Closed LHSnow closed 1 month ago
Pagination should be working as confirmed in #136, but unlike GitHub, GitLab has very strict limitations. Will figure out what I can do to work around the complexity issue. The query itself is simple, only retrieving the size
and rawTextBlob
properties, but I guess the paths
array is too long.
I run a fairly large static site; Hugo renders close to 4000 html pages. As I was experimenting a possible migration from DecapCMS (or at least give the users an option), I ran into "unexpected error" when loading site data. Gitlab backend reports
Query being (likely from gitlab.js#L422)
This looks like the Query tries to load every single markdown file in the repository in one query, rather than what the comment says, paginated.
The query error then causes
TypeError: Cannot read properties of undefined (reading 'project'
in the following statement, crashing the entire app.Update: Reducing the config to only a single collection (the one with fewest pages) also reduces the Query complexity, but as the
media_folder
(with 800+ images) is also taken into consideration the error still occurs.