Re: [PATCH 1/2] crypt32: Import system root certs into volatile registry keys instead of dedicated root store.

[Jacek Caban]

It was me sending changed Donat's patches with messed From header. Sorry
about that. Same for 2/2 patch.

Jacek

[PATCH 1/2] crypt32: Import system root certs into volatile registry keys instead of dedicated root store.

[Donat Enikeev]

Hi Jacek,

Thank you for the adjustments.

There are probably too much questions for the one patch, so only if you 
have another free minute, I would be really grateful if you clarify a 
background under removal of `root_certs_imported` flag, intended to 
skip import if it's supposed to be done already. I do understand that 
this flag will not prevent re-importing by another process, but within 
one given - expected to, as semaphore will not exist by a moment of 
next request for HKLM\Root store, while certificates are all in 
registry already. Without the flag, it will re-import all system 
certificates each time an application is calling for HKLM\Root store. 
It's not a big deal of course, just extra electricity burning by 
certification management applications, using all stores heavily (like 
`certmgr`). On the other hand, I could miss something substantial 
again, if so - just ignore the above, I would hope to figure it out 
eventually

Best,
Donnie

On Пн, ноя 7, 2016 at 7:43 , Jacek Caban <jacek(a)codeweavers.com> 
wrote:
> It was me sending changed Donat's patches with messed From header. 
> Sorry
> about that. Same for 2/2 patch.
>
> Jacek

[PATCH 1/2] crypt32: Import system root certs into volatile registry keys instead of dedicated root store.

[Jacek Caban]

Hi Donnie,

The flag is still there, it just skips the whole function.

Jacek

On 07.11.2016 18:23, Donat Enikeev wrote:
> Hi Jacek,
>
> Thank you for the adjustments.
>
> There are probably too much questions for the one patch, so only if
> you have another free minute, I would be really grateful if you
> clarify a background under removal of `root_certs_imported` flag,
> intended to skip import if it's supposed to be done already. I do
> understand that this flag will not prevent re-importing by another
> process, but within one given - expected to, as semaphore will not
> exist by a moment of next request for HKLM\Root store, while
> certificates are all in registry already. Without the flag, it will
> re-import all system certificates each time an application is calling
> for HKLM\Root store. It's not a big deal of course, just extra
> electricity burning by certification management applications, using
> all stores heavily (like `certmgr`). On the other hand, I could miss
> something substantial again, if so - just ignore the above, I would
> hope to figure it out eventually
>
> Best,
> Donnie
>
> On Пн, ноя 7, 2016 at 7:43 , Jacek Caban <jacek(a)codeweavers.com> wrote:
> > It was me sending changed Donat's patches with messed From header. Sorry
> > about that. Same for 2/2 patch.
> >
> > Jacek

[PATCH 1/2] crypt32: Import system root certs into volatile registry keys instead of dedicated root store.

[Donat Enikeev]

> The flag is still there, it just skips the whole function.

Argh, overlooked. Indeed, better in your way. Thank you!

Donnie

On Пн, ноя 7, 2016 at 9:30 , Jacek Caban <jacek(a)codeweavers.com> 
wrote:
> Hi Donnie,
>
> The flag is still there, it just skips the whole function.
>
> Jacek
>
> On 07.11.2016 18:23, Donat Enikeev wrote:
> > Hi Jacek,
> >
> > Thank you for the adjustments.
> >
> > There are probably too much questions for the one patch, so only if
> >  you have another free minute, I would be really grateful if you
> >  clarify a background under removal of `root_certs_imported` flag,
> >  intended to skip import if it's supposed to be done already. I do
> >  understand that this flag will not prevent re-importing by another
> >  process, but within one given - expected to, as semaphore will not
> >  exist by a moment of next request for HKLM\Root store, while
> >  certificates are all in registry already. Without the flag, it will
> >  re-import all system certificates each time an application is 
> > calling
> >  for HKLM\Root store. It's not a big deal of course, just extra
> >  electricity burning by certification management applications, using
> >  all stores heavily (like `certmgr`). On the other hand, I could miss
> >  something substantial again, if so - just ignore the above, I would
> >  hope to figure it out eventually
> >
> > Best,
> >  Donnie
> >
> > On Пн, ноя 7, 2016 at 7:43 , Jacek Caban 
> > <jacek(a)codeweavers.com> wrote:
> > > It was me sending changed Donat's patches with messed From header. 
> > > Sorry
> > >  about that. Same for 2/2 patch.
> > >
> > > Jacek