Improve handling of http parameters and arguments for gmp #1355
Conversation
|
Aside from the things I've noted above I've also noticed that the "Rows Per Page" setting is now ignored. |
Yes the web server should be only a proxy between GSA and gvmd and therefore all request preparation and params should be done in the js code now. |
Abstract adding arguments into a c structure and functions.
Add gsad_gmp_arguments.c and gsad_gmp_request.c as depdendencies for gsad.
* Extract a get_entities function from get_many and handle only default arguments in get_many additionally. * Update get_tasks use get_many and get_reports_gmp to use get_entities directly. * Use gmp_arguments_t to avoid char hacking for filters * Drop extra get_reports function * Drop special handlings from get_many
* Update get_assets to use gmp_arguments * Require asset_type param
Don't request tasks and filters in get_alerts.
Don't remove filter params before its values are copied. The char data is feed when params_remove is called. Therefore the freed memory has been accessed here.
Don't add the special handling of the plural type to get_many. Instead pass the correct command types to the function.
This function isn't used in GSA.
By removing the support for the extra_xml param at get_many and get_entities the extra functions for each get commands have become obsolete too.
Don't request report formats, tasks and filters in get_alert. This has to be and is done in GSA directly
| } | ||
|
|
||
| void | ||
| gmp_arguments_free (gmp_arguments_t *arguments) |
There was a problem hiding this comment.
The _free in the function name makes it sound like it will immediately free the memory of the data structure, so maybe _unref would be better.
I don't want to hold this back any longer, so I'd be fine if you changed this in another PR or we discussed this later.
There was a problem hiding this comment.
Oh I thinks this is a leftover. I couldn't agree on using referencing or freeing with myself ;-) I'll change it in a future PR. Thanks a lot!
Checklist: