Phil Stracchino writes:
On 10/4/20 4:09 PM, Loïc Dachary wrote:
Hi,
It turns out to be a feature added in 1.3.1 [0]. I wonder if the MySQL instructions listed in this comment[1] are safe to use? I'm reluctant to upgrade because the machine is for production and using packages from Debian GNU/Linux buster is recommended.
Cheers
[0] https://gitlab.com/mailman/hyperkitty/-/issues/146 [1] https://gitlab.com/mailman/hyperkitty/-/issues/146#note_423111350
To be honest, they don't look like very well written queries.
They're not elegantly written, but they look correct to me (they're *supposed* to be a bulldozer :) and they worked for that user. If one doesn't have 1.3.1 and is unwilling to upgrade, it should be safe enough.
- For extra safety you can take a snapshot of the database.
- Make absolutely sure you have the right mailinglist_id.
And that procedure should work.
@Loïc Oerhaps you should review your policy on upgrading with respect to this particular application (choose Mailman suite or HyperKitty, depending). We're not seeing many regressions, but over time we are fixing inconveniences where documented features work as documented but are insufficient to get your work done, so you have to drop into "hack the Mailman command shell" or even "hack the backend RDBMS" mode. Ie, we're adding features that allow you to do the work with code we've written and tested, rather than some third party script.
This is something only you can decide, of course, and I'm not going to advocate one way or the other. If you're pretty comfortable with direct access to your MySQL database, that way lowers risk of regressions by not upgrading the software. If you're not, upgrading lowers the risk you face with lightly tested third party procedures.
Steve