Feature #1371
open
Entry model
0%
Description
I am working on integrating Ferret in Redmine. I wan't to be able to index all relevant models and there is no problem with that point. I also want to be able to index the content of files in the repositories.
I think that the proper way to do that is to have a model that represents an entry in the repository (similar to Redmine::Scm::Entry). This model could be created and maintained when syncing the changesets and linked with Changes.
Here is how my model looks like :
+--------------------+ +----------------+ | Entry | | Change | +--------------------+ +----------------+ | name | | entry_id | | content_type +---------+ | | removed | | | | source_revision | | | | source_id | | | | parent_id | | | +--------------------+ +----------------+
The source references the entry it is copied from if needed.
Was there discussions about such a model ? Is this interesting Redmine devs for other needs like attaching metadatas or discusions to an entry ?
Updated by Pierre Paysant-Le Roux almost 15 years ago
Sorry, it's not a defect but a feature...
Updated by Jean-Philippe Lang almost 15 years ago
Pierre, does your entry model reprent a file or a file at a given revision ?
Also, the term Entry can represent many things, so I would propose to name this model ScmEntry or something like this.
Updated by Pierre Paysant-Le Roux almost 15 years ago
Jean-Philippe Lang wrote:
Pierre, does your entry model reprent a file or a file at a given revision ?
Also, the termEntrycan represent many things, so I would propose to name this modelScmEntryor something like this.
My first idea was to represent a file. A RepositoryEntry is created when a file is added in the repository. Each time the file is modified, the Change model references the corresponding RepositoryEntry.
If a file is created from a copy, the RepositoryEntry references it trough source_id.
When a file is removed, the corresponding RepositoryEntry is removed (or marked as removed).
Maintaining this representation of the repository results in lots of calculation when retrieving changesets.
Updated by Toshi MARUYAMA