extra patches path #17520
extra patches path #17520
Conversation
|
just tested this with my patch from conan-io/conan-center-index#26260 - worked great! at first I tried to pass the new config value on command line, but got
is it intended that this feature is global and not per-profile? And I didn't see the new message |
|
Thanks for testing this @kambala-decapitator !
Yes, it is not possible. The
Consider for example a |
Co-authored-by: Andrey Filipenkov <[email protected]>
|
I should've used |
fix docstring of vars()
- was inconsistent before: returned False if the pattern was not found (with strict off), otherwise None
* better traces and msgs * fix tests
* working in workspace_api * fixed loader + run() error
…io#17507) * proposing new --format=file.ext feature * change approach and some simplifications * remove unused * minor changes * minor changes * simplify a bit * fix format * minor changes * fix * move colorama deinit/reinit * add out-file only if formatters * revert * wip * do not deinit/reinit if not tty * clean import * minor changes --------- Co-authored-by: czoido <[email protected]>
* new test for remote_login * unittest->pytest
* Keep sessions around between each ConanRequester Avoid creating a new session for each ConanRequester instance, this helps a lot with performance, as it avoids the overhead duplicate handshakes * Patch new cached value on ConanRequester tests * Modify adapter as needed * Fix tests * Name * Update conans/client/rest/conan_requester.py * Refactor requester tests * Reinit is not part of this PR * Refactor request usage out of app * Make max_retries a local variable
Add rest of posible ConanBasicApp usages
As the API actually accepts the remote to be `None`, this should also be reflected in the type information in order to support users.
* Updated to use maxsplit=1 * Add tests, unquote paths with spaces --------- Co-authored-by: Abril Rincón Blanco <[email protected]>
* refactor moving model conans->conan * fix tests
| @@ -69,5 +69,8 @@ def trim_conandata(conanfile, raise_if_missing=True): | |||
| if version_data is not None: | |||
| result[k] = {version: version_data} | |||
|
|
|||
| # Update the internal conanfile data too | |||
There was a problem hiding this comment.
It's not an issue in the current CCI CI, as the current flow is to first export each version separatedly, and upload it to a staging repo from which workers can then download it to build them, so the create/install is done on an already trimmed recipe conandata
There was a problem hiding this comment.
The possible issue would be if some custom command or something was reading conanfile.conandata after the export that trimmed it and somehow reading other versions information too. I doubt this happens ,but just in case
Changelog: Feature: Allow applying patches on "create" time for conan-center-index like layouts from an external centralized folder.
Docs: https://github.com/conan-io/docs/pull/XXXX
Close #17360
Some constraints:
export_conandata_patches()callconandata.ymlwith apatchesentry (otherwiseexport_conandata_patchesfails anyway)trim_conandata(). If not it will most likely crash.