diff --git a/swh/lister/cgit/lister.py b/swh/lister/cgit/lister.py --- a/swh/lister/cgit/lister.py +++ b/swh/lister/cgit/lister.py @@ -169,6 +169,25 @@ ) return None + # ensure the active tab is the summary tab + # on some repository, the default tab can be the about one where there is no clone info + if ( + "summary" not in repository_url + ): # avoid infinite loop if the repository doesn't exist + tab = bs.find("table", {"class": "tabs"}) + if tab: + active_tab = tab.find("a", {"class": "active"}) + if active_tab and active_tab.text != "summary": + logger.debug( + "%s : Active tab is not the summary, trying to load the summary page", + repository_url, + ) + return self._get_origin_from_repository_url( + f"{repository_url}/summary" + ) + else: + logger.debug("No summary tab found on %s", repository_url) + # origin urls are listed on the repository page # TODO check if forcing https is better or not ? # @@ -177,6 +196,7 @@ urls = [x["href"] for x in bs.find_all("a", {"rel": "vcs-git"})] if not urls: + logger.debug("No git urls found on %s", repository_url) return None # look for the http/https url, if any, and use it as origin_url diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/Readme.md b/swh/lister/cgit/tests/data/https_git.acdw.net/Readme.md new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/Readme.md @@ -0,0 +1 @@ +These files are a partial dump of http://git.savannah.gnu.org/cgit diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/cgit b/swh/lister/cgit/tests/data/https_git.acdw.net/cgit new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/cgit @@ -0,0 +1,40 @@ + + + +friendware by acdw + + + + + +

+ + + + +

+index

+ + + + + +

Name	Description	Idle	Links
sfeed	My sfeed scripts +	28 min.	summary log tree
foo	Non existing repository +	28 min.	summary log tree

+ +

+ + diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/foo b/swh/lister/cgit/tests/data/https_git.acdw.net/foo new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/foo @@ -0,0 +1,33 @@ + + + + + + + + + +

+ + + + +

+index

No repositories found

+ +

+ + diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/foo_summary b/swh/lister/cgit/tests/data/https_git.acdw.net/foo_summary new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/foo_summary @@ -0,0 +1,33 @@ + + + + + + + + + +

+ + + + +

+index

No repositories found

+ +

+ + diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed b/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed @@ -0,0 +1,49 @@ + + + +sfeed - My sfeed scripts + + + + + + + + +

+ + + + +

+about summary refs log tree commit diff stats

sfeed

Turns out, sfeed is cool! You can see what this repo generates at https://acdw.casa/planet/.

+ +

+ + diff --git a/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed_summary b/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed_summary new file mode 100644 --- /dev/null +++ b/swh/lister/cgit/tests/data/https_git.acdw.net/sfeed_summary @@ -0,0 +1,63 @@ + + + +sfeed - My sfeed scripts + + + + + + + + +

+ + + + +

+about summary refs log tree commit diff stats

+ + + + + + + + + + + + + + + +

Branch	Commit message	Author	Age
main	Add APOD	Case Duckworth	38 min.


Age	Commit message	Author
38 min.	Add APOD HEAD main	Case Duckworth
4 days	Change fresh item colors	Case Duckworth
4 days	Indentation	Case Duckworth
5 days	Add Tab Completion	Case Duckworth
5 days	Add Lonnie Johnson	Case Duckworth
7 days	Add miniature calendar; metafilter	Case Duckworth
9 days	Add active listening	Case Duckworth
10 days	Add tilde.town blog	Case Duckworth
12 days	Add zserge	Case Duckworth
12 days	Remove duplicate	Case Duckworth
[...]

Clone
https://git.acdw.net/sfeed

+ +

+ + diff --git a/swh/lister/cgit/tests/test_lister.py b/swh/lister/cgit/tests/test_lister.py --- a/swh/lister/cgit/tests/test_lister.py +++ b/swh/lister/cgit/tests/test_lister.py @@ -265,3 +265,16 @@ # we should have 16 repos (listed on 3 pages) assert len(repos) == 3 assert len(flattened_repos) == 16 + + +def test_lister_cgit_summary_not_default(requests_mock_datadir, swh_scheduler): + """cgit lister returns git url when the default repository tab is not the summary""" + + url = "https://git.acdw.net/cgit" + + lister_cgit = CGitLister(swh_scheduler, url=url) + + stats = lister_cgit.run() + + expected_nb_origins = 1 + assert stats == ListerStats(pages=1, origins=expected_nb_origins)