Hi,

1.
Approx. 2 years ago some Guixers proposed to create profiles in order to 
have not too many packages in the main profile, which makes it easier to 
update the profile. If during updating the profile one package creates 
trouble, the whole updating process stops. I had those trouble several 
times.

2.
If I put all packages in the main profile, I would have approx. 30 
packages more and altogether approx. 120 packages at the moment, and 
there will be more in future.
I don´t use "guix home"

3.
May be those Guixers have special needs and it works easier for them.
Of course, those Guixers know what they are doing unlike me.

4.
If I had all packages in my main profile, would it be possible to update 
only some packages?
e.g. icecat and others quite regularly are updated, so it would be good 
to update them more often.

thanks for your help

Gottfried



Am 09.12.24 um 18:15 schrieb Ian Eure:
> Hi gfp,
> 
> gfp <gfp@posteo.at> writes:
> 
>> I have got 17 Profiles in:
>> /home/gfp/Projekte/Calibre
>>           /Emacs
>>           /Libreoffice
>>           etc.
>>
>>  Would that mean to sorce all 17 profiles in one go? or only two  at a 
>> time?
> 
> I agree with Steve: you shouldn’t use profiles like this.  By activating 
> all these profiles at once, you’re effectively creating a single profile 
> with the combined packages of all 17.  But since this is done in an 
> ad-hoc way, outside Guix, it’s missing functionality, such as telling 
> you about conflicts, building the XDG MIME database, font cache, etc.  
> This is very likely going to cause further bugs and problems.
> 
> If you put the software you regularly use into your main profile, the 
> system works better and your problem goes away without needing to mess 
> with shell startup.
> 
> Engineering is all about making tradeoffs.  With the downsides you’ve 
> noted, and the complexity of managing it, I’d expect there to be a very 
> strong and specific upside from such a system which can’t be replicated 
> with a one-profile setup.  I may have missed it, but I don’t see such a 
> rationale here.
> 
> I strongly urge you to reconsider your approach.
> 
>   -- Ian