Vincent van Ravesteijn wrote:
>
>> have you checked its not dataloss from multiple windows view as i have
>> warned before?
>> if you have two windows, which have opened the same file and then try to
>> hide buffer in one of
>> the windows, you will ask for save. this imho shouldn't happen, secondly
>> if hit don't save i get reload and lose my changes of this document in the
>> second window?
>>
> Yes, you shouldn't get the askForSave thingie,.. (but there's much more
> wrong with this kind of thing. It was also changed a lot in r27525 with the
> very interesting comment that it's the second patch from Tomasso, .. Abdel
> #(*#*($ ?
>> i haven't followed the whole thread but this feel wrong to me.
> Could you be more specific.
>> it looks like crippling hide feature
> Could you be more specific.
the advantage of having more windows in one instance is sharing opened
documents and its convenient to have visible only certain dynamic subset of
shared documents. for this hide button is a good thing.
now you made impossible to hide buffer from the second window without
saving buffer which is what 'feels wrong'.
to ask for discarding changes opened in another window would lead imho
to more dataloss problems the the issue you are trying to solve.
>> and introducing new bugs
> Could you be more specific.
>> (i just got crash when trying the scenario outlined above)
> Could you be more specific.
DocIterator.cpp(95): Should not happen, but it does.
Paragraph.cpp(1363): pos: 6 size: 3
lassert.cpp(21): ASSERTION pos <= size() VIOLATED IN Paragraph.cpp:1364
Assertion triggered in void lyx::doAssert(const char*, const char*, long int)
by failing check "false" in file lassert.cpp:23
to sum up:
1. new file1,new file2
2. write some text in file 2
3. new window,open both files 1,2
4. try to hide file 2
-> impossible without saving (cripling #1)
-> if file has its name introducing danger of dataloss
(lyx asks for something when hiding and i unfortunatelly
press discard expecting hiding wont affect window 1 as it
was with the current workflow.) (cripling #2)
5. save file 2
6. type something in file 2
7. hide file 2, discard
8. kaboom when focus in window 1 ("introducing new bugs")
> We agreed on the list that we better wanted to make sure that hidden
> buffers were not dirty.
i would be more happy if the condition instead of "hidden buffers were not
dirty" was "buffers hidden _in all windows_ were not dirty."
>If you didn't like that you had enough
> opportunities to step up.
thats unfair. i was already rasing the multiple window issue and warned
the the fixes should consider this.
pavel
