First, thanks for making this tool, so great to have a nice tool for doc generation.
I wanted to generated the feed/package to publish a docset on a server but without having the docset installed in my local XCode. If I set --install-docset to false, then the publish folder is not generated at all, even if --create-docset is set to true.
Am I missing something ? Can I achieve what I want ?
Thanks a lot,
Whenever you disable certain generation phase, all subsequent are also disabled. And vice versa: if you enable certain step, then all previous are also automatically enabled. However make sure you're using latest development branch version as it contains a fix regarding these options not yet on master branch.
Thanks, I thought that was the case based on another thread. I'm indeed using the latest dev branch so the fix should be in there.
I understand it makes sense to create the docset in order to be able to install it in XCode but I would see install in XCode and publish more as "being on the same level" rather than publish being subsequent to install in XCode. In this case, I would want the publish without the install.
Do you think it makes sense and could be added as a feature request ?
I get it - didn't understand correctly the first time... It could be done - each phase relies on data generated by previous so as long as files are there, it should work. Although I'm writting this from memory, so I'm not sure if there's dependency on parsed data in subsequent steps... Changed the title and will leave it as a feature request.
We want to generate DocSets in the same way on a server where the user running appledoc doesn't have access to install to Xcode. Is this feature on the roadmap?
One option: Make GBDocSetInstallGenerator.m:64 a GBLogWarn instead of an error.
I started work on project revamping to modernize it and make it easier to maintain (see my blog post about it). I'll try and include this functionality for the redesigned version from the start. Patching current one would certainly be possible, but being one man shop with other obligations, seems like too much work for the moment being. While working on redesign, I'll only accept patches from contributors and perhaps do smaller fixes myself, so the code should be quite stable - so it's likely you can get away with it for a while by commenting the suggested like on your build and have it work for you.
PS I'll upload my preliminary work on experimental branch shortly, so from then on all progress will be transparent. I'd encourage all to give it a go so I can get early feedback on it (even if it won't produce useful results from the start)