Added couple search options today... (use these at the search box, top-right of the obex table)
# followed by the object number (Can be comma or space delim'd list)
@ followed by an author name (or some characters)
Examples:
#2784
#2784,#2783, etc...
#OB2784,#ob2783
@Wuerfel_21
@ada
Caveats/Notes:
Object number search is exact match relationship (ie. #278 would not return all those objects starting with that prefix). With some time for indexing adjustment that could be improved.
Author search works with either the full (exact match) user-id, or a partial free-text search based on the imported "git author" names or user names. It's not perfect, but it's a start that works for most simple-searches. There are some funny result combinations possible that can be ironed out later
The zips remove a lot of useful functionality and restrict exposure to AI tools!
In regard to AI, Stephen is working on some neat future solutions for IDE integration that would also benefit from access to unzipped code.
-
For the consumer - Every OBEX download page includes a zip download button - so downloading an entire object zip is always a choice. And with unzipped sources being uploaded, we can also also choose to browse the code online, copy/paste code sections or download single files, etc..
A code adjustment was made to the dropdown tags list near the end of the OBEX entry/edit form.
Sometimes that dropdown would fail validation and prevent an object being posted or updated (ie. submitted).
When it happened, the dropdown wasn't very clear about being the devil on the page either!
That bug has now been squashed
Though... Please let me know if anyone discovers this or any other issue reoccurring!
ps. I took a look at adding multi-delete for the attachments box; to prevent posting back on each attachment action. After some analysis I can see it will take a longer free day to get that dealt with, so that task remains at the top of the todo list. It's not so much just stopping the post-back bubbling, but rather adjusting some actions that are called to ensure integrity of things like the file store, archive and zip builder that happen behind the scenes.
Comments
Added couple search options today... (use these at the search box, top-right of the obex table)
# followed by the object number (Can be comma or space delim'd list)
@ followed by an author name (or some characters)
Examples:
#2784
#2784,#2783, etc...
#OB2784,#ob2783
@Wuerfel_21
@ada
Caveats/Notes:
Object number search is exact match relationship (ie. #278 would not return all those objects starting with that prefix). With some time for indexing adjustment that could be improved.
Author search works with either the full (exact match) user-id, or a partial free-text search based on the imported "git author" names or user names. It's not perfect, but it's a start that works for most simple-searches. There are some funny result combinations possible that can be ironed out later
Tip / Reminder (for all OBEX users)
Please upload folders instead of zip archives
The zips remove a lot of useful functionality and restrict exposure to AI tools!
In regard to AI, Stephen is working on some neat future solutions for IDE integration that would also benefit from access to unzipped code.
-
For the consumer - Every OBEX download page includes a zip download button - so downloading an entire object zip is always a choice. And with unzipped sources being uploaded, we can also also choose to browse the code online, copy/paste code sections or download single files, etc..
Update report:
A code adjustment was made to the dropdown tags list near the end of the OBEX entry/edit form.
Sometimes that dropdown would fail validation and prevent an object being posted or updated (ie. submitted).
When it happened, the dropdown wasn't very clear about being the devil on the page either!
That bug has now been squashed
Though... Please let me know if anyone discovers this or any other issue reoccurring!
ps. I took a look at adding multi-delete for the attachments box; to prevent posting back on each attachment action. After some analysis I can see it will take a longer free day to get that dealt with, so that task remains at the top of the todo list. It's not so much just stopping the post-back bubbling, but rather adjusting some actions that are called to ensure integrity of things like the file store, archive and zip builder that happen behind the scenes.