From kfm-devel Mon Nov 06 11:06:01 2006 From: Thiago Macieira Date: Mon, 06 Nov 2006 11:06:01 +0000 To: kfm-devel Subject: Re: Suggestion for Konquerer developers and delivering "stripped" Message-Id: <200611061206.09292.thiago () kde ! org> X-MARC-Message: https://marc.info/?l=kfm-devel&m=116281134912281 MIME-Version: 1 Content-Type: multipart/mixed; boundary="--nextPart1376955.YVDg9sUM5S" --nextPart1376955.YVDg9sUM5S Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline OnThe Road wrote: >Guys, >I don't know what you are thinking, but if you have >any hope at all of going mainstream, you need to come >up with an alternate strategy for getting debug info. And we don't know what you're thinking, but you're way off the mark.=20 Either way, thanks for your comments; let's address your remarks. Setting things straight first: KDE does not ship binaries. At all. Any=20 binaries. KDE ships source code. Distributions ship binaries. So, if you have a qualm or suggestion about=20 how to handle debug builds or debug info, it's to your distribution that=20 you should be talking. It's they who make the builds and it's they who=20 decide how/if to package the debug info. By the way, one recent trend in distribution policy is to generate=20 separate debug packages. If a user is interested in submitting a bug=20 report, turning his application into one with proper backtrace is a few=20 mouse-clicks away, by installing a new package (which, by the way,=20 doesn't impact at all run time!). >To get Konqueuer to crash, using (XUbuntu 6.06), just >go to cnn.com, select the business button. That's it! >Gone...every,.. single,.. time... No crash here. I got redirected to http://money.cnn.com/index.html?cnn=3Dye= s=20 and the page loaded just fine. KDE 3.5.5 r594555 >Expecting a "main stream" (i.e. non-geek) user to get >source, rebuild with debug options, install, just so >they can send a backtrace is completely unrealistic. Agreed. That would not be a realistic expectation. >Wake up!.. For stripped releases, you should >incorporate a "log option" that turns on the >generation of log files (yes even in a released >version, where the debug symbols have been stripped). A log option that would log what? The websites being visited? >If you are unreasonable and unwilling to do this, at >least make a debug binary build available, that's the >least I would expect. That's not our job to do, but, like I said before, it's already available=20 from the major distributions. >BTW, I have been developing C/C++ for over 20 years >and I must say that software should never crash.=20 Agreed, but it happens. Or you're going to tell me that in those 20 years,= =20 you have never written software that crashed, at least once? >It=20 >might not work correctly, but when something just >crashes, it indicates poor architecture, framework and >especially error handling (or lack there of). Have you >ever heard of exception handling? If not, I would >suggest that you read up on it (and think about >implementing it). Exception handling isn't the solution for all problems under the Sun. I=20 suggest you read up on our architecture too before making wild=20 suggestions. >I realize that no one is getting paid for the open >source stuff, however, that is no excuse for shitty >software. Ok, now you can go away. If you want me to continue addressing your=20 remarks, rewrite your email with a cool head and without swearing. If you consider our software "shitty", then we can't address your=20 concerns. You have already made up your mind about our quality and=20 nothing we do will be good enough. I suggest you try something else that=20 suits your needs instead. If you've just written this out of anger because you lost data or=20 something recent, then we can talk when you're no longer angry. [snip the rest of the message] =2D-=20 =A0 Thiago Macieira =A0- =A0thiago (AT) macieira.info - thiago (AT) kde.org =A0 =A0 PGP/GPG: 0x6EF45358; fingerprint: =A0 =A0 E067 918B B660 DBD1 105C =A0966C 33F5 F005 6EF4 5358 --nextPart1376955.YVDg9sUM5S Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQBFTxchM/XwBW70U1gRAsjfAJ4/Ch+qHY+C3DXY4JSYRuORE1ndIACeIUhT dL9Alp0YBLUnrZJ840iardY= =6Vs1 -----END PGP SIGNATURE----- --nextPart1376955.YVDg9sUM5S--