Get these globs right.

Author: Patrick Thomson

build/example_repos.bzl

@@ -4,28 +4,34 @@ load(
     "new_git_repository",
 )
 
-all_example_repos = {
+_all_example_repos = {
     "numpy": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "058851c5cfc98f50f11237b1c13d77cfd1f40475",
         "repo": "numpy/numpy",
         "since": "",
-        "excludes": ["data structures"],
+        "excludes": [],
     },
     "python": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "c6be53e1c43f870f5364eef1499ee1b411c966fb",
         "repo": "thealgorithms/python",
-        "since": "",
-        "excludes": [],
+        "since": "1548508158 +0800",
+        "excludes": [
+            "**/data structures/*",
+            "**/binary tree/*",
+            "**/graphs/*",
+            "**/Random Forest*/*",
+            "**/* */*",
+        ],
     },
     "flask": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "0b5b4a66ef99c8b91569dd9b9b34911834689d3f",
         "repo": "pallets/flask",
@@ -34,7 +40,7 @@ all_example_repos = {
     },
     "httpie": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "358342d1c915d6462a080a77aefbb20166d0bd5d",
         "repo": "jakubroztocil/httpie",
@@ -43,16 +49,16 @@ all_example_repos = {
     },
     "keras": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "e59570ae26670f788d6c649191031e4a8824f955",
         "repo": "keras-team/keras",
-        "since": "",
+        "since": "1548927621 +0530",
         "excludes": [],
     },
     "requests": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "64bde6582d9b49e9345d9b8df16aaa26dc372d13",
         "repo": "requests/requests",
@@ -61,7 +67,7 @@ all_example_repos = {
     },
     "scikit-learn": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "d0f63a760d9993a7f68cfc5e1a075700d67c53d3",
         "repo": "scikit-learn/scikit-learn",
@@ -70,16 +76,16 @@ all_example_repos = {
     },
     "scrapy": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "65d631329a1434ec013f24341e4b8520241aec70",
         "repo": "scrapy/scrapy",
-        "since": "",
+        "since": "1548908933 -0300",
         "excludes": [],
     },
     "pytorch": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "c865d46736db4afff51690a712e35ed8e3899490",
         "repo": "pytorch/pytorch",
@@ -88,11 +94,11 @@ all_example_repos = {
     },
     "certbot": {
         "data": [
-            "**/*.rb",
+            "**/*.py",
         ],
         "commit": "bb8222200a8cbd39a3ce9584ce6dfed6c5d05228",
         "repo": "certbot/certbot",
-        "since": "",
+        "since": "1549052531 -0800",
         "excludes": [],
     },
     "spec": {
@@ -106,18 +112,18 @@ all_example_repos = {
     },
     "desktop": {
         "data": [
-            "**/*.rb",
+            "**/*.[tj]s",
         ],
-        "commit": "d1324f56d02dd9afca5d2e9da545905",
+        "commit": "d1324f56d02dd9afca5d2e9da545905a7d41d671",
         "repo": "desktop/desktop",
-        "since": "",
+        "since": "1523834029 +1000",
         "excludes": [],
     },
 }
 
-all_repo_deps = ["@" + k + "//:src" for (k, v) in all_example_repos.items()]
+semantic_external_test_repositories = ["@" + k + "//:src" for (k, v) in _all_example_repos.items()]
 
-def example_repo(name, data, excludes, commit, repo):
+def _example_repo(name, data, excludes, commit, repo, since):
     new_git_repository(
         name = name,
         build_file_content = """
@@ -129,8 +135,9 @@ filegroup(
 """.format(data, excludes),
         commit = commit,
         remote = "https://github.com/{}.git".format(repo),
+        shallow_since = since,
     )
 
 def declare_example_repos():
-    for k, v in all_example_repos.items():
-        example_repo(name = k, data = v["data"], excludes = v["excludes"], commit = v["commit"], repo = v["repo"])
+    for k, v in _all_example_repos.items():
+        _example_repo(name = k, data = v["data"], excludes = v["excludes"], commit = v["commit"], repo = v["repo"], since = v["since"])

docs/build.md

@@ -10,7 +10,7 @@ The Semantic project supports builds with the Bazel build system. This is unconv
 
 Assuming you're on macOS, run the script located at ~script/bootstrap-bazel~. This uses Homebrew to install Bazel and creates the `.bazel-cache` directory.
 
-The first time you run `bazel build`
+The first time you run `bazel build`, it'll take some time, as Bazel will compile all of Stackage. Fear not: you won't have to do this again.
 
 ## `cabal` → `stack` cheatsheet
 
@@ -36,7 +36,7 @@ If this seems complicated, don't worry: most of the time you'll be able to skip
 ## Things to know
 
 1. **Don't generally run `bazel clean`**. Since Bazel builds are reproducible, there's very little reason to clean, unless somehow your whole cache got irrevocably corrupted.
-2. **You can load a REPL for any target by appending `@repl`.**
+2. **You can load a REPL for any target by appending `@repl`.** (with the exception of the language packages, due to [this](https://github.com/tweag/rules_haskell/issues/1377)).
 3. **Some packages come with GHC and are not loaded from Stackage**. These include `base`, `containers`, and [others](https://downloads.haskell.org/~ghc/latest/docs/html/users_guide/8.10.1-notes.html?highlight=bytestring#included-libraries). To depend on those packages, you use `//:base`, `//:containers`, etc. They are specified in the `BAZEL.build` at the project root. You probably won't need to add any more.
 
 ## Quick reference links
@@ -58,4 +58,6 @@ The default `.bazelrc` file imports a `.bazelrc.local` file if it's present; use
 
 ## Custom rules
 
-We have two custom rules, defined in `build/common.bzl`. The first, `tree_sitter_node_types_archive`, uses the `http_archive` rule to download a specified tree-sitter grammar's `node-types.json` file. These calls declare new top-level targets, so they're only present in the top-level `WORKSPACE` file. The second, `semantic_language_library`, takes care of the boilerplate associated with declaring a target for a `semantic-LANG` language package (as these packages' contents are identical, their target declarations are almost identical).
+We have two common custom rules, defined in `build/common.bzl`. The first, `tree_sitter_node_types_archive`, uses the `http_archive` rule to download a specified tree-sitter grammar's `node-types.json` file. These calls declare new top-level targets, so they're only present in the top-level `WORKSPACE` file. The second, `semantic_language_library`, takes care of the boilerplate associated with declaring a target for a `semantic-LANG` language package (as these packages' contents are identical, their target declarations are almost identical).
+
+For the purposes of setting up the examples upon which the `parse-examples` test depends, we have code in `build/example_repos.bzl` which defines them, checks them out, and computes the set of target names. You shouldn't need to change or modify this, unless you're adding new repos.

semantic/BUILD.bazel

@@ -21,7 +21,7 @@ load(
 )
 load(
     "//:build/example_repos.bzl",
-    "all_repo_deps",
+    "semantic_external_test_repositories",
 )
 
 haskell_toolchain_library(name = "ghc-prim")
@@ -197,7 +197,7 @@ haskell_test(
         "test/fixtures/javascript/**/*.js",
         "test/fixtures/typescript/**/*.ts",
         "test/fixtures/cli/*.json",
-    ]) + all_repo_deps,
+    ]) + semantic_external_test_repositories,
     deps = semantic_common_dependencies + [
         ":fixtureshim",
         "//:base",