three days mostly lost!

[juan_gandhi]

Yes, I am a fan of IntelliJ. Still. But last week a "method breakpoint" somehow penetrated my code. How the fuck did I know it makes the code 30 times slower? I spent interesting time looking for the causes of possible network delays; I found stupid bugs in java.text classes (made by one of my colleagues); I bugged unsuccessfully my colleagues.

Until I discovered in a mail thread that METHOD BREAKPOINTS DRASTICALLY SLOW INTELLIJ.

Drastically is the word. Imaging you have bugs to fix, and you cannot even run the server, for 3 days in a row.

Comments

[xryun-porosyan.livejournal.com]

Offtopic: а почему Google search engine в Firefox когда я печатаю по русски "по" в качестве первого варианта сразу же предлагает мне "по хую"?

[deni-ok.livejournal.com]

У меня "погода".
Видимо такова личная история поисков ;)
Персонификация, однако...

[xryun-porosyan.livejournal.com]

Пааапрааашу!
Я хуи не искал :)

[ivan-gandhi.livejournal.com]

Это у тебя по хую; у меня - по принуждению. (Стоит задуматься, однако.)

Просто наш Фил видит тебя насквозь. Рентген.

[xryun-porosyan.livejournal.com]

Я сейчас с работы попробовал, и тоже пишет "по принуждению" А из дома: "по хую"


ГЕНИАЛЬНО!!!!

[ivan-gandhi.livejournal.com]

See here (http://ivan-ghandhi.livejournal.com/579283.html).

[nine_k]

That's why I strictly prefer logging in such cases.

Debugging is only good when you have a rather small piece of code that behaves strangely right before your eyes, or for getting stack trace / program state from crash dumps.

IMHO, of course.

[ivan-gandhi.livejournal.com]

Sure... and run it from command line.

In my case, I spent two days figuring out how to run it from cl.

In general, there's a better way: unittesting. But in our team we are just starting having really using it. Unfortunately.

[nicity.livejournal.com]

особенность использования JDI,
на входе и выходе в каждый метод проверяется, то ли это метод, на который стоит брекпоинт