Tags: zimbatm/upm
Tags
Bump golang.org/x/net from 0.17.0 to 0.23.0 (replit#263) * Bump golang.org/x/net from 0.17.0 to 0.23.0 Bumps [golang.org/x/net](https://github.com/golang/net) from 0.17.0 to 0.23.0. - [Commits](golang/net@v0.17.0...v0.23.0) --- updated-dependencies: - dependency-name: golang.org/x/net dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> * update go mods dep hash --------- Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Toby Ho <[email protected]>
Ensure that we handle the optionality of the trailing newline in requ… …irements.txt (replit#237)
[bug] Only dereference into [tool.poetry] if defined (replit#236)
Stripping out pkgs.glibc from imports (replit#231) [Reports online](https://discourse.nixos.org/t/conflict-between-glibc-and-glibclocales-tl-dr-i-messed-it-up/7608) suggest that glibc isn't intended to behave like a normal package, so we shouldn't automatically install it.
chore/gen_pypi_cache overhaul (replit#225) * Actually log error when reading BQ * Adding "test-one" command for easier debugging * Typo * Attempt to demystify magic * Clarity comment Initially I didn't realize that the whole error info was being accumulated for serialization into retval * Relax requirement for intermediate modules * Replace install_diff.py with os.WalkDir * Nicer time info in the summary * Add a lower bound to the modules we accumulate * Adding -timeout to constrain how much time we spend building packages * Omit empty * Retry installing via pip up to 5 times * Better error message * Bumping download_stats.json * exec.Command can return nil apparently * "test" should not also "updatepkgs" Conflation of these reduces clarity of the domain of responsibility * Regenerating pkgs.json * Updating README to clarify subcommand execution order
Run upm on Replit (replit#89) Co-authored-by: Faris Masad <>
Update python module guesser mappings (replit#86) * [WIP] Generate python module mapping as a single step * Use json for cache and do codegen * Renamed files * Linting * Fix concurreny bugs * Limit package index size for testing * Support whl, zip, and egg files together * Print error messages to stderr * Dump generated code to file * Fetch entire pypi package index * Actually save errors for printing * Improve locking performance * Add cli flags * Update generate command to use new flags * Actually add packages into cache * Commit bigquery backing data * Linting * Cache zip files on disk * Cleaned up error logging * Linting * Fetch pypi stats from bigquery * Improve feedback * Tweak dist parsing * Install packages to determine provided modules * WIP * organized sources; allow gen_pypi_map.go to have multiple commands * WIP * WIP * Modified guess algorithm using dependencies list. Regression tests. * Added cache files * updated go.mod and go.sum * remove unneeded file * updated map consumer * Added back downloads map * fixed pypi_map.gen.go target * update vendorSha256 * Removed cache directory from repo. Now it is temporary but we have a persistent pkg.json * added cache dir to ignores * cleaned up naming * removed outdated comments * used relative path * Updated path to module_to_pypi.json * added readmes and updated test.py to read from download_stats.json * Added module_to_pypi.legacy.json for regression tests * updated regression test to say whether we've added the package * up to 10k for regression tests * added package test error case * fix * updated readme * updated test runs * fixed name handling * fixed another naming issue * updated test to get more precise results Co-authored-by: Zach Anderson <[email protected]>
Add --no-update to poetry lock (replit#83)
Update supported languages We should auto-generate this table based on backends
PreviousNext