Hi Timo

On Mon, Oct 20, 2025 at 06:41:17PM +0200, Timo Rothenpieler via ffmpeg-devel wrote:
> On 20.10.2025 18:14, Michael Niedermayer via ffmpeg-devel wrote:
> > Hi everyone
> > 
> > We currently have 2 issue trackers (trac and forgejo),
> > 
> > Previous discussion:
> > "[RFC] Issue tracker" and "trac ticket statistics"
> > https://lists.ffmpeg.org/archives/list/ffmpeg-devel@ffmpeg.org/thread/HS46O6K62MHRHJDJC45XLRMR64RE5IB7/#LMXCTJDF5SIUSX43X7YZ4347U5YHWW33
> > 
> > I think there are 2 questions, the first is,
> > which issue tracker do we want to use going forward?
> > And the 2nd question is, about migrating tickets.
> > 
> > First lets establish which tracker we want to use going forward.
> > It seems there where 2 options suggested.
> > 
> > "T", Trac, with a potential future switch to redmine (similar to trac)
> > 
> > "F", Forgejo
> 
> 
> F for Forgejo, simply because maintaining two applications, specially one as
> old and "brittle" as trac is a pain.
> 
> trac currently needs subtantially more server resources than Forgejo, while
> having roughly similar amounts of requests.

> It also frequently goes down if the traffic goes too hard, and needs a
> reboot to come back.

trac never needed a reboot previously, not that i remember.
I do remember when it used sqlite and under load it sometimes showed
database locked errors but it always recovered from this with no action from the admin
at all. just waiting, and i think these errors stoped after tuning sqlite a bit

what exactly did get stuck ?
The current setup is different from what it was


> 
> I also personally feel like we don't really need/use any of the more
> advanced features in trac. At least I have never looked at votes before, or
> heard anyone talk about them before Forgejo came up.

I was just in the recent days using the upvotes as a means to find the
most popular bugs and feature requests for the sponsoring RFC

Sponsoring or not, I want to know which bugs and feature requests affect
our users most.
This does matter. Has nothing to do with sponsoring.
Plain and simple, if i have time to work on something, i would want to
spend it on something that has impact and "sort by vote count" is the simple awnser.


> The only clear advantage it has is the much more refined search, but that's
> something other people have noticed to be lacking in Forgejo as well, so
> there is ongoing effort to improve it.

There also was ongoing work on libavcodec to have a generic "MPEG" encoder
that would handle h264 and later standards ... very long ago


> 
> 
> Also, redmine is not something one can just switch to.
> It looks similar, but is an entirely seaprate thing.
> There is also no clear way to migrate to it, so it'd be another disruptive
> switch.
> There is https://www.redmine.org/projects/redmine/wiki/redminemigrate
> but it explicitly says the migration from Trac is obsolete and does not
> work.

IIUC the convert script needs a patch and or a older intermediate redmine

They removed trac import support as everyone who needed it no longer needs it

you can look at for example
https://www.redmine.org/issues/33722

In a comment:
"I already migrated from Trac to Gitlab via Redmine by using this patch."

So the convert worked in the past.

thx

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

When you are offended at any man's fault, turn to yourself and study your
own failings. Then you will forget your anger. -- Epictetus