**Describe the bug**
I need to use command `rpm --root /path/to/image/fakeroot -ivh xxx.rpm xxx.rpm`
as normal user to install RPM packages into my fakeroot for my embedded system.
When running with RPM 4.19.1.1, the above command runs perfectly.
However, after I upgrade my Fedora 40 to Fedora 41,
@dmnks commented on this pull request.
> +}
+
+rpmRC keystore_openpgp_cert_d::load_keys(rpmtxn txn, rpmKeyring keyring)
+{
+return load_keys_from_glob(txn, keyring, "%{_keyringpath}/*/*");
+}
+
+rpmRC keystore_openpgp_cert_d::delete_key(rpmtxn txn, rpmPubkey key)
+{
+rpmRC rc = RPMRC_NOT
@dmnks commented on this pull request.
>
-if (!rc && replace) {
- /* find and delete the old pubkey entry */
- if (delete_key(txn, fp, keyfmt) == RPMRC_NOTFOUND) {
- /* make sure an old, short keyid version gets removed */
- delete_key(txn, fp+32, keyfmt);
-
@dmnks commented on this pull request.
>
-if (!rc && replace) {
- /* find and delete the old pubkey entry */
- if (delete_key(txn, fp, keyfmt) == RPMRC_NOTFOUND) {
- /* make sure an old, short keyid version gets removed */
- delete_key(txn, fp+32, keyfmt);
-
@pmatilai commented on this pull request.
> @@ -4,6 +4,8 @@
#include "system.h"
+#include
Oh, this shouldn't be here, no vectors are used in the patch.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3439#pullrequestreview-242987
@pmatilai pushed 0 commits.
--
View it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442/files/36903e9b7cca205a0efa84138290082a1a8ae467..2aed2f5e3c08e0f083041957f53353fd8abc469b
You are receiving this because you are subscribed to this thread.
Message ID:
___
That AC proposal neglets to mention just how exactly this behavior is
triggered, which is rather critical piece of information. Are new APIs and cli
switches being added - which ones?
Stuff like "move new instance to proper place" is an uninteresting
implementation detail, what we're really int
Replying to the first question myself:
`rpmtxnImportPubkey()` is where `replace` is decided, based on whether the key
can be found. That kinda alleviates the need for an atomic rename since only
*new* keys would be inserted with it disabled.
That still doesn't ensure true atomicity but this is
Oh and FWIW, these are intentionally not returning filesystem::path items as
those have some peculiar charasteristics that don't play well with our
codebase. Like appending an absolute path to an existing path resetting the
whole path to the part that was supposed to be appended, ie `"aa" / "/bb
No, other than "use it where it makes sense". Quite apparently it doesn't in
all sorts of scenarios.
Beyond that something like having entire rpm use filesystem::path for all paths
is sooo far in the future it's not worth speculating at this point
:sweat_smile:
--
Reply to this email
The one debug leftover dropped and more elaborate commit message on the main
commit (about why filesystem::path is not used etc)
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#issuecomment-2470502384
You are receiving this because y
The pun was pretty funny though :smile:
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#issuecomment-2470291751
You are receiving this because you are subscribed to this thread.
Message ID: __
@Conan-Kudo commented on this pull request.
> @@ -306,79 +307,34 @@ char *rpmCleanPath(char * path)
return path;
}
-/* Merge 3 args into path, any or all of which may be a url. */
+/* Merge 3 args into path */
Did we drop support for URLs?
--
Reply to this email directly or view it on
@dmnks commented on this pull request.
>
-if (!rc && replace) {
- /* find and delete the old pubkey entry */
- if (delete_key(txn, fp, keyfmt) == RPMRC_NOTFOUND) {
- /* make sure an old, short keyid version gets removed */
- delete_key(txn, fp+32, keyfmt);
-
@dmnks commented on this pull request.
> +rpmRC keystore_openpgp_cert_d::load_keys(rpmtxn txn, rpmKeyring keyring)
+{
+return load_keys_from_glob(txn, keyring, "%{_keyringpath}/*/*");
+}
+
+rpmRC keystore_openpgp_cert_d::delete_key(rpmtxn txn, rpmPubkey key)
+{
+rpmRC rc = RPMRC_NOTFOUND
The latest version you've pushed looks fine. The only remaining questions are:
* Do we care that `replace` can be set to `0` and thus not use `rename(2)` for
insertion? The latter is mandated by the spec, although it seems redundant,
with the writelock in place
*
[How](https://github.com/rpm-sof
@ffesti commented on this pull request.
> @@ -6,6 +6,7 @@
#include
#include
#include
+#include
Why do we needs this here. With all the new functions going into
rpmmacro_internal.hh ? Or should this be in some later patch?
--
Reply to this email directly or view it on GitHub:
https://g
Ack, let's merge this then.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3437#issuecomment-2470234777
You are receiving this because you are subscribed to this thread.
Message ID: ___
Rpm-mai
Closed #3341 as completed via 6e19c1602c1118938d556453ee33357ea0006d3d.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/3341#event-15262766543
You are receiving this because you are subscribed to this thread.
Message ID:
__
Closed #3341 as completed via #3437.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/3341#event-15262766568
You are receiving this because you are subscribed to this thread.
Message ID:
___
R
Merged #3437 into master.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3437#event-15262766156
You are receiving this because you are subscribed to this thread.
Message ID:
___
Rpm-maint mail
@pmatilai pushed 6 commits.
515406906e9cca7a1b79015b9f6b02f325269134 Don't try to redefine isblank()
c4e29e0d06e6af14255261d2f3c7b277de71ffef Add internal C++ counterparts for
path manipulation functions
1af476167a26b0ad9ecd0c29203bf5c3baf8b8dc Eliminate uses of rpmCleanPath() in
librpmbuild
@Conan-Kudo approved this pull request.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#pullrequestreview-2429419631
You are receiving this because you are subscribed to this thread.
Message ID: __
Merged #3444 into master.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3444#event-15263662496
You are receiving this because you are subscribed to this thread.
Message ID:
___
Rpm-maint mail
Is there a path (pardon the pun) to us being able to adopt
`std::filesystem::path` for file paths?
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#issuecomment-2470268424
You are receiving this because you are subscribed to this thre
Well, it kinda does. By not replacing you make sure you fail if a key was added
in the mean time instead of just overwriting it without merging.
Also as far as rpm is concerned we do have the transaction log (which is
represented by the txn parameter passed down) whihc protects parallel access
I would suggest to not get hung up on the names especially as long as they are
only internal. We might even still change the name in the configuration as soon
as we know what back ends we'll end up with and what exactly they do.
--
Reply to this email directly or view it on GitHub:
https://gith
The interesting part of this PR are new rpm::join_path(), rpm::expand_path()
and rpm::normalize_path() functions which are more powerful C++ native
counterparts of rpmGenPath(), rpmGetPath() and rpmCleanPath(), and hopefully
with more meaningful names too.
rpmGenPath() and rpmGetPath() use the
@Conan-Kudo commented on this pull request.
> @@ -306,79 +307,34 @@ char *rpmCleanPath(char * path)
return path;
}
-/* Merge 3 args into path, any or all of which may be a url. */
+/* Merge 3 args into path */
So I guess stuff like `rpm -ivh
https://kojipkgs.fedoraproject.org//packages
Related: #3341
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/3444
-- Commit Summary --
* Test reading openpgpg backend with sq
-- File Changes --
M tests/rpmsigdig.at (8)
-- Patch Links --
https://github.com/rpm-so
@pmatilai commented on this pull request.
> @@ -306,79 +307,34 @@ char *rpmCleanPath(char * path)
return path;
}
-/* Merge 3 args into path, any or all of which may be a url. */
+/* Merge 3 args into path */
Yes.
--
Reply to this email directly or view it on GitHub:
https://github.com
@pmatilai commented on this pull request.
> @@ -6,6 +6,7 @@
#include
#include
#include
+#include
Oh, that's just a debugging leftover. Thanks for spotting, I'll drop.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#discussi
After digging into rpm's git log, I found following commit:
```
commit 9c96c5d4ca376b998be9919e70f74d0995c4df2e
Author: Panu Matilainen
Date: Fri Sep 15 14:38:35 2023 +0300
Axe the experimental non-privileged chroot support from b4c832ca
While the chroot call itself works nicely fo
@dmnks commented on this pull request.
>
-if (!rc && replace) {
- /* find and delete the old pubkey entry */
- if (delete_key(txn, fp, keyfmt) == RPMRC_NOTFOUND) {
- /* make sure an old, short keyid version gets removed */
- delete_key(txn, fp+32, keyfmt);
-
@pmatilai commented on this pull request.
> @@ -306,79 +307,34 @@ char *rpmCleanPath(char * path)
return path;
}
-/* Merge 3 args into path, any or all of which may be a url. */
+/* Merge 3 args into path */
Not really, the URL support never worked at all (as is explained in the commit
@pmatilai commented on this pull request.
> @@ -306,79 +307,34 @@ char *rpmCleanPath(char * path)
return path;
}
-/* Merge 3 args into path, any or all of which may be a url. */
+/* Merge 3 args into path */
So it was a lucky case of being SO broken nobody could have possibly used it
f
This is part of M4, not specific to our test-suite, but it's a question
that comes up often, and everyone (myself included) tends to forget how
it's done.
You can view, comment on, or merge this pull request online at:
https://github.com/rpm-software-management/rpm/pull/3443
-- Commit Summary
@dmnks pushed 1 commit.
ba96f1a73ec1b4a306b9b2daef6094898f6ad30d Add note about square brackets to
tests README
--
View it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3443/files/a26c6a332711f5ebaf75c2c4faaecb7aeb084bf7..ba96f1a73ec1b4a306b9b2daef6094898f6ad30d
You are receiv
Merged #3443 into master.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3443#event-15263376259
You are receiving this because you are subscribed to this thread.
Message ID:
___
Rpm-maint mail
@pmatilai pushed 1 commit.
36903e9b7cca205a0efa84138290082a1a8ae467 Deprecate rpmCleanPath() in the
public API
--
View it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442/files/8b831e4e58fcdc64958e367d7214cee4e5089fdc..36903e9b7cca205a0efa84138290082a1a8ae467
You are receivi
@pmatilai commented on this pull request.
> @@ -6,6 +6,7 @@
#include
#include
#include
+#include
Dropped.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#discussion_r1838081294
You are receiving this because you are subscri
Also FWIW I'm totally okay with splitting the latter commits to another PR if
this seems a bit much at a time.
--
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/3442#issuecomment-2470557022
You are receiving this because you are subscribed
42 matches
Mail list logo
