Closed howardjones closed 3 years ago
Log($"Converting Markdown to HTML for {files.Count} files");
Meant I could at least tell if it thought it was working.
As far as I can see, alphabetically. And actually a few sub-directories do have a .order file, but it doesn't include all the pages, and the index in the wiki itself shows all the pages, so it's unclear what the .order file is actually doing. I think those files first, and then alphabetical, which might explain an issue I had with sorting in the wiki, now I think about it :-) I think it depends on whether you add pages directly to the git repo, or via the wiki UI.
I did get a simple file search working last night. It doesn't do the ordering as described above though, yet.
Hmmm, well I have a version that I can run build in VSCode that does what I want (convert all .md files found, in alphabetical order), but only by shipping around 176 files... trying to publish with -p:PublishSingleFile=true -p:PublishTrimmed=true
crashes with "Found multiple entries with the same BundleRelativePath". I don't know what a BundleRelativePath is, or where that might come from. I don't see anything duplicated in the .csproj file.
The contents tree of the PDF is a bit strange too - lots, but not all, of the internal headers, in a funny order. Will probably investigate further there.
how is the order of the pages then determined? Random? Alphabetically?
FWIW:
https://docs.microsoft.com/en-us/azure/devops/project/wiki/wiki-file-structure?view=azure-devops
When there's no .order file the pages are sorted alphabetically.
Closing. Once I got the file search working, the output wasn't really what I'd hoped for. Some python and pandoc did it for me in the end.
What was the difference/what is missing?
I thought I'd try this on our Project wiki, only to get a blank document! Two observations: 1) verbose mode isn't very verbose (e.g. would be good if it told you what/how many files it had processed) 2) our wiki doesn't have any .order files, so the exporter doesn't do anything
I'm going to try and write a replacement/fallback for ReadOrderFiles that just walks the directories.