Workaround hard to parse $'' strings

Signed-off-by: Alfred Wingate <parona@protonmail.com>

pyproject.toml

@@ -16,7 +16,8 @@ description = "Ebuild upstream scan utility."
 license = {text = "GPL-2.0"}
 dependencies = [
     "portage",
-    "beautifulsoup4>=4.8.2"
+    "beautifulsoup4>=4.8.2",
+    "packaging"
 ]
 dynamic = ["version"]
 

src/euscan/__init__.py

@@ -53,6 +53,11 @@ BLACKLIST_PACKAGES = [
 SCANDIR_BLACKLIST_URLS = [
     "mirror://rubygems/(.*)",  # Not browsable
     "mirror://gentoo/(.*)",  # Directory too big
+    "https://dev.gentoo.org/(.*)",  # There shouldn't be releases here
+    # Waste of time to go through
+    "https://crates.io/(.*)",
+    "https://api.nuget.org/(.*)",
+    "https://myget.org/(.*)",
 ]
 
 BRUTEFORCE_BLACKLIST_PACKAGES = [

src/euscan/handlers/gitea.py

@@ -0,0 +1,70 @@
+# Copyright 2020-2024 src_prepare group
+# Distributed under the terms of the GNU General Public License v2
+
+import json
+import re
+
+import portage
+
+from euscan import helpers, mangling, output
+
+HANDLER_NAME = "gitea"
+CONFIDENCE = 100
+PRIORITY = 90
+
+# Forgejo strives to be compatible with Gitea API
+# https://forgejo.org/2024-02-forking-forward/
+
+_gitea_instances = [
+    "codeberg.org",
+    "git.osgeo.org",
+    "gitea.com",
+    "gitea.ladish.org",
+    "gitea.osmocom.org",
+    "gitea.treehouse.systems",
+]
+
+gitea_patterns = [
+    re.compile(rf"https://(?P<domain>{domain})/(?P<repository>[^/]+/[^/]+)")
+    for domain in _gitea_instances
+]
+
+
+def can_handle(pkg, url=None):
+    return url and any([re.search(pattern, url) for pattern in gitea_patterns])
+
+
+def scan_url(pkg, url, options):
+    "https://docs.gitea.com/api/1.20/#tag/repository/operation/repoListReleases"
+
+    match = [
+        re.search(pattern, url)
+        for pattern in gitea_patterns
+        if re.search(pattern, url) is not None
+    ][0]
+
+    domain = match.group("domain")
+    repository = match.group("repository")
+
+    output.einfo(f"Using Gitea API in {domain}: {repository}")
+
+    request = helpers.urlopen(f"https://{domain}/api/v1/repos/{repository}/releases")
+
+    data = json.load(request)
+
+    versions = [release["tag_name"] for release in data]
+
+    cp, ver, rev = portage.pkgsplit(pkg.cpv)
+
+    ret = []
+    for up_pv in versions:
+        pv = mangling.mangle_version(up_pv, options)
+        if helpers.version_filtered(cp, ver, pv):
+            continue
+        urls = " ".join(
+            mangling.mangle_url(release["tarball_url"], options)
+            for release in data
+            if release["tag_name"] == up_pv
+        )
+        ret.append((urls, pv, HANDLER_NAME, CONFIDENCE))
+    return ret

src/euscan/handlers/gitlab.py

@@ -0,0 +1,82 @@
+# Copyright 2020-2024 src_prepare group
+# Distributed under the terms of the GNU General Public License v2
+
+import json
+import re
+
+import portage
+
+from euscan import helpers, mangling, output
+
+HANDLER_NAME = "gitlab"
+CONFIDENCE = 100
+PRIORITY = 90
+
+_gitlab_instances = [
+    "gitlab.com",
+    "gitlab.freedesktop.org",
+    "invent.kde.org/",
+    "gitlab.gnome.org",
+    "gitlab.kitware.com",
+    "gitlab.xfce.org",
+    "code.videolan.org",
+    "gitlab.xiph.org",
+]
+
+gitlab_patterns = [
+    # Regular expression adapted from pkgcheck
+    # https://docs.gitlab.com/ee/user/reserved_names.html
+    re.compile(
+        rf"https://(?P<domain>{domain})/(?P<repository>((?!api/)\w[^/]*/)+(?!raw/)\w[^/]*)"
+    )
+    for domain in _gitlab_instances
+]
+
+
+def can_handle(pkg, url=None):
+    return url and any([re.search(pattern, url) for pattern in gitlab_patterns])
+
+
+def scan_url(pkg, url, options):
+    "https://docs.gitlab.com/ee/api/releases/index.html"
+
+    match = [
+        re.search(pattern, url)
+        for pattern in gitlab_patterns
+        if re.search(pattern, url) is not None
+    ][0]
+
+    domain = match.group("domain")
+    repository = match.group("repository")
+
+    output.einfo(f"Using GitLab REST API in {domain}: {repository}")
+
+    request = helpers.urlopen(
+        f"https://{domain}/api/v4/projects/{repository.replace('/', '%2F')}/releases"
+    )
+
+    data = json.load(request)
+
+    versions = [release["tag_name"] for release in data]
+
+    cp, ver, rev = portage.pkgsplit(pkg.cpv)
+
+    ret = []
+    for up_pv in versions:
+        pv = mangling.mangle_version(up_pv, options)
+        if helpers.version_filtered(cp, ver, pv):
+            continue
+        urls = " ".join(
+            [
+                mangling.mangle_url(source["url"], options)
+                for source in [
+                    release["assets"]["sources"]
+                    for release in data
+                    if release["tag_name"] == up_pv
+                ][0]
+                # prefer tar.bz2
+                if source["format"] == "tar.bz2"
+            ]
+        )
+        ret.append((urls, pv, HANDLER_NAME, CONFIDENCE))
+    return ret

src/euscan/handlers/gnome.py

@@ -20,7 +20,7 @@ HANDLER_NAME = "gnome"
 CONFIDENCE = 100
 PRIORITY = 90
 
-GNOME_URL_SOURCE = "http://ftp.gnome.org/pub/GNOME/sources"
+GNOME_URL_SOURCE = "https://download.gnome.org/sources"
 
 
 def can_handle(_pkg, url=None):
@@ -38,7 +38,7 @@ def guess_package(cp, url):
 
 
 def scan_url(pkg, url, options):
-    "http://ftp.gnome.org/pub/GNOME/sources/"
+    "https://download.gnome.org/sources/"
     package = {
         "data": guess_package(pkg.cpv, url),
         "type": "gnome",
@@ -55,7 +55,7 @@ def scan_pkg(pkg, options):
     content = fp.read()
     fp.close()
 
-    cache = json.loads(content, encoding="ascii")
+    cache = json.loads(content)
 
     if cache[0] != 4:
         output.eerror("Unknow cache format detected")

src/euscan/handlers/pypi.py

@@ -1,11 +1,13 @@
 # Copyright 2011 Corentin Chary <corentin.chary@gmail.com>
-# Copyright 2020-2023 src_prepare group
+# Copyright 2020-2024 src_prepare group
 # Distributed under the terms of the GNU General Public License v2
 
+import json
 import re
-import xmlrpc.client
+import urllib.error
 
 import portage
+from packaging.version import parse
 
 from euscan import helpers, mangling, output
 
@@ -29,7 +31,7 @@ def guess_package(cp, url):
 
 
 def scan_url(pkg, url, options):
-    "http://wiki.python.org/moin/PyPiXmlRpc"
+    "https://peps.python.org/pep-0691/"
 
     package = guess_package(pkg.cpv, url)
     return scan_pkg(pkg, {"data": package})
@@ -38,15 +40,23 @@ def scan_url(pkg, url, options):
 def scan_pkg(pkg, options):
     package = options["data"]
 
-    output.einfo("Using PyPi XMLRPC: " + package)
+    output.einfo("Using PyPi JSON API: " + package)
 
-    client = xmlrpc.client.ServerProxy("https://pypi.python.org/pypi")
+    try:
-    versions = client.package_releases(package)
+        fp = helpers.urlopen(f"https://pypi.org/pypi/{package}/json/")
+    except urllib.error.URLError:
+        return []
+    except OSError:
+        return []
 
-    if not versions:
+    if not fp:
-        return versions
+        return []
 
-    versions.reverse()
+    data = json.loads(fp.read())
+
+    versions = list(data["releases"].keys())
+
+    versions.sort(key=parse, reverse=True)
 
     cp, ver, rev = portage.pkgsplit(pkg.cpv)
 
@@ -55,7 +65,12 @@ def scan_pkg(pkg, options):
         pv = mangling.mangle_version(up_pv, options)
         if helpers.version_filtered(cp, ver, pv):
             continue
-        urls = client.release_urls(package, up_pv)
+        urls = " ".join(
-        urls = " ".join([mangling.mangle_url(infos["url"], options) for infos in urls])
+            [
+                mangling.mangle_url(file["url"], options)
+                for file in data["releases"][up_pv]
+                if file["packagetype"] == "sdist"
+            ]
+        )
         ret.append((urls, pv, HANDLER_NAME, CONFIDENCE))
     return ret

src/euscan/scan.py

@@ -153,6 +153,9 @@ def scan_upstream(query, on_progress=None):
     else:
         uris = pkg.environment("SRC_URI")
 
+    # Roundabout way to handle $'' strings
+    uris = uris.encode("raw_unicode_escape").decode("unicode_escape")
+
     cpv = pkg.cpv
 
     uris = parse_src_uri(uris)