This branch has been obsoleted by #17. Please close this and review that one.
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/11#issuecomment-146226805___
Rpm-maint mailing list
It's not clear that `features.h` is included when it is supposed to be, and in
fact it wasn't. This should correct that.
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/28
-- Commit Summary --
* Ensure features.h is
So, as it turns out, pretty much all libc implementations except for legacy
ones implement it as `fstat64()`, so we will use `fstat64()` unless otherwise
necessary.
Also, fix typo in checking for `_D_EXACT_NAMLEN` definition.
You can view, comment on, or merge this pull request online at:
Closed #28.
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/28#event-463602765___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint
So, I gave it another go, and while the scriptlets seem to work now,
interesting issues can come up.
It causes segfaults in Yum and DNF, and I suspect anything that uses bindings
to talk to RPM.
---
Reply to this email directly or view it on GitHub:
Per the recommendation of Nick Coghlan and Toshio Kuratomi, `pythonXegg(M)` is
being renamed to `pythonX.Ydist(M)`.
An option has also been added to add a `pythonXdist(M)` Provides for
distributions that may prefer to have it. The option `--majorver-provides` is
intended for use if only one
@proyvind Could the changelog be merged into the spec in an SRPM build or
something? Otherwise, there would need to be a way to declare that a particular
source in the spec provides the changelog.
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly
@proyvind This current approach means that you should probably have a
"ChangelogFile" or some similar property that can be used in the preamble to
override the default file name (but not the path).
---
You are receiving this because you commented.
Reply to this email directly or view it on
This PR contains both #35 and #46
After reviewing the changes from @soig in #46, I've merged it into a new pull
request, after doing some tweaking and git sorcery. This obsoletes #35 and #46.
From #35
Per the recommendation of Nick Coghlan and Toshio Kuratomi, `pythonXegg(M)` is
being
Major change: introduction of attr file to enable the dependency generator
by default.
This pull request contains @soig's attr file to enable the dependency generator
by default, adapted to also read python wheel data too.
Do we want to enable this dependency generator by default in RPM,
Closed #35.
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/35#event-517361915___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint
This pull request has been superseded by #49.
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/35#issuecomment-172375774___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
> @@ -560,6 +560,7 @@ dnl Checks for library functions.
> AC_CHECK_FUNCS(putenv)
> AC_CHECK_FUNCS(mempcpy)
> AC_CHECK_FUNCS(fdatasync)
> +AC_CHECK_DECLS(fdatasync, [], [], [#include ])
`fdatasync()` was just replaced with `fsync()` in
6151ac9a298a9fe560cf8f899dc0b0e67453446c.
---
Reply to
> @@ -230,9 +230,9 @@ rpmvar_DATA =
>
> install-exec-hook:
> @rm -f $(DESTDIR)$(bindir)/rpmquery
> - @LN_S@ ../../bin/rpm $(DESTDIR)$(bindir)/rpmquery
> + @LN_S@ rpm $(DESTDIR)$(bindir)/rpmquery
This is a problem here. I have no idea where "rpm" exists in this symlink
creation
It's totally fine to submit this as a full pull request. If necessary, the
maintainers of the project can cherry pick from a pull request. It makes life
easier, in my opinion, to have them all in one pull request, anyway.
---
Reply to this email directly or view it on GitHub:
@cgwalters I'm not sure if I would say the Groups were a failure, unless you're
willing to elaborate on why.
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/48#issuecomment-176185394___
> @@ -1120,6 +1122,23 @@ done \
> %{__patch} %{-p:-p%{-p*}} %{-q:-s}\
> %{__bzr} commit %{-q} -m %{-m*}
>
> +# Subversion
> +%__scm_setup_svn(q)\
> +%{__svnadmin} create .svnrepos\
> +%{__svn} mkdir %{-q} -m "Create directory structure."
> file://`pwd`/.svnrepos/trunk\
> +%{__svn} checkout
@ascherer Eck, I didn't override the macro properly to test, that's why... It's
all good to me :+1:
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/56#issuecomment-176159440___
Rpm-maint
> @@ -1120,6 +1122,23 @@ done \
> %{__patch} %{-p:-p%{-p*}} %{-q:-s}\
> %{__bzr} commit %{-q} -m %{-m*}
>
> +# Subversion
> +%__scm_setup_svn(q)\
> +%{__svnadmin} create .svnrepos\
> +%{__svn} mkdir %{-q} -m "Create directory structure."
> file://`pwd`/.svnrepos/trunk\
> +%{__svn} checkout
@ascherer It doesn't look like -S is supported anymore in `%autosetup`. Does
`%autosetup -S git` or something like that still work correctly?
---
Reply to this email directly or view it on GitHub:
> @@ -86,20 +90,7 @@ char * stpncpy(char * dest, const char * src, size_t n);
> #define xstrdup(_str) rstrdup((_str))
> #define _free(_ptr) rfree((_ptr))
>
> -/* Retrofit glibc __progname */
> -#if defined __GLIBC__ && __GLIBC__ >= 2
> -#if __GLIBC_MINOR__ >= 1
> -#define __progname
> @@ -230,9 +230,9 @@ rpmvar_DATA =
>
> install-exec-hook:
> @rm -f $(DESTDIR)$(bindir)/rpmquery
> - @LN_S@ ../../bin/rpm $(DESTDIR)$(bindir)/rpmquery
> + @LN_S@ rpm $(DESTDIR)$(bindir)/rpmquery
I would suggest making a new pull request that adds a configure switch for
changing
Looks good to me :+1:
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/58#issuecomment-184071948___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint
NetBSD and OS X aren't the only BSD derivatives with support for
`setprogname()` and `getprogname()`, so this PR adds the definitions for the
rest of them.
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/59
-- Commit Summary
@ffesti The actual problem is that `/bin/rpm` isn't valid when you install to
`/usr/local`, and the symlinks unconditionally assume rpm is being installed
systemwide.
---
Reply to this email directly or view it on GitHub:
@ffesti What I was suggesting is that perhaps the following might solve the
problem better:
```
@rm -f $(DESTDIR)$(rpmbindir)/rpmquery
@LN_S@ -f $(rpmbindir)/rpm $(DESTDIR)$(rpmbindir)/rpmquery
@rm -f $(DESTDIR)$(rpmbindir)/rpmverify
@LN_S@ -f $(rpmbindir)/rpm $(DESTDIR)$(rpmbindir)/rpmverify
@petere After testing the patch, it seems like it's fine to me. I'm okay with
it entering as-is. :+1:
---
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/60#issuecomment-186955079___
Rpm-maint
These days, normally RPM is built with NSS. However, it is desirable in some
environments to use BeeCrypt as an alternative to NSS.
In my case, I wanted to use BeeCrypt instead of NSS for RPM for OS X. In my
quest to compile RPM for OS X, I have found a few things that needed to be
fixed. The
The reference in `rpm.pc.in` is invalid and never gets filled in. By changing
it to the correct reference for `configure.ac` substitution, it should work as
expected.
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/64
--
In the `%apply_patches` that inspired `%autopatch`, patch application respects
the fuzz settings that are used for `%patch`. `%autopatch` and `%autosetup`
weren't using this, which led to an inconsistent patch application behavior.
You can view, comment on, or merge this pull request online at:
If it's okay, could this also be applied to the rpm 4.13 branch? It's biting us
hard in Mageia, and it'd be nice if this patch was included in the released rpm
4.13.
---
Reply to this email directly or view it on GitHub:
This pull request adds support for x86_64 for Darwin (Mac OS X) and adjusts the
logic slightly so that in the event an architecture is not defined, it throws a
warning.
Depending on how the architecture is set up, RPM may or may not properly detect
it, which is why this is a warning instead of
Looks good to me!
@ffesti @lkardos: I think this is sufficiently trivial that it can go straight
in...
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@ffesti With this patch, I can build RPM on OS X again, so I'm happy with it.
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Looks good to me as well, just a matter of @ffesti or @lkardos doing the final
review to pull it in.
---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Conan-Kudo requested changes on this pull request.
I'd prefer that the new extra scripts be disabled by default for now, rather
than enabled by default.
In addition, instead of using `dont_`, I'd suggest using `enable_` or
`disable_` and wording it appropriately. That's consistent with other
@fweimer I see, that makes much more sense.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/144#issuecomment-278350760___
I would hazard to say I would _almost never_ recommend switching to
multithreaded stuff by default, after accounting for how it would potentially
break DeltaRPM. Unless there's a way to make _that_ deterministic, I see no
pathway that would allow for widespread usage of multithreaded
@toabctl I'm not particularly a fan of the late byte-compilation technique
Debian uses, and I'd rather not propagate that down to everyone.
However, @bmwiedemann's idea of getting the .py files set to
`$SOURCE_DATE_EPOCH` to embed that in .pyc files is interesting.
--
You are receiving this
So, I've had a chance to test this on MacOS... A few points of feedback:
* The configure script _does not_ catch that OpenSSL 0.9.8zh is incompatible,
and the build fails due to undeclared things (`EVP_PKEY_CTX` and `pkey_ctx`
were mentioned before it bailed out). It needs some better checks to
Conan-Kudo commented on this pull request.
> @@ -0,0 +1,14 @@
+#!/bin/sh
+
+# If using normal root, avoid changing anything.
+if [ -z "$RPM_BUILD_ROOT" -o "$RPM_BUILD_ROOT" = "/" ]; then
+ exit 0
+fi
+
+find "$RPM_BUILD_ROOT" \( -type f -o -type l \) -name \*.la -print0 |
Yay!
--
You
Conan-Kudo commented on this pull request.
> @@ -0,0 +1,15 @@
+#!/bin/sh
+
+# If using normal root, avoid changing anything.
+if [ -z "$RPM_BUILD_ROOT" -o "$RPM_BUILD_ROOT" = "/" ]; then
+ exit 0
+fi
+
+INFODIR=`rpm --eval %{_infodir}/dir`
+
+dir="$RPM_BUILD_ROOT/$INFODIR"
+
+if [ -f $dir
I've encountered a similar issue when I was building pythonX-rpm by hand on
PCLinuxOS when I was attempting to upgrade them from 4.8.1 to 4.12.0.1. There
didn't appear to be an obvious way to fix this, as the setuptools is wont to
refer to system libraries for linking rather than the ones
@daurnimator I'd prefer to see a PR for this to make it easier to review (even
if it's marked as "WIP" for now).
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@ffesti Could you please merge this?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/154#issuecomment-279715293___
Rpm-maint
@proyvind If you are changing them to be disabled by default, then I don't
particularly care if they are there, I suppose.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Conan-Kudo approved this pull request.
The changes look good to me for Linux, and it properly fails to allow usage of
OpenSSL backend from the configure step on macOS.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@Redfoxmoon3 OpenSSL support has now been merged into rpm git master with
commit 64028f9a1c25ada8ffc7a48775f526600edcbf85.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Perhaps @hyc could provide some insight into whether the situation has improved
any since the discussions that occurred in Red Hat Bugzilla.
>From my point of view, the main issues I see are the following:
* LMDB is not independently released (it's part of the OpenLDAP source tree)
* The
Reference: https://bugzilla.redhat.com/show_bug.cgi?id=1368673
Reported-and-tested-by: Igor Gnatenko
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/80
-- Commit Summary --
* pythondistdeps.py: Ensure
Looks excellent to me!
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/82#issuecomment-241387139___
Rpm-maint mailing list
Looks good to me
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/85#issuecomment-244376171___
Rpm-maint mailing list
Looks great to me! :+1:
@ffesti @ignatenkobrain: What do you think? Can it be merged now?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Looks good to me!
@ffesti What do you think?
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/17#issuecomment-245801457___
Rpm-maint mailing list
@legionus I'd prefer to see a more detailed description in the commit itself
for future reference.
Otherwise, looks good to me!
What do you think, @ffesti?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
This PR changes `rpm2archive` so that in the event `RPMERR_ITER_END` is
returned back from `process_package()`, it will consider it successful and
return `EXIT_SUCCESS`.
Thus, shell scripts will be able to use rpm2archive the same way as most tools.
You can view, comment on, or merge this pull
Conan-Kudo approved this pull request.
Looks good to me.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@ffesti @ignatenkobrain Anyone have a chance to re-review after the rebase?
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
Tomas Orsava from the Fedora Python SIG requested that the dependency generator
support only using `pythonXdist(M)` format for both Provides and Requires, so
now this capability
exists.
You can view, comment on, or merge this pull request online at:
@pavlinamv Would it be possible to support the date+time format where the year
comes before the time? It's rather strange to see the year mentioned after the
time.
For example, your example date would be structured as:
Mon Jan 6 2016 09:02:22 CEST
--
You are receiving this because you are
No one is quite sure why there's a redundant `-python` suffix, but the module
isn't named that, and typically we want the name in the metadata to be the same
as the name of the module.
This has no effect on Python code itself, as it doesn't change the name of the
installed module used in
@proyvind The most simple case would be for SUSE, who uses `%{name}.changes`
for their changelog file name, though it's not currently in the correct form
for RPM changelogs (that can easily change, though). But there may be other
reasons (such as reusing changelog file or using a generated
These provides are specifically for packages providing AppStream files, which
are either going to be `*.appdata.xml` or `*.metainfo.xml` files in
`/usr/share/appdata` or `/usr/share/metainfo`.
The upstream AppStream specification mandates *.metainfo.xml files installed
into
@soig Removed the reference to Mageia, then.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/101#issuecomment-262324925___
Hmm, apparently there is a difference...
https://www.freedesktop.org/software/appstream/docs/chap-Metadata.html
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@ximion Would `as-metainfo` work better as a name than `appstream` or
`metainfo`? I want it to be clear what this actually is (a form of AppStream
data).
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
I think I want to just leave it as `appstream()`, as then we won't get bitten
again by changes in what actually bits of AppStream are called.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@mlschroe That's not how @ximion explained it to me. He seemed to indicate that
AppStream covers both aspects, and supersedes the old AppData spec.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
As of 57f94a582602f0353cdb17a02dc12c4461d4f32d, it's now possible
to have proper changelogs with dates and times properly set.
Thus, it makes sense to render this information by default.
You can view, comment on, or merge this pull request online at:
@pmatilai It would also be a pure extension of the existing changelog date
stamp, as it would purely append new data in the date stamp, rather than mix it
up a bit.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Conan-Kudo approved this pull request.
Looks good to me.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
I'm a bit confused. What exactly are MetaTags supposed to do? You mention
comps, are these supposed to be more like [SUSE's
patterns](https://build.opensuse.org/package/view_file/system:install:head/patterns-openSUSE/patterns-openSUSE.spec?expand=1)?
I can't quite tell what this is supposed to
For those looking from GitHub and the URL is being weirdly borked, here's the
bug: https://bugs.mageia.org/show_bug.cgi?id=9832
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@proyvind @soig @ignatenkobrain Maybe it'd be better if it was not on by
default?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Does doing the following work?
```python
rpm.expandMacro('%define _sourcedir "/home/cqi/my\ package/systemd/"')
```
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@pmatilai so you'd want instead a `-S patchbackup` backend instead of using
`-B` at `%autosetup`/`%autopatch`?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@pmatilai I'm tempted to not modify the default backend because of reasons
mentioned by @soig and @ignatenkobrain in #109. But let me see what I can do
about introducing an alternative backend that creates patch backups.
--
You are receiving this because you are subscribed to this thread.
I think most of us are keenly aware that Berkeley DB 5.x is dead. While Oracle
has Berkeley DB 6.x, it is licensed AGPLv3, which seems to make some people
rather skittish.
As a potential alternative, why not use
[LMDB](https://symas.com/products/lightning-memory-mapped-database/)? Projects
Conan-Kudo approved this pull request.
Looks good to me!
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@pmatilai Well, can you merge it and then delete the branch?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
The correct change is to get the file/libmagic project to ship a file for
pkg-config to use and stop relying on creaky manual tests.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
I believe @sgallagher from the Fedora Modularity group also was interested in
this...
Snippet from `#rpm.org` conversation back in October requesting it:
```
[Monday, October 17, 2016] [12:54:05 PM EDT] ffesti: I know this is
going to be a controversial question, but how hard would it be to
Conan-Kudo approved this pull request.
Looks fantastic to me!
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
This patch adds support for `armv5tl`, which we have been using in Mageia to
support an ARM port for a while now, and is an officially supported
architecture for Mageia 6.
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/179
I'd defer to @pmatilai and @ffesti on this, but from my point of view, I'd just
name them something like `cfg_enable`, `cfg_disable`, `cfg_with`,
`cfg_without`...
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@ignatenkobrain `armv5tl` is the base armv5 architecture without any
extensions. There's not much "real" hardware using "just" this, but it provides
maximal compatibility across the 32-bit ARM ecosystem today, as it's a
soft-float arch with just the common subset of instructions supported by
@lkundrak Could you please also add the required information to `doc/rpm.8`?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Also, we use armv5tl to support Raspberry Pi 1 and older random ARM hardware.
Modern 32-bit ARM stuff is simultaneously supported with armv7hl variant.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@myungjoo If I recall correctly, armv8 arches are the 32-bit descriptions of
aarch64. So why not just add support for armv8 arch names?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Thanks!
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/179#issuecomment-288380953___
Rpm-maint mailing list
@jaymzh Could you add information to `doc/rpm.8` about the command line option?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@n3npq Pro tip:
You can preserve formatting of diffs by wrapping it in triple backticks and
declaring format type, like so:
```diff
--- foo
+++ foo
@@ -1 +1 @@
-bar
+baz
```
Use ` ```diff ` before the first line, and use ` ``` ` after the last line.
--
You are receiving this because you are
Note that technically you cannot mix armsfp with armhfp packages as they have
different libc ABIs. I'm not sure if there is a better way to do this in RPM,
though.
As it is, though, at least the first commit in this PR is needed regardless,
since it fixes the arch compatibility tree changes
It wouldn't necessarily matter one way or the other, as the system rpm is
installing into the chroot, so it would be outside of the personality scope,
wouldn't it?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Also, the only reason they can't be mixed is because on most RPM systems, they
share `/usr/lib`. If they were in a sysroot-style FHS or Debian-style FHS, that
would not be the case.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on
That's more in @xsuchy's domain. But as far as I know, mock uses
`personality(2)` syscall to set target arch, and without the commit to declare
compat, rpm bombs out saying that the packages are for a different architecture.
--
You are receiving this because you are subscribed to this thread.
Not that I see anything particularly wrong with this, but, umm, why do you want
this?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
The with/without macros don't quite work the same way, as they don't assume
autotools underneath. Your enable/disable macros do.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@n3npq For what it's worth, to make it all unformatted text, use ` ``` ` at the
top of the block, and ` ``` ` at the bottom of the block. That way, it looks
like patches or code. :)
In addition, you can tell it to be diff style by doing ` ```diff ` for the top
of the block.
--
You are
We never needed anything because the x86_64 architecture never further mutated.
Unless AArch64 also mutates like the 32-bit ARM architectures did, I don't see
why we need `%arm64`.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on
1 - 100 of 1109 matches
Mail list logo