Обсуждение: Test results for 1.6RC2 rev: 5636M

Поиск
Список
Период
Сортировка

Test results for 1.6RC2 rev: 5636M

От
Erwin Brandstetter
Дата:
Hi developers! Hi Dave!

Testing 1.6RC2 rev: 5636M, client Win XP.

# One more cornercase for "Find and Replace". Try to replace multiple 
occurrences of  'abc' with 'abcd' and see for yourself ..
Seems like "Replace" continues the search at the _beginning_ of the 
newly inserted text. If the search string is included in the replace 
string, we end up in an endless loop.
Might be easy to fix by continuing the search _after_ the replaced 
string. ("Find" seems to do it right.)


dpage@postgresql.org wrote:
> Figured out a way to check for modification more sensibly, so we can 
> now disable the buttons on column change.

That's very good news! Not only have you fixed the issue at hand, you 
also got rid of other unnecessary save dialogues (we have been 
discussing before). The latter is even the greater improvement, IMO.
I have my doubts you want to hear of this, but I have still found an 
(even less important) cornercase. :)
If I stay _on_ the not edited cell and just press <esc> to leave edit 
mode, I can still trigger the previously discussed issues.


>> That takes care of it, mostly.
>> Unlike with text fields the new quoting in numeric fields does not 
>> escape quotes in the pasted data.
>> If there is no specific reason for the deviation, you might consider 
>> adapting the behaviour for text fields to prevent weird effects.
>
> Will be escaped properly now.

That seems to work just fine now.


One other thing I noticed on my tour:
In the help pages, "The pgAdmin Development Team" dates back to 
"*Version 1.3, 15 July 2003". Maybe time for an update?

*Other than that, I have found no new oddities that might have been 
introduced by the latest fixes. I have revisited most of the issues I 
was involved in over the last few weeks and everthing was as expected, 
i.e. mostly fixed or mostly harmless. I couldn't even reproduce the 
"pane-shuffling"-crash. Wasn't aware that this one had been fixed, too.

This is very likely the best pgAdmin I have ever been working with. :)


Regards
Erwin


Re: Test results for 1.6RC2 rev: 5636M

От
Dave Page
Дата:
Erwin Brandstetter wrote:
> Hi developers! Hi Dave!
> 
> Testing 1.6RC2 rev: 5636M, client Win XP.
> 
> # One more cornercase for "Find and Replace". Try to replace multiple 
> occurrences of  'abc' with 'abcd' and see for yourself ..
> Seems like "Replace" continues the search at the _beginning_ of the 
> newly inserted text. If the search string is included in the replace 
> string, we end up in an endless loop.
> Might be easy to fix by continuing the search _after_ the replaced 
> string. ("Find" seems to do it right.)

OK, got that one.

> dpage@postgresql.org wrote:
>> Figured out a way to check for modification more sensibly, so we can 
>> now disable the buttons on column change.
> 
> That's very good news! Not only have you fixed the issue at hand, you 
> also got rid of other unnecessary save dialogues (we have been 
> discussing before). The latter is even the greater improvement, IMO.
> I have my doubts you want to hear of this, but I have still found an 
> (even less important) cornercase. :)
> If I stay _on_ the not edited cell and just press <esc> to leave edit 
> mode, I can still trigger the previously discussed issues.

Yeah, that ones a bit more tricky to solve :-(

Thanks, Dave


Re: Test results for 1.6RC2 rev: 5636M

От
Erwin Brandstetter
Дата:
Di Dave!

Testing 1.6.0 rev: 5641

dpage@postgresql.org wrote:
> Erwin Brandstetter wrote:
>> # One more cornercase for "Find and Replace". Try to replace multiple 
>> occurrences of  'abc' with 'abcd' and see for yourself ..
>> Seems like "Replace" continues the search at the _beginning_ of the 
>> newly inserted text. If the search string is included in the replace 
>> string, we end up in an endless loop.
>> Might be easy to fix by continuing the search _after_ the replaced 
>> string. ("Find" seems to do it right.)
>
> OK, got that one.

Almost. :) You nailed the case abc -> abcd. But you missed the case abc 
-> dabc.
Looks like you continue to search one character after beginn of replaced 
text. The proper solution would be to continue the search at the end of 
the inserted text (= beginn of the rest of the text).


Regards
Erwin


Re: Test results for 1.6RC2 rev: 5636M

От
Dave Page
Дата:
Erwin Brandstetter wrote:
> Di Dave!
> 
> Testing 1.6.0 rev: 5641
> 
> dpage@postgresql.org wrote:
>> Erwin Brandstetter wrote:
>>> # One more cornercase for "Find and Replace". Try to replace multiple 
>>> occurrences of  'abc' with 'abcd' and see for yourself ..
>>> Seems like "Replace" continues the search at the _beginning_ of the 
>>> newly inserted text. If the search string is included in the replace 
>>> string, we end up in an endless loop.
>>> Might be easy to fix by continuing the search _after_ the replaced 
>>> string. ("Find" seems to do it right.)
>>
>> OK, got that one.
> 
> Almost. :) You nailed the case abc -> abcd. But you missed the case abc 
> -> dabc.
> Looks like you continue to search one character after beginn of replaced 
> text. The proper solution would be to continue the search at the end of 
> the inserted text (= beginn of the rest of the text).

That code was somewahat complex to say the least, so I've rewritten it 
in a far more simple way. From an end user point of view, the only 
noticable difference should be that at 'From the Top' search is now 
*always* from the top, however, it will automatically switch back to 
'From the Cursor' following a successful 'From the Top' search, so 
repeatedly pressing Find/Replace will do what you expect.

I'll send you a .exe to test - feedback appreciated as always!

Thanks, Dave