xmonad and launcher prompt memory

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

xmonad and launcher prompt memory

Carsten Mattner-2
Carlos,

I'm wondering whether using the launcher prompt may be the reason
for xmonad starting at 4 to 5 megs freshly and climbing to around 7 megs.
I didn't try without the launcher promot and don't really complain but
am curious whether there's some caching and possible space leak
in either the launcher prompt or xmonad core or one of the popular
action hooks.

Gwern, this is not an objection to merging, just an observation.

This is on ghc-7.6.1-i386.

Any comments?


Reply | Threaded
Open this post in threaded view
|

xmonad and launcher prompt memory

Carlos López Camey
Hello Carsten,

After I started the launcher prompt and did some queries, the resident
memory shown in `top` went from 5620 bytes in a restarted xmonad to
6440 bytes. After doing more queries it stopped at 7280 bytes.

I observed the same behavior with the shell prompt (Prompt.Shell), but
I am not sure if it is due to the new patch. Could someone that
doesn't have the patch applied that a look?

If the new patch is the reason, the guilty data type might be
XPOperationMode in XPState:

> data XPOperationMode = XPSingleMode ComplFunction XPType | XPMultipleModes (W.Stack XPType)

where

> data XPType = forall p . XPrompt p => XPT p
> type ComplFunction = String -> IO [String]

I am not sure what should I do to get the XPState garbage collected
when a window prompt is destroyed. Help?

cheers

2012/9/19 Carsten Mattner <carstenmattner at gmail.com>:

> Carlos,
>
> I'm wondering whether using the launcher prompt may be the reason
> for xmonad starting at 4 to 5 megs freshly and climbing to around 7 megs.
> I didn't try without the launcher promot and don't really complain but
> am curious whether there's some caching and possible space leak
> in either the launcher prompt or xmonad core or one of the popular
> action hooks.
>
> Gwern, this is not an objection to merging, just an observation.
>
> This is on ghc-7.6.1-i386.
>
> Any comments?