Correct handling of tagged prereleases

.circleci/config.yml

@@ -182,17 +182,12 @@ commands:
   # --------------------------------------------------------------------------
   # Build Commands
 
-  setup_prerelease_commit_hash:
+  setup_prerelease:
     steps:
       - run:
-          name: Store commit hash and prerelease
+          name: Store prerelease suffix
           command: |
-            if [[ $CIRCLE_BRANCH == release || -n $CIRCLE_TAG ]]; then
-              echo -n > prerelease.txt;
-            else
-              date -u +"nightly.%Y.%-m.%-d" > prerelease.txt;
-            fi
-            echo -n "$CIRCLE_SHA1" > commit_hash.txt
+            "scripts/prerelease_suffix.sh" nightly "$CIRCLE_TAG" > prerelease.txt
 
   install_and_check_minimum_requirements:
     parameters:
@@ -1006,6 +1001,7 @@ jobs:
     <<: *base_ubuntu2404_xlarge
     steps:
       - build
+      - run: cat prerelease.txt
 
   # x64 ASAN build, for testing for memory related bugs
   b_ubu_asan: &b_ubu_asan
@@ -1115,7 +1111,7 @@ jobs:
     <<: *base_ubuntu_clang_large
     steps:
       - checkout
-      - setup_prerelease_commit_hash
+      - setup_prerelease
       - run_build_ossfuzz
       - persist_ossfuzz_executables_to_workspace
       - matrix_notify_failure_unless_pr
@@ -1225,7 +1221,7 @@ jobs:
     <<: *base_ubuntu2404_small
     steps:
       - checkout
-      - setup_prerelease_commit_hash
+      - setup_prerelease
       - run:
           name: Install build system dependencies
           command: |

docs/conf.py

@@ -126,10 +126,12 @@ def get_github_username_repo(url):
 with open('../CMakeLists.txt', 'r', encoding='utf8') as f:
     version = re.search('PROJECT_VERSION "([^"]+)"', f.read()).group(1)
 # The full version, including alpha/beta/rc tags.
-if not os.path.isfile('../prerelease.txt') or os.path.getsize('../prerelease.txt') == 0:
+if os.path.isfile('../prerelease.txt') and os.path.getsize('../prerelease.txt') == 0:
     release = version
+elif os.path.isfile('../prerelease.txt'):
+    with open('../prerelease.txt', 'r', encoding='utf8') as prerelease_file:
+        release = version + '-' + prerelease_file.read()
 else:
-    # This is a prerelease version
     release = version + '-develop'
 
 # The language for content autogenerated by Sphinx. Refer to documentation

scripts/build.sh

@@ -10,8 +10,10 @@ else
     BUILD_TYPE="$1"
 fi
 
-if [[ $(git tag --points-at HEAD 2> /dev/null) == v* ]]; then
-    touch "${ROOTDIR}/prerelease.txt"
+# Intentionally not using prerelease_suffix.sh here. We do not want lingering prerelease.txt in
+# dev builds, accidentally overriding version when someone runs the build manually.
+if [[ $(git tag --points-at HEAD 2> /dev/null) =~ ^v[0-9.]+$ ]]; then
+    echo -n > prerelease.txt
 fi
 
 mkdir -p "$BUILDDIR"

scripts/ci/build.sh

@@ -9,23 +9,12 @@ prerelease_source="${1:-ci}"
 
 cd "${ROOTDIR}"
 
-if [[ $CIRCLE_BRANCH == release || -n $CIRCLE_TAG || -n $FORCE_RELEASE ]]; then
-    echo -n > prerelease.txt
-else
-    # Use last commit date rather than build date to avoid ending up with builds for
-    # different platforms having different version strings (and therefore producing different bytecode)
-    # if the CI is triggered just before midnight.
-    TZ=UTC git show --quiet --date="format-local:%Y.%-m.%-d" --format="${prerelease_source}.%cd" > prerelease.txt
-fi
-
-if [[ -n $CIRCLE_SHA1 ]]; then
-    echo -n "$CIRCLE_SHA1" > commit_hash.txt
-fi
+"${ROOTDIR}/scripts/prerelease_suffix.sh" "$prerelease_source" "$CIRCLE_TAG" > prerelease.txt
 
 mkdir -p build
 cd build
 
-[[ -n $COVERAGE && $CIRCLE_BRANCH != release && -z $CIRCLE_TAG ]] && CMAKE_OPTIONS="$CMAKE_OPTIONS -DCOVERAGE=ON"
+[[ -n $COVERAGE && -z $CIRCLE_TAG ]] && CMAKE_OPTIONS="$CMAKE_OPTIONS -DCOVERAGE=ON"
 
 # shellcheck disable=SC2086
 cmake .. -DCMAKE_BUILD_TYPE="${CMAKE_BUILD_TYPE:-Release}" $CMAKE_OPTIONS

scripts/ci/build_emscripten.sh

@@ -45,17 +45,7 @@ function build() {
 
     cd "${ROOT_DIR}"
 
-    if [[ $CIRCLE_BRANCH == release || -n $CIRCLE_TAG || -n $FORCE_RELEASE || $(git tag --points-at HEAD 2> /dev/null) == v* ]]; then
-        echo -n > prerelease.txt
-    else
-        # Use last commit date rather than build date to avoid ending up with builds for
-        # different platforms having different version strings (and therefore producing different bytecode)
-        # if the CI is triggered just before midnight.
-        TZ=UTC git show --quiet --date="format-local:%Y.%-m.%-d" --format="${prerelease_source}.%cd" > prerelease.txt
-    fi
-    if [[ -n $CIRCLE_SHA1 ]]; then
-        echo -n "$CIRCLE_SHA1" > commit_hash.txt
-    fi
+    "${SCRIPT_DIR}/prerelease_suffix.sh" "$prerelease_source" "$(git tag --points-at HEAD 2> /dev/null)" > prerelease.txt
 
     # Disable warnings for unqualified `move()` calls, introduced and enabled by
     # default in clang-16 which is what the emscripten docker image uses.

scripts/ci/build_win.sh

@@ -8,20 +8,7 @@ FORCE_RELEASE="${FORCE_RELEASE:-}"
 CIRCLE_TAG="${CIRCLE_TAG:-}"
 
 cd "$ROOTDIR"
-if [[ $FORCE_RELEASE != "" || $CIRCLE_TAG != "" ]]; then
-    echo -n > prerelease.txt
-else
-    # Use last commit date rather than build date to avoid ending up with builds for
-    # different platforms having different version strings (and therefore producing different bytecode)
-    # if the CI is triggered just before midnight.
-    # NOTE: The -local suffix makes git not use the timezone from the commit but instead convert to
-    # local one, which we explicitly set to UTC.
-    # NOTE: git --date is supposed to support the %-m/%-d format too, but it does not seem to
-    # work on Windows. Without it we get leading zeros for month and day.
-    last_commit_date=$(TZ=UTC git show --quiet --date="format-local:%Y-%m-%d" --format="%cd")
-    last_commit_date_stripped=$(date --date="$last_commit_date" "+%Y.%-m.%-d")
-    echo -n "${prerelease_source}.${last_commit_date_stripped}" > prerelease.txt
-fi
+"${ROOTDIR}/scripts/prerelease_suffix.sh" "$prerelease_source" "$CIRCLE_TAG" > prerelease.txt
 
 mkdir -p build/
 cd build/

scripts/create_source_tarball.sh

@@ -2,18 +2,36 @@
 set -euo pipefail
 
 REPO_ROOT="$(dirname "$0")"/..
+# shellcheck source=scripts/common.sh
+source "${REPO_ROOT}/scripts/common.sh"
+
 cd "$REPO_ROOT"
 version=$(scripts/get_version.sh)
 commit_hash=$(git rev-parse --short=8 HEAD)
-commit_date=$(git show --format=%ci HEAD | head -n 1 | cut - -b1-10 | sed -e 's/-0?/./' | sed -e 's/-0?/./')
 
-# File exists and has zero size -> not a prerelease
-if [[ -e prerelease.txt && ! -s prerelease.txt ]]; then
-    version_string="$version"
+if [[ -e prerelease.txt ]]; then
+    prerelease_suffix=$(cat prerelease.txt)
+    if [[ $prerelease_suffix == "" ]]; then
+        # File exists and has zero size -> not a prerelease
+        version_string="$version"
+    elif [[ $prerelease_suffix == pre.* ]]; then
+        # Tagged prerelease -> unambiguous, so commit hash not needed
+        version_string="${version}-${prerelease_suffix}"
+    else
+        # Nightly/develop/other prerelease -> include commit hash
+        version_string="${version}-${prerelease_suffix}-${commit_hash}"
+    fi
 else
+    # Nightly/develop/other prerelease -> include commit hash + default prerelease suffix
+    commit_date=$(TZ=UTC git show --quiet --date="format-local:%Y.%-m.%-d" --format="%cd")
     version_string="${version}-nightly-${commit_date}-${commit_hash}"
 fi
 
+# The only purpose of commit_hash.txt is to make it possible to build the compiler without git.
+# It is not meant as an override of the real hash.
+[[ ! -e commit_hash.txt ]] || \
+    fail "commit_hash.txt is present in the repository root, but will not be used to override the commit hash for the source package."
+
 TEMPDIR=$(mktemp -d -t "solc-src-tarball-XXXXXX")
 SOLDIR="${TEMPDIR}/solidity_${version_string}/"
 mkdir "$SOLDIR"

scripts/prerelease_suffix.sh

@@ -0,0 +1,29 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+(( $# <= 2 )) || { >&2 echo "Usage: $0 [PRERELEASE_SOURCE] [GIT_TAG]"; exit 1; }
+prerelease_source="${1:-nightly}"
+git_tag="${2:-}"
+FORCE_RELEASE="${FORCE_RELEASE:-}"
+
+GNU_DATE="date"
+if [[ "$OSTYPE" == "darwin"* ]]; then
+    GNU_DATE=gdate
+fi
+
+if [[ $FORCE_RELEASE != "" || $git_tag =~ ^v[0-9.]+$ ]]; then
+    echo -n
+elif [[ $git_tag =~ ^v[0-9.]+-pre. ]]; then
+    echo -n "pre.${git_tag#*-pre.}"
+else
+    # Use last commit date rather than build date to avoid ending up with builds for
+    # different platforms having different version strings (and therefore producing different bytecode)
+    # if the CI is triggered just before midnight.
+    # NOTE: The -local suffix makes git not use the timezone from the commit but instead convert to
+    # local one, which we explicitly set to UTC.
+    # NOTE: git --date is supposed to support the %-m/%-d format too, but it does not seem to
+    # work on Windows. Without it we get leading zeros for month and day.
+    last_commit_date=$(TZ=UTC git show --quiet --date="format-local:%Y-%m-%d" --format="%cd")
+    last_commit_date_stripped=$("$GNU_DATE" --date "$last_commit_date" "+%Y.%-m.%-d")
+    echo -n "${prerelease_source}.${last_commit_date_stripped}"
+fi