Closed Gregory-Pereira closed 2 years ago
/lgtm /assign @Gregory-Pereira
New changes are detected. LGTM label has been removed.
nice work. was the s2i_assemble.patch copied from other folders? and another minor question is changes to ubi file intentional ?
Yes I copied the file for the s2i_assemble.patch living in the f34-py39
folder, and went to change anything specific to fedora version 34 or python version 3.9, but there were none. I did this because I wasn't exactly sure how these were meant to be implemented / generated, if there is a better way I am all ears.
I did not want to change the UBI assemble patches however, there was some formatting issue in them that made the pre-commit fail. I am guessing it was something in this commit that caused it, and I am not sure if changes were forced into master, the pre-commit was down, or the configuration file for precommit changed since then.
- Yes I copied the file for the s2i_assemble.patch living in the
f34-py39
folder, and went to change anything specific to fedora version 34 or python version 3.9, but there were none. I did this because I wasn't exactly sure how these were meant to be implemented / generated, if there is a better way I am all ears.
The reason why I asked this is because, the s2i_assemble.patch
is created based on the difference of the assemble
present in the base image and addition we are trying to do from our assemble script.
As the assemble
bit can differ from base image to base image, it is good to verify this once.
*. Easiest way to verify: Try to build the image from the dockerfile. (if successful the patch is in good shape)
assemble
from the base image: usually located at usr/libexec/s2i/assemble
Remove the bits we don't want , (this can be referenced from other patches)state the copied assemble script in a/assemble
state the removed bit assemble script in b/assemble
diff -u a/assemble b/assemble > assemble.patch
the resultant patch is what we require.
- I did not want to change the UBI assemble patches however, there was some formatting issue in them that made the pre-commit fail. I am guessing it was something in this commit that caused it, and I am not sure if changes were forced into master, the pre-commit was down, or the configuration file for precommit changed since then.
Thanks for the changes :)
/hold
hold till the image build is confirmed.
The s2i_assemble.patch script fails with f35 base image to be patched:
/usr/libexec/s2i ~
File assemble is read-only; trying to patch anyway
patching file assemble
Hunk #2 FAILED at 23.
1 out of 2 hunks FAILED -- saving rejects to file assemble.rej
please try to create the s2i_assemble.patch as instructed above if facing any issue, let me know
Got it, thanks Harshad, taking a look now.
Can confirm that the f35-py310
overlay now builds correctly:
$ podman build .
STEP 1/8: FROM registry.fedoraproject.org/f35/python3:0-41.container
...
STEP 8/8: USER 1001
--> Using cache 83dc730284d93fd5378c557bf4a3f1fc58647b0f8279eaab5aeb147de067fc78
--> 83dc730284d
83dc730284d93fd5378c557bf4a3f1fc58647b0f8279eaab5aeb147de067fc78
/hold
Working on formatting the README.rst
file.
/unhold Ready for review. To preview formatted doc changes, see my fork
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: fridex
The full list of commands accepted by this bot can be found here.
The pull request process is described here
Related Issues and Dependencies
Related to https://github.com/thoth-station/s2i-thoth/issues/231
Includes linting changes for all
s2i_assemble.patch
files, which is outside this PR's scope, but otherwisepre-commit
check won't pass. Additionally added documentation toREADME.rst
so people don't try to manually build the patch like I did...This introduces a breaking change
This Pull Request implements
Implements the files required to build a Thoth image of fedora35 with python3.10 using S2I.
Description
Solely creates the base image, a solver image still needs to be built from this.