aboutsummaryrefslogtreecommitdiff
path: root/cmd/podman
Commit message (Collapse)AuthorAge
...
* | | Merge pull request #14539 from Luap99/completion5OpenShift Merge Robot2022-06-09
|\ \ \ | | | | | | | | shell completion: fix problems with container path completion
| * | | shell completion: fix problems with container path completionPaul Holzinger2022-06-09
| | |/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When you try to complete a path which exists and it is a file the completion logic did not check the parent dir for other matching file names. To fix that we have to check if the current completion is not a dir and use the parent dir in this case. See the updated test for an example why this is required. Also make sure directories are correctly completed, the shell always adds the "/" as suffix to signal the user that this path is a directory. In this case we do not want to automatically add a space. When the path is a regular file we want the space after the suggestion since there is nothing more to complete. This better matches the normal default shell completion. The test were changed to not assume any particular ordering since this is irrelevant for the shell completion script and there is no guarantee about the ordering. Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* / | compat api: fix regressions from "Swagger refactor/cleanup"Paul Holzinger2022-06-09
|/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For some reason commit 5b79cf15a022 moved the container create options parsing from cmd/podman/common to pkg/api/handlers. However it did not remove the old code. Unfortunately it moved the code from an outdated version and did not update it before this commit was merged. Therefore a couple of regressions were introduced. I manually compared both versions and found three missing bugfixes. I fixed the network test again that was changed in bce97a3b5dd1. We want bridge as default even as rootless. Sine the test is not run as rootless in CI the regression was not caught. Also the no hosts test never worked since it was missing the import check if the hosts file exists. I don't think we can check for the volume parsing change since this only works on windows/wsl. Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* | Merge pull request #14466 from mheon/fix_9075OpenShift Merge Robot2022-06-03
|\ \ | | | | | | Improve robustness of `podman system reset`
| * | Improve robustness of `podman system reset`Matthew Heon2022-06-03
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Firstly, reset is now managed by the runtime itself as a part of initialization. This ensures that it can be used even with runtimes that would otherwise fail to be created - most notably, when the user has changed a core path (runroot/root/tmpdir/staticdir). Secondly, we now attempt a best-effort removal even if the store completely fails to be configured. Third, we now hold the alive lock for the entire reset operation. This ensures that no other Podman process can start while we are running a system reset, and removes any possibility of a race where a user tries to create containers or pull images while we are trying to perform a reset. [NO NEW TESTS NEEDED] we do not test reset last I checked. Fixes #9075 Signed-off-by: Matthew Heon <mheon@redhat.com>
* | | Merge pull request #14460 from cipherboy/align-docker-podman-load-outputOpenShift Merge Robot2022-06-02
|\ \ \ | | | | | | | | Align docker load and podman load output
| * | | Align docker load and podman load outputAlexander Scheel2022-06-02
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The comma-separated podman load output isn't conducive for using the subsequent images. For tarballs with multiple images, the comma separator must be manually identified and a suitable range identified. Docker CLI on the other hand, has one image identifier per line: Loaded image: repo1/name1:latest Loaded image: repo1/name1:tag1 Loaded image: repo2/name2:tag1 (as of Docker version 20.10.16, build aa7e414). Switch `podman load` to this format for consistency and usability. [NO NEW TESTS NEEDED] Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* | | | changed megabyte to mebibyteKarthik Elango2022-06-02
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In podman run --help, the message said megabyte, gigabyte, etc. In reality podman takes mebibytes, gibibytes, etc. [CI:DOCS] Signed-off-by: Karthik Elango <kelango@redhat.com>
* | | | Merge pull request #14457 from Luap99/completion4OpenShift Merge Robot2022-06-02
|\ \ \ \ | |_|_|/ |/| | | shell completion for paths inside the image/container
| * | | shell completion for paths inside the image/containerPaul Holzinger2022-06-02
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add shell completion for paths inside the container or image. Currently podman run IMAGE [TAB] only uses the default shell completion which suggests paths on the host. This is fine for some cases but often the user wants a path which only exists in the image/container. This commits adds support for that. Both podman create/run can now complete the paths from the image, podman cp ctr:... now completes paths from the actual container. Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* | | | Merge pull request #14301 from rhatdan/volumeOpenShift Merge Robot2022-06-02
|\ \ \ \ | |/ / / |/| | | Support setting image_volume_mode in containers.conf
| * | | Support setting image_volume_mode in containers.confDaniel J Walsh2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | Fixes: https://github.com/containers/podman/issues/14230 Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* | | | Merge pull request #14417 from Luap99/machine-sshOpenShift Merge Robot2022-06-02
|\ \ \ \ | |_|_|/ |/| | | podman machine ssh: set correct exit code
| * | | podman machine ssh: set correct exit codePaul Holzinger2022-05-30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Forward the ssh exit code to the podman caller. This is useful for scripts. Use the same logic as podman unshare. Fixes #14401 Signed-off-by: Paul Holzinger <pholzing@redhat.com>
| * | | fix bad import path for cmd/podman/utilsPaul Holzinger2022-05-30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Libpod or packages under /pkg should never import from /cmd/... This will quickly result in import cycles and weird code paths. Also there is no reason to use this special code we can just use syscall.SIGHUB as SIGNAL. Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* | | | podman volume export/import: give better errorPaul Holzinger2022-05-30
|/ / / | | | | | | | | | | | | | | | | | | | | | | | | When the volume does not exist we should output an error stating so and not some generic one. Fixes #14411 Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* | | Merge pull request #14320 from flouthoc/build-honor-squash-and-layersOpenShift Merge Robot2022-05-27
|\ \ \ | | | | | | | | build: allow using `cache` explicitly with `--squash-all` using `--layers`
| * | | vendor: bump buildah to v1.26.1-0.20220524184833-5500333c2e06Aditya R2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | Bump buildah to v1.26.1-0.20220524184833-5500333c2e06 Signed-off-by: Aditya R <arajan@redhat.com>
| * | | build: allow using cache explicitly with --squash-all using --layersAditya R2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Buildah already supports using `--layers` with `--squash` after https://github.com/containers/buildah/pull/3674 if user wants to do so hence podman must honor similar configuration in `--squash-all` behaviour if user wants to using cache. PS: We cannot alter behaviour of `podman build --squash` for docker-compat reasons hence this feature can be easily supported by `--squash-all`. Closes: https://github.com/containers/buildah/issues/4011 Signed-off-by: Aditya R <arajan@redhat.com>
| * | | cmd, build: remove redundant squash processing logicAditya R2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Same block contains similar lines above this is not needed as this looks redundant. [NO NEW TESTS NEEDED] [NO TESTS NEEDED] Signed-off-by: Aditya R <arajan@redhat.com>
* | | | Merge pull request #14382 from jwhonce/wip/init_todoOpenShift Merge Robot2022-05-26
|\ \ \ \ | | | | | | | | | | Remove TODO comment
| * | | | Remove TODO commentJhon Honce2022-05-26
| |/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Allowing custom flags to provider has the potential to break all the hand-crafted commands currently in use. This could become a support nightmare. ```release-note NONE ``` [NO NEW TESTS NEEDED] Signed-off-by: Jhon Honce <jhonce@redhat.com>
* | | | Merge pull request #14374 from umohnani8/todo-3OpenShift Merge Robot2022-05-26
|\ \ \ \ | |_|/ / |/| | | Combine the CheckAllLatest CID and PodID functions
| * | | Combine the CheckAllLatest CID and PodID functionsUrvashi Mohnani2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These two functions were doing the exact same thing just with cidfile and pod-id-file separately. Combine the functionality to one function to remove repetative code. Fix the TODO in cmd/podman/validate/args.go [NO NEW TESTS NEEDED] Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
* | | | Merge pull request #14373 from umohnani8/todo-2OpenShift Merge Robot2022-05-26
|\ \ \ \ | |_|/ / |/| | | Fix TODO in pod/ps.go and parse/net.go
| * | | Fix TODO in parse/net.goUrvashi Mohnani2022-05-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix up the parseEnv function to differentiate between a label and env when parsing. Don't do a system lookup when parsing labels. [NO NEW TESTS NEEDED] Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
| * | | Remove TODO from pods/ps.goUrvashi Mohnani2022-05-26
| |/ / | | | | | | | | | | | | | | | | | | | | | The TODO has already been fixed. Filters is now a []string [NO NEW TESTS NEEDED] Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
* | | Merge pull request #14366 from jwhonce/wip/rm_todoOpenShift Merge Robot2022-05-25
|\ \ \ | |/ / |/| | Support remote deadlock errors in rm
| * | Support remote deadlock errors in rmJhon Honce2022-05-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Refactor test for deadlock by comparing error text vs. actual ErrWillDeadlock constant. When running with --remote the error constant will always be not equal to the error returned by the API. ```release-note NONE ``` [NO NEW TESTS NEEDED] Signed-off-by: Jhon Honce <jhonce@redhat.com>
* | | Remove unused archive flag from diff commandsJhon Honce2022-05-25
|/ / | | | | | | | | | | | | | | | | | | | | | | | | * Option left in images/diff.go CLI as comment implies it is needed for backwards compatibility. ```release-note NONE ``` [NO NEW TESTS NEEDED] Signed-off-by: Jhon Honce <jhonce@redhat.com>
* | Merge pull request #14357 from rhatdan/codespellOpenShift Merge Robot2022-05-25
|\ \ | | | | | | Fix codespell errors
| * | Fix codespell errorsDaniel J Walsh2022-05-25
| | | | | | | | | | | | | | | | | | [NO NEW TESTS NEEDED] Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* | | Merge pull request #14333 from rhatdan/podOpenShift Merge Robot2022-05-25
|\ \ \ | |/ / |/| | Allow podman pod create --share +pid
| * | Allow podman pod create --share +pidDaniel J Walsh2022-05-24
| | | | | | | | | | | | | | | | | | Fixes: https://github.com/containers/podman/issues/13422 Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* | | Merge pull request #14334 from rhatdan/pod1OpenShift Merge Robot2022-05-24
|\ \ \ | | | | | | | | Allow podman pod create to accept name argument
| * | | Allow podman pod create to accept name argumentDaniel J Walsh2022-05-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | I am constantly attempting to add the podname to the last argument to podman pod create. Allowing this makes it match podman volume create and podman network create. It does not match podman container create, since podman container create arguments specify the arguments to run with the container. Still need to support the --name option for backwards compatibility. Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* | | | Merge pull request #14346 from edsantiago/helpmsg_testsOpenShift Merge Robot2022-05-24
|\ \ \ \ | |_|/ / |/| | | help-message system test: catch more cases
| * | | help-message system test: catch more casesEd Santiago2022-05-24
| | |/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Look for and prevent lower-case arg descriptions: podman cmd [arg] - Look for and prevent optional-mandatory misordering: podman cmd [ARG] ARG - Tighter whitespace checks (and fix podman pod ps) - simplify a no-longer-necessary mess! #8635 fixed the horrible "CONTAINER | IMAGE" strings (with spaces), so there's no longer a need to special-case those. The one-extra-arg check is now much cleaner. Minor refactoring. Signed-off-by: Ed Santiago <santiago@redhat.com>
* | | Merge pull request #14330 from Luap99/completion3OpenShift Merge Robot2022-05-24
|\ \ \ | | | | | | | | cmd/podman/common/completion.go: fix FIXMEs
| * | | cmd/podman/common/completion.go: fix FIXMEsPaul Holzinger2022-05-24
| |/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is no good way to recommend labels for podman container runlabel. Add the missing max-size log option. These are the only documented options so the completion should not suggest something different. Add proper --stop-signal completion. It will now complete all supported signal names both upper and lowercase depending on the user input. Also it work with and without the SIG prefix. Fixing the TODOs in this file are more complicated since they describe bigger features. [NO NEW TESTS NEEDED] Signed-off-by: Paul Holzinger <pholzing@redhat.com>
* | | Merge pull request #14327 from rhatdan/commonOpenShift Merge Robot2022-05-24
|\ \ \ | | | | | | | | Use containers/common/pkg/util.StringToSlice
| * | | Use containers/common/pkg/util.StringToSliceDaniel J Walsh2022-05-23
| | | | | | | | | | | | | | | | | | | | | | | | [NO NEW TESTS NEEDED] Just code cleanup for better reuse Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
* | | | Merge pull request #14348 from jwhonce/wip/root_fixmeOpenShift Merge Robot2022-05-24
|\ \ \ \ | |_|_|/ |/| | | [CI:DOCS] Remove TODO re: storage-driver
| * | | Remove TODO re: storage-driverJhon Honce2022-05-24
| | |/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently this list is not exported from c/storage and the group decided this will not be changed. ```release-note NONE ``` Signed-off-by: Jhon Honce <jhonce@redhat.com>
* | | podman image mount: print pretty tableValentin Rothberg2022-05-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | Make sure that `podman image mount` prints a pretty table unless there is only argument passed and without a custom format. Fixing a TODO item brought me to the specific code location and revealed the fart in the logic. Signed-off-by: Valentin Rothberg <vrothberg@redhat.com>
* | | autocomplete `podman search --filter`Valentin Rothberg2022-05-24
|/ / | | | | | | Signed-off-by: Valentin Rothberg <vrothberg@redhat.com>
* | Merge pull request #14263 from Luap99/completionOpenShift Merge Robot2022-05-23
|\ \ | |/ |/| shell completion --format: various improvements
| * shell completion --format: add help to function with argsPaul Holzinger2022-05-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | From a template users POV it is not importent when they use a struct field or method. They only notice the difference when the function requires arguments. So lets be nice and let the user know that this method requires arguments via the help text. This is how it now looks like when the completion descriptions are enabled on bash: ``` $ bin/podman ps --format {{.Created.A {{.Created.AddDate (This is a function and requires 3 arguments) {{.Created.After (This is a function and requires 1 argument) {{.Created.Add (This is a function and requires 1 argument) {{.Created.AppendFormat (This is a function and requires 2 arguments) ``` Signed-off-by: Paul Holzinger <pholzing@redhat.com>
| * shell completion --format: support maps and functionsPaul Holzinger2022-05-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently we only support structs in a template string like this: `{{.var1.test.` -> this meams that test must be a struct field on var1. Now with this var1 and test could also be either a map or function which returns a struct. A actual example: `podman container inspect --format {{.NetworkSettings.Networks.netname.` Now we can complete the struct fileds after netname. Note that this cannot complete map keys since they are empty by default, so it is impossible to get them in the completion logic. Also this fixes a panic with embeeded nil structs Fixes #14223 Signed-off-by: Paul Holzinger <pholzing@redhat.com>
| * shell completion --format: fix embedded struct handlingPaul Holzinger2022-05-18
| | | | | | | | | | | | | | | | | | | | | | | | When a struct is embeeded it is possible that we end up with same names but different types, this results in incorrect completions. The go template logic always preferes the actual field/method name before the one from the embedded one. Thefore the completion logic should do the same. First get all method/fields names from the struct and then only add the field names from the embedded struct when they are not already present in the list. Signed-off-by: Paul Holzinger <pholzing@redhat.com>
generated by cgit v1.2.3-54-g00ecf (git 2.39.0) at 2024-12-29 14:54:48 +0000