From: Jordi 'I. M. <jmi...@or...> - 2001-01-10 10:15:51
|
I have not been able to use CVS since this morning. SSH was giving a warning about some hacker doing a man-in-the-middle attack. This morning I spoke to my admin and found out that he had changed the network's configuration so now there's a Linux machine rounting all traffic. ( The man in-the-middle ). sigh.... Ok, on to the point. As usual when I wait too much before updating my API's version when I do I don't recognize the code. ( Even mine ). One good thing about being a web developer is that I can spend the whole day debugging the API and still I'm doing my job :) I'm trying to fix the label and button widgets in the hope that this will lead to fix many other bugs. Now, tracking down one of these pestering event-not-happening bugs I've reached this place inside event.js if (is.ns4 && is.platform=="other" && type=="mouseup") this.invokeEvent("click",e); I'm reasonabily sure that this -is.platform- condition should not be there. Coded that way, click events only fire when you're running a Linux/Unix Netscape, don't they ? By removing the condition, the button widget worked fine in NS( did you notice that clicking the text in the button was causing the button to animate but the resizing did not happen ? ). IE, as Pascal already stated, has another problem. I don't want to directly update CVS with this change because I have a terrible feeling that this code was placed there because some platform issue and I may have broken something else. Would the person that coded that condition please tell me if it is a mistake or not ? Another issue. Same function, several lines before: if (is.ns4 && is.platform=="other") { if (type=="mousedown") { if (this.dbltimer!=null) type="dblclick"; else this.dbltimer=setTimeout(this+'.dbltimer=null',300); } } Again it seems that Win NS won't fire any doubleclick events. Since this code has been there since Dan's first DynAPI2 release and whenever I remove it, it reappears again, my best guesses are: - 1: The condition should be platform!='other'. I know that win NS handles doubleclick events correctly ( I've tryed removing the condition and did well ). Then maybe some Linux/Unix doesn't but the condition is miswritten. - 2: There's something that I have missed and dblclick events under Windows are handled in another way. - 3: I'm dumb I'll continue to investigate ( I will dedicate the rest of the day to debug this release ) In the meantime it would be useful if you all posted all the bugs you can think of so I don't forget something. ---- I like OOP as much as anyone and I do recognize the benefits of modularity, classes, inheritance, etc, etc. However, we must keep in mind that sometimes the ground we're steping is not solid enought to build a skyscraper on top of it. I understand that programming Delphy and such is easier, and I do understant that JAVA's object model is cool and scalable and you can reuse code but, how many years did it took until technology was mature enought to support things as abstract and platform-independant as JAVA ( I lived the old ASM days, one mistake-one reboot, registers, pointers.... ) ? Concepts such as code size still count. You may have a fast connection but still a website 10k smaller is a better website. Lots more will be said. |