Bug#252148: zenity is S L O W

[giacomo boffi]

Ross Burton writes:

 > On Tue, 2004-06-01 at 17:40, giacomo boffi wrote:

hi Ross, thank you for your prompt reply

i'm sorry for the repeated bug report, my wrong.

 > > please look at the attached little benchmark, you can see that
 > > instantiating a zenity dialog takes MORE than 10 times the cpu
 > > needed to load xemacs, execute its startup files, to instantiate
 > > a fully decorated (menus, toolbars, pixmaps) frame and to close
 > > it, and what's more important, it takes more than FIVE seconds of
 > > wall clock
 > 
 > For me that code results in the Zenity dialog appearing instantly.

 i never intended to say that zenity doesn't do very well its job

 i just said that, under particular circumstances that i'm willing to
 help investigate, zenity takes an unorderly amount of time to do its
 job

 > Note that the duration that time will return includes the amount of
 > time you take to actually press the dialog, so does not represent a
 > sensible benchmark.

 Ross, please look at the user+sys time in my tests! that's not me
 delaying the button press, it's zenity doing its mumblemumble using
 more than five secs of CPU *before* showing a window, while, e.g.,
 xemacs uses less than 0.5s to initialize itself, open a full featured
 frame and exit. to quote yourself: "something weird going on" 

 > It's probably a good idea to run strace zenity... 2>log and attach
 > the log, so we can see if there is something weird going on with
 > your system.

 uh, sorry, but strace'ing will be delayed until thursday morning
 CEST, as tomorrow it's a holiday here in italy, and i cannot have
 physical access to the system in my office

 i will send the log asap

 good night,
                                                                gb