#398 closed enhancement (wontfix)
View Source: search support
| Reported by: | sascha_silbe | Owned by: | timClicks |
|---|---|---|---|
| Priority: | Normal | Milestone: | |
| Component: | Sugar | Version: | Git as of bugdate |
| Severity: | Minor | Keywords: | sugar-love |
| Cc: | smparrish | Distribution/OS: | Unspecified |
| Bug Status: | New |
Description
It would be very useful to be able to search inside the "View Source" window.
Change History (12)
comment:1 Changed 14 years ago by tomeu
- Bug Status changed from Unconfimed to New
- Milestone changed from Unspecified by Release Team to 0.86
- Priority changed from Unspecified by Maintainer to Normal
- Severity changed from Unspecified to Minor
comment:2 Changed 14 years ago by homunq
- Keywords sugar-love added
comment:3 Changed 14 years ago by erikos
- Milestone changed from 0.86 to 0.88
- Owner changed from marcopg to tomeu
- Status changed from new to assigned
comment:4 Changed 13 years ago by timClicks
- Owner changed from tomeu to timClicks
- Status changed from assigned to accepted
comment:5 Changed 13 years ago by tomeu
- Milestone changed from 0.88 to 0.90
comment:6 follow-up: ↓ 7 Changed 13 years ago by timclicks
comment:7 in reply to: ↑ 6 Changed 13 years ago by timclicks
Replying to timclicks:
- Problems
- Non-greedy search. Stops after encountering the first match.
This isn't actually correct. There function stops at an optional limit. So this approach looks fine to use.
<silbe> timClicks: can you time-bound gtksourceview2.iter_forward_search(), e.g. by letting it search only up to a maximum distance / certain point in the text?
<timClicks> yes
<timClicks> it has a limit argument
<timClicks> i'll check, but I'm fairly certain it's byte-bound, not time-bound
<silbe> timClicks: then my first suggestion would be to run it (if necessary repeatedly) in the "idle" loop.
<silbe> timClicks: doing it this way evades almost all of the pitfalls of multithreading / multiprocessing
<timClicks> i see
<timClicks> does that loop only do something if nothing else requests cpu time?
- timClicks thinks of boinc/seti@home as an analogy
<silbe> timClicks: if you like, you can measure the time the call took and tweak the number of bytes, or call it repeatedly until the time limit is reached. The only important thing is to bound the maximum time it takes well enough to make the GUI feel responsive.
<timClicks> i see
<timClicks> i wanted the gui to be such that you didn't need to hit Ctrl+F
<timClicks> e.g. if you have view source open, you start typing
<timClicks> once you push esc, the string clears
<silbe> timClicks: it's called whenever the GUI main loop is "idle" (i.e. not processing user actions or invoking callbacks); it doesn't depend on the behaviour of other processes or threads.
<silbe> timClicks: ah, so iterative search. very nice.
<silbe> timClicks: oh, and please change to cursor (tomeu posted a link on sugar-devel some days ago) while it's searching / doing highlighting in the background so the user knows the input was accepted
comment:8 follow-up: ↓ 10 Changed 12 years ago by smparrish
- Cc smparrish added
timclicks can you provide an update as to where you are with this.
comment:9 Changed 12 years ago by smparrish
- seeta_dev set to Shanjit, Dipankar
comment:10 in reply to: ↑ 8 Changed 12 years ago by timclicks
Replying to smparrish:
timclicks can you provide an update as to where you are with this.
smparrish I haven't updated anything since the last comment. Feel free to take over.
comment:11 Changed 10 years ago by dnarvaez
- Resolution set to wontfix
- Status changed from accepted to closed
Patches welcome but this doesn't seem to be high priority and it has not been implemented for 3 years.
Update - Implementation notes: