build_generation: Add LLM approach #979
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
arthurscchan
force-pushed
the
llm-build-script-generation
branch
from
b06fe30 to
9c24a86
Compare
arthurscchan
force-pushed
the
llm-build-script-generation
branch
from
de9f9af to
6c2a665
Compare
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
arthurscchan
force-pushed
the
llm-build-script-generation
branch
from
50509b3 to
ee5c775
Compare
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Contributor
Author
|
/gcbrun skip |
Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
arthurscchan
force-pushed
the
llm-build-script-generation
branch
from
f22ec69 to
509835c
Compare
Contributor
Author
|
/gcbrun skip |
DavidKorczynski
approved these changes
Apr 18, 2025
Contributor
DavidKorczynski
left a comment
DavidKorczynski
left a comment
There was a problem hiding this comment.
This looks great for a start. I think we should make sure to give the agent more control of the environment in the near future.
DavidKorczynski
added a commit
that referenced
this pull request
Apr 21, 2025
#979 caused issues when using `oss-fuzz-generator` from the installed package, because `requirements.txt` is not found in the package. Fixing this by avoiding to copy it in, and just install the library we need. Signed-off-by: David Korczynski <david@adalogics.com>
AdamKorcz
pushed a commit
that referenced
this pull request
Apr 21, 2025
#979 caused issues when using `oss-fuzz-generator` from the installed package, because `requirements.txt` is not found in the package. Fixing this by avoiding to copy it in, and just install the library we need. Signed-off-by: David Korczynski <david@adalogics.com>
DavidKorczynski
pushed a commit
that referenced
this pull request
Apr 23, 2025
Following #979, it was observed that the container tools execute the generated build script within the Docker image using a single command. If the **final** command in that script succeeds, the Docker execution will return an exit code of `0`—even if earlier commands have failed. This leads to false positives in interpreting the build result. This PR addresses the issue by prepending `set -e` to the command sequence passed to the container tool. This ensures the script exits immediately upon any command failure, preserving the non-zero exit code to correctly indicate build script failures. Additionally, this PR improves parts of the prompt templates and refines the retry-wait logic, resulting in more reliable LLM-driven build script generation. --------- Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
gabe-sherman
pushed a commit
to FuturesLab/oss-fuzz-gen-plain
that referenced
this pull request
Jan 6, 2026
This PR initialis the LLM approach for automatic build script generation of C/C++ projects. This is done by adding a separate path for the new LLM agent. Adding an extra flag --agent could trigger the new LLM agent flow for automatic build script generation. --------- Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
gabe-sherman
pushed a commit
to FuturesLab/oss-fuzz-gen-plain
that referenced
this pull request
Jan 6, 2026
google#979 caused issues when using `oss-fuzz-generator` from the installed package, because `requirements.txt` is not found in the package. Fixing this by avoiding to copy it in, and just install the library we need. Signed-off-by: David Korczynski <david@adalogics.com>
gabe-sherman
pushed a commit
to FuturesLab/oss-fuzz-gen-plain
that referenced
this pull request
Jan 6, 2026
Following google#979, it was observed that the container tools execute the generated build script within the Docker image using a single command. If the **final** command in that script succeeds, the Docker execution will return an exit code of `0`—even if earlier commands have failed. This leads to false positives in interpreting the build result. This PR addresses the issue by prepending `set -e` to the command sequence passed to the container tool. This ensures the script exits immediately upon any command failure, preserving the non-zero exit code to correctly indicate build script failures. Additionally, this PR improves parts of the prompt templates and refines the retry-wait logic, resulting in more reliable LLM-driven build script generation. --------- Signed-off-by: Arthur Chan <arthur.chan@adalogics.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR initialis the LLM approach for automatic build script generation of C/C++ projects. This is done by adding a separate path for the new LLM agent. Adding an extra flag --agent could trigger the new LLM agent flow for automatic build script generation.