "Order allow,deny Allow from
... Deny from ..." :)
I wondered if a new variab
Gleb Popov wrote on 2024/04/12 02:57:
> Should be fixed now by
> https://cgit.FreeBSD.org/ports/commit/?id=665b0219e541019b849fd5a1efdf032081dcf610
Thank you very much.
This would mean that there is more support for the use of _WITH_${feature} and
_WITHOUT_${feature} in the future.
It is expecte
Gleb Popov wrote on 2024/04/11 15:51:
> Would like me to fix it or you'll do that yourself?
Fixing this could conceivably break the functionality added in
d697653cffd19ddaf3cdee2589a830c72ab39777, so it is better to fix it carefully
or leave it alone.
There is a workaround on my part.
make -C /
Hello.
Since everyone does not need to debug all ports, I assume the definition is
made like WITH_DEBUG_PORTS+=audio/libopenshot-audio .
This has changed the variable defined from WITH_DEBUT to _WITH_DEBUG, so the
part that switches when WITH_DEBUG is defined is malfunctioning.
e.g. https: //cgi
Hello.
Tomoaki AOKI wrote on 2024/04/08 23:59:
> Is it described somewhere in the Porter's Handbook?
A fragment of the string path:dir[:target] is found when it is searched.
But will the binaries in the stage be executed directly and will the necessary
libraries be found...?
Especially in a mini
Hello Yuri-san.
science/nwchem has a pre-build target, is there a problem with ${SETENV} used
there?
New reports from bugzilla seems to cite some problems with ports with such
targets.
Regards.
Since gmake 4.4 is still x.x.1, it must still need a transition period from 4.3.
So, Yuri-san prepared devel/gmake43.
Currently this would conflict with devel/gmake.
To use this, BUILD_DEPENDS must be gmake:devel/gmake, not
gmake>=4.4.1:devel/gmake.
Or it would require construction using DEFAULT_
Hello.
I don't think this is relevant at all, but gmake does seem to have a mechanism
inside that stalls due to load average.
It seems to be possible to set it with options -l and --load-average, but I
don't think that matters, since if it is left at the default negative number,
it doesn't seem
Miroslav Lachman wrote on 2024/03/15 07:24:
> And this line should explain the order of patch and fix-shebang
> _PATCH_REAL_SEQ = ${:Uask-license} ${:Upatch-message} ${:Upatch-depends}
> ${:Ufix-shebang} ${:Udo-patch}
Regarding this, the use of :U out of the blue without a variable name doesn't
Hello.
Various answers have already been given, however, here is a command that can
produce interesting values on this.
make -C /usr/ports/ports-mgmt/pkg -V _TARGETS_STAGES
make -C /usr/ports/ports-mgmt/pkg -V _SANITY_SEQ -V _PKG_SEQ -V _FETCH_SEQ -V
_EXTRACT_SEQ -V OMITTED_HEREAFTER:P
The -d
Hello.
:On began to be used to expand the variable PATCHES in readline-8.2.10.
This determined that porttree itself would not support versions with older
make, such as 12.4-STABLE :)
However, it seems that the indexer for INDEX-12 is still at work, creating a
broken index.
The following result i
Hello.
This is related to Re: How do I clear no-longer-usable packages from poudriere?
There are some fixes for build failures when clang17 and version-script are
combined in porttree as well.
Nothing seemed to be wrong with clang15.
It seems that since clang16, the warning to be strict has been
Thank you for the various extractions.
In the official poudriere this still seems to be yes.
The version of it is 14.0-RELEASE-p5 amd64.
In my environment this is yes.
The version of it is still 12.4-STABLE amd64 :)
Since dns/libidn2 has --enable-ld-version-script in CONFIGURE_ARGS, the problem
I recently saw a similar problem and thought I had looked into something, but
this was it.
https://lists.freebsd.org/archives/freebsd-ports/2024-February/005546.html
David Wolfskill wrote on 2024/03/05 10:27:
> albert(14.0-S)[14] pdf2ps pdf/bike.pdf >/tmp/bike.ps
> ld-elf.so.1: /usr/local/lib/lib
David Wolfskill wrote on 2024/03/04 09:11:
> I am. however, also concerned about recently-installed packages that
> appear to have references to obsolete libraries, such as libc.so.6. (I
> had written earlier that I had not had libc.so.6 on my systems since 18
> February. That is technically true
Hello.
David Wolfskill wrote on 2024/03/04 03:33:
> ld-elf.so.1: /usr/local/lib/libtasn1.so.6: version LIBTASN1_0_3 required by
> /usr/local/lib/libgnutls.so.30 not defined
I don't know what it.
If you want to have them rebuilt only for those packages in poudriere, the
command will be as follow
Hello.
It seems that py39-setuptools_scm-6.4.2 must not be installed to install
py39-setuptools-scm-8.x.
So we cannot use -o of portmaster or portupgrade.
The only way seems to be to delete it beforehand with pkg delete -f.
Then reinstall the ports that have lost their dependencies, and they will
Tatsuki Makino wrote on 2024/02/27 08:48:
> Tatsuki Makino wrote on 2024/02/23 17:58:
>> Might need pkg delete -f -g alsa-plugins-\* because it causes file conflicts?
> In the case of pkg-1.20.9, It seems that alsa-plugins-* is not a glob pattern
> that picks out the intended pa
Tatsuki Makino wrote on 2024/02/23 17:58:
> Might need pkg delete -f -g alsa-plugins-\* because it causes file conflicts?
In the case of pkg-1.20.9, It seems that alsa-plugins-* is not a glob pattern
that picks out the intended package.
Others like alsa-plugins-\?\* or alsa-plugins-\[^-\]\*
Hello.
Michael Butler wrote on 2024/02/26 01:41:
> On 2/25/24 11:23, David Wolfskill wrote:
>> ld-elf.so.1: /usr/local/lib/libfstrm.so.0: version LIBFSTRM_0.2.0 required
>> by /usr/local/lib/libdns-9.18.24.so not defined
>> /usr/local/etc/rc.d/named: ERROR: named-checkconf for
>> /usr/local/etc/
Hello.
It seems that subpackaging of alsa-plugins (audio/alsa-plugins) has been
reverted.
What is the post-processing when such switching takes place?
Might need pkg delete -f -g alsa-plugins-\* because it causes file conflicts?
poudriere leaves a subpackage in the package repository. Will these
By the way, did anyone mention that setting CHECK_CHANGED_DEPS to no is a
workaround for those who don't want to build qt5-webengine or www/qt5-webengine
or qt5-webengine over and over again in poudriere?
This configuration exists in /usr/local/etc/poudriere.conf .
If we don't build the whole thi
Hello.
I have yet to read the instructions by pizzamig@freebsd presented by
arrowd@freebsd... :)
( https://lists.freebsd.org/archives/freebsd-ports/2024-February/005321.html )
John Marino (FreeBSD) wrote on 2024/02/21 00:18:
> The port builds fine.
Yes, however, it seems that all subpackage opti
Hello.
Luca Pizzamiglio wrote on 2024/02/21 05:26:
> The aforementioned `make describe` issues affect the `pkg version` issue
> you highlighted.
Yes, perhaps there is a mistake in INDEX-*, and pkg version -P will not cause
the problem (Unconfirmed as I am still in a INDEX-12 environment :) ).
R
Tatsuki Makino wrote on 2024/02/18 04:58:
> If subpackage is considered to be a FLAVOR that can be optionally changed
> (but it seems not :) ), then something is still wrong around here.
> https://github.com/freebsd/poudriere/blob/4ca881e06c33f85870127a57f9457ae6c1b69d74/src/share/
Hello.
Aryeh Friedman wrote on 2024/02/19 10:23:
> On Sun, Feb 18, 2024 at 3:21 PM Mark Millard wrote:
>>
>> I fully agree that poudriere's systematic behavior
>> rebuilds more than the likes of portmaster [but fails
>> less often].
>>
> The reason why no alternative has been produced yet is make
Hello.
It would be easier for later visitors to understand if the background of the
"working" directory were also explained.
Moin Rahman wrote on 2024/02/18 03:34:
>> On Feb 17, 2024, at 7:26 PM, bob prohaska wrote:
>> The options section of named.conf contains:
>>directory "/usr/
Hello.
Robert R. Russell wrote on 2024/02/18 02:03:
> Would that missing code cause audio/alsa-plugins to show a need for a
> rebuild with new dependencies every time poudriere runs?
If subpackage is considered to be a FLAVOR that can be optionally changed (but
it seems not :) ), then something
Hello.
I don't know what that means, but poudriere will put the BROKEN port in the
list of Ignored ports.
If the reason is BROKEN, write "is marked as broken" in the nearby Search: text
box and they will be filtered.
The BROKEN ports will always try to build if TRYBROKEN is defined or if
poudri
Hello.
I am delayed in many things on my end, so I don't have time for this kind of
thing :)
But...
I think the following conversions need to be modified to work correctly.
PKGNAME -> port origin -> PKGNAME
For example,
alsa-plugins-pulseaudio -> audio/alsa-plugins -> alsa-plugins-...
Current
Jan Beich wrote on 2024/02/03 13:03:
> See https://github.com/freebsd/poudriere/issues/1113
Thank you.
In the case of FLAVOR, the package name was also dynamically changed by a
variable (e.g.
${PYTHON_PKGNAMEPREFIX}sphinx>=0:textproc/py-sphinx@${PY_FLAVOR}), but how does
subpackage work...?
R
Hello.
I don't know what a subpackage is, but poudriere repeats the build that seems
to be one of the causes.
Since subpackage are going to be used, what else could be the cause?
1. I'm still using 12.4-STABLE :)
2. My poudriere has been fixed a bit :)
3. Dependencies are written in the style of
Hello everyone, especially those who prefer to rebuild rust over and over by
poudriere :)
Good news.
The following commits would have rebuilt gcc12.
https://cgit.freebsd.org/ports/commit/?id=ea307f60573edaffb8152c7ab9ab668a4a8785c7
However, it is mandatory to rebuild gcc12 by the following commit
mailing list to make it one example.
Regardless of the old gnu-configure, we may be affected by this for similar
reasons.
So, we need to check the list above :)
Regards.
From 793d298fc33fb0e017b3e371ec702a8f433e2da6 Mon Sep 17 00:00:00 2001
From: Tatsuki Makino
Date: Sun, 21 Jan 2024 12:00:00 +
1 - 100 of 333 matches
Mail list logo
