Control: retitle -1 hnb: please make sync() operations optional
Control: tag -1 - patch
Hi Xavier,
xavier renaut wrote:
> * What led up to the situation?
>
> working on heavy loaded system (in particular IO)
>
>
> * What exactly did you do (or not do) that was effective (or
> ineffective)?
>
> if you press enter after editing a note, it's doing a "sync"
>
> * What was the outcome of this action?
>
> super slow to be able to work again, after each enter it ask to sync
>
> * What outcome did you expect instead?
>
> immediate return and being able to work
Try to use eatmydata, e.g. call "eatmydata hnb".
> can we make this a config parameter ? I believe it's a very rare
> for an app to call sync after each small edit.
Depends. It's quite common in e.g. dpkg or other citical applications.
> it's supposing the computer will reboot without flushing it's cache
> any moment now.
Yeah, hence it IMHO makes sense that it calls sync after an auto save
operation.
> which is not usually the case.
Sure it does not happen that often (mostly laptops running out of
power), but if it happens, you're glad about it.
> only mysql or other db's have such a "data integrity" feature.
No, it's way more common than you think. IIRC even some web browsers
do that for history and stuff, e.g. Firefox uses sqlite which again is
a database, hence it works like a database.
Hardcoding autosave_sync to 0 is IMHO no proper solution, hence
removing the patch tag. But implementing it as option seems to be a
nice feature even though there is eatmydata which kills all these sync
operations via LD_PRELOAD.
Regards, Axel
--
,''`. | Axel Beckert <a...@debian.org>, http://people.debian.org/~abe/
: :' : | Debian Developer, ftp.ch.debian.org Admin
`. `' | 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
`- | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
