You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For the 2.0 we expected to do a lot of breaking changes (especially the new model), but due the sheer complexity to keep the retrocompatibility, I think that a more step by step approach should be considered.
The first step - worthy of a 2.0 release - consist in dropping mysql and hsqldb. With the availability of postgresql in basically every big cloud provider, the necessity to support others database is clearly reduced. This has already been done in the v2 branch. [PR done]
POSTPONED: Additionally the pgsql switch, we can use the skip locked functionality where possible: this is the first step for removing the quartz scheduler and will improve the contention.
Switch to spring boot 2 and spring 5: to keep up with spring and enjoy the new features (already done in the v2 branch). [PR done]
POSTPONED: Integrate the new payment api: some work has already been done, need to be completed though.
Update source level compatibility to JDK 11. The final v2.0 will most likely target JDK12
use the skip locked functionality where possible: this is the first step for removing the quartz scheduler and will improve the contention. -> WIP in task Explore how to remove quartz scheduler #361
try to make true the "waiting queue" centric design
a better way to load fonts for the pdf generation (should we add other fonts in the docker image?)
email handling: instead of processing them in bulk, use select for update skip locked for a single row + use pgsql notify on creation of the mail, the first instance that will be able to select will send the email + job for the remaining one in case of failures
For the 2.0 we expected to do a lot of breaking changes (especially the new model), but due the sheer complexity to keep the retrocompatibility, I think that a more step by step approach should be considered.
A more realistic roadmap
2.0M0 DONE!
skip locked
functionality where possible: this is the first step for removing the quartz scheduler and will improve the contention.2.0M1 DONE!
skip locked
functionality where possible: this is the first step for removing the quartz scheduler and will improve the contention. -> WIP in task Explore how to remove quartz scheduler #3612.0M2 DONE!
Public side work. The goal is to reduce the maintenance burden of the current frontend and expose new venues for integration.
2.0M3 DONE!
2.0M4
2.0M5
2.0M6
Work originally planned for M3, postponed due to "force majeure"
Admin side work. The goal is to finally cleanup the admin side.
It may be possible to do a gradual conversion.
Interesting tasks that could be done in the 2.0-2.3 timeframe:
The text was updated successfully, but these errors were encountered: