fix/distro-aware-package-check (#164)

Reviewed-on: #164

tools/build

@@ -23,6 +23,8 @@ import sys
 import logging
 import subprocess
 import requests
+from requests.adapters import HTTPAdapter
+from urllib3.util.retry import Retry
 from pathlib import Path
 from typing import List, Optional
 from concurrent.futures import ThreadPoolExecutor, as_completed
@@ -33,6 +35,36 @@ import yaml
 from cerberus import Validator
 
 
+# ==================== HTTP SESSIONS ====================
+
+def _make_session(retries: int = 3, backoff_factor: float = 0.5) -> requests.Session:
+    """
+    Build a requests Session that retries on transient failures.
+
+    Retries up to `retries` times on connection errors, read timeouts, and
+    5xx / 429 responses, with exponential backoff (0s, 0.5s, 1s, 2s …).
+    Only GET requests are retried to stay safe.
+    """
+    session = requests.Session()
+    retry = Retry(
+        total=retries,
+        backoff_factor=backoff_factor,
+        status_forcelist=[429, 500, 502, 503, 504],
+        allowed_methods=["GET"],
+        raise_on_status=False,
+    )
+    adapter = HTTPAdapter(max_retries=retry)
+    session.mount("https://", adapter)
+    session.mount("http://", adapter)
+    return session
+
+# Shared sessions — one per upstream so connections are pooled and reused
+# across all package checks (including parallel builds in ThreadPoolExecutor).
+# requests.Session is thread-safe for concurrent requests.
+_gitea_session = _make_session()
+_github_session = _make_session()
+
+
 # ==================== VALIDATION SCHEMA ====================
 
 # Cerberus schema for metadata.yaml validation based on PackageMetadata and Build dataclasses
@@ -81,6 +113,11 @@ METADATA_SCHEMA = {
         'required': False,
         'empty': False
     },
+    'dist_tag': {
+        'type': 'boolean',
+        'required': False,
+        'default': False
+    },
     'builds': {
         'type': 'list',
         'required': True,
@@ -94,7 +131,7 @@ METADATA_SCHEMA = {
                     'minlength': 1,
                     'schema': {
                         'type': 'string',
-                        'allowed': ['almalinux/el8', 'almalinux/el9'],
+                        'allowed': ['almalinux/el8', 'almalinux/el9', 'fedora/42', 'fedora/43', 'fedora/44'],
                         'empty': False
                     }
                 },
@@ -146,6 +183,7 @@ class PackageMetadata:
     maintainer: str = ""
     homepage: str = ""
     license: str = ""
+    dist_tag: bool = False
     builds: List[Build] = None
 
     def __post_init__(self):
@@ -274,15 +312,14 @@ def get_github_latest_release(repo: str) -> Optional[dict]:
 
     try:
         github_token = get_github_token()
+        _github_session.headers.update({
+            'Authorization': f'token {github_token}',
+            'Accept': 'application/vnd.github.v3+json',
+        })
 
         url = f"https://api.github.com/repos/{repo}/releases/latest"
-        headers = {
-            'Authorization': f'token {github_token}',
-            'Accept': 'application/vnd.github.v3+json'
-        }
-
         logger.debug(f"Checking GitHub releases: {url}")
-        response = requests.get(url, headers=headers, timeout=30)
+        response = _github_session.get(url, timeout=15)
 
         if response.status_code == 200:
             release = response.json()
@@ -389,15 +426,14 @@ def get_github_releases_by_pattern(repo: str, pattern: str) -> Optional[dict]:
 
     try:
         github_token = get_github_token()
+        _github_session.headers.update({
+            'Authorization': f'token {github_token}',
+            'Accept': 'application/vnd.github.v3+json',
+        })
 
         url = f"https://api.github.com/repos/{repo}/releases"
-        headers = {
-            'Authorization': f'token {github_token}',
-            'Accept': 'application/vnd.github.v3+json'
-        }
-
         logger.debug(f"Checking GitHub releases with pattern '{pattern}': {url}")
-        response = requests.get(url, headers=headers, timeout=30)
+        response = _github_session.get(url, timeout=15)
 
         if response.status_code == 200:
             releases = response.json()
@@ -504,14 +540,45 @@ def normalize_version(version: str) -> str:
     return ''.join(normalized_parts)
 
 
+def get_rpm_dist_tag(distro: str) -> str:
+    """
+    Map a distro path to its RPM dist tag string.
+
+    Examples:
+        'almalinux/el9' -> 'el9'
+        'fedora/43'     -> 'fc43'
+    """
+    if distro.startswith('almalinux/'):
+        return distro.split('/', 1)[1]
+    if distro.startswith('fedora/'):
+        return 'fc' + distro.split('/', 1)[1]
+    return ''
+
+
+def effective_release(base_release: str, distro: str, use_dist_tag: bool) -> str:
+    """Return the release string to use for building and registry checks.
+
+    When use_dist_tag is True the dist tag is appended so different distros
+    produce distinct entries in the package registry (e.g. '1.el9', '1.fc43').
+    """
+    if not use_dist_tag:
+        return base_release
+    tag = get_rpm_dist_tag(distro)
+    return f"{base_release}.{tag}" if tag else base_release
+
+
 def check_package_exists(package_name: str, version: str, release: str) -> bool:
     """
     Check if a package version exists in the Gitea package registry.
 
+    Distro disambiguation is handled by the caller via the release string:
+    when dist_tag is enabled the caller appends the dist tag to release
+    (e.g. '1.el9', '1.fc43') so each distro maps to a unique version entry.
+
     Args:
         package_name: Name of the package
         version: Version string
-        release: Release number
+        release: Effective release (may include dist tag suffix)
 
     Returns:
         True if package exists, False otherwise
@@ -536,17 +603,17 @@ def check_package_exists(package_name: str, version: str, release: str) -> bool:
             f"{package_type}/{package_name}/{full_version}"
         )
 
-        headers = {'Authorization': f'token {gitea_token}'}
+        _gitea_session.headers.update({'Authorization': f'token {gitea_token}'})
 
         logger.debug(f"Checking package existence: {url}")
-        response = requests.get(url, headers=headers, timeout=30)
+        response = _gitea_session.get(url, timeout=10)
 
         if response.status_code == 200:
             package_info = response.json()
-            # Package exists if we get package info back
             exists = bool(package_info.get('id'))
             logger.debug(f"Package {package_name}:{full_version} {'exists' if exists else 'not found'}")
             return exists
+
         elif response.status_code == 404:
             logger.debug(f"Package {package_name}:{full_version} not found (404)")
             return False
@@ -1241,6 +1308,7 @@ class Builder:
 
                 package_name = metadata.get('name', package_dir.name)
                 build_configs = metadata.get('builds', [])
+                use_dist_tag = metadata.get('dist_tag', False)
 
                 if not build_configs:
                     self.logger.warning(f"No builds in metadata.yaml for {package_name}")
@@ -1259,7 +1327,8 @@ class Builder:
                             if repositories and base_image and version and release:
                                 # Use the first repository as the distro identifier
                                 build_distro = repositories[0] if repositories else 'unknown'
-                                packages.append(PackageInfo(package_name, version, str(release), package_dir, build_distro, base_image))
+                                rel = effective_release(str(release), build_distro, use_dist_tag)
+                                packages.append(PackageInfo(package_name, version, rel, package_dir, build_distro, base_image))
                 else:
                     # Build for specific distro
                     for build_config in build_configs:
@@ -1271,7 +1340,8 @@ class Builder:
 
                             # Check if the target distro matches any repository
                             if distro in repositories and base_image and version and release:
-                                packages.append(PackageInfo(package_name, version, str(release), package_dir, distro, base_image))
+                                rel = effective_release(str(release), distro, use_dist_tag)
+                                packages.append(PackageInfo(package_name, version, rel, package_dir, distro, base_image))
                                 break
                     else:
                         # If no matching distro found, log a warning
@@ -1324,6 +1394,8 @@ class Builder:
             with open(metadata_file, 'r') as f:
                 metadata = yaml.safe_load(f)
 
+            use_dist_tag = metadata.get('dist_tag', False)
+
             # Find base image and validate version/release for the specified distro
             build_configs = metadata.get('builds', [])
             base_image = None
@@ -1367,7 +1439,8 @@ class Builder:
             self.logger.error(f"Error reading metadata.yaml for {package}: {e}")
             return False
 
-        package_info = PackageInfo(package, version, release, package_dir, distro, base_image)
+        rel = effective_release(release, distro, use_dist_tag)
+        package_info = PackageInfo(package, version, rel, package_dir, distro, base_image)
         return self._build_package(package_info, dry_run, force, native, buildah)
 
     def build_all(self, dry_run: bool = False, force: bool = False, parallel: int = 4, distro: str = 'el/9', native: bool = False, buildah: bool = False) -> bool: