|
http://cgit.freedesktop.org/xorg/app/sessreg/tree/sessreg.c?id=sessreg-1.0.8#n153 Because pointers and C strings aren't hard enough.
|
# ? Nov 5, 2014 19:56 |
|
|
# ? May 29, 2024 06:10 |
|
How does this thread generally feel about Qt? I sometimes work on a large Qt-based application, and it seems to be pretty coherent overall, but are there some good examples of horror there that I just haven't stumbled upon yet?
|
# ? Nov 5, 2014 19:58 |
|
Qt is great compared to a lot of frameworks out there *cough*gtk*cough*
|
# ? Nov 5, 2014 20:01 |
|
Subjunctive posted:Rule check: how often am I allowed to complain about Perforce in here? If Perforce is the worst VCS you've worked with, you're not allowed to complain about it. (Though IIRC, you're at Google, and they have or used to have a bunch of scripts to monitor the latencies of their Perforce servers and just restart them when they inevitably start running slow and stop scaling. I can see how this is more painful at Google than at smaller places.)
|
# ? Nov 5, 2014 20:03 |
|
Skuto posted:http://cgit.freedesktop.org/xorg/app/sessreg/tree/sessreg.c?id=sessreg-1.0.8#n153 I like the descriptive variable names personally.
|
# ? Nov 5, 2014 20:14 |
|
I don't think I've seen *++* used before.
|
# ? Nov 5, 2014 20:23 |
|
hobbesmaster posted:Qt is great compared to a lot of frameworks out there *cough*gtk*cough* GTK+ owns
|
# ? Nov 5, 2014 20:25 |
|
Skuto posted:http://cgit.freedesktop.org/xorg/app/sessreg/tree/sessreg.c?id=sessreg-1.0.8#n153 http://patchwork.freedesktop.org/patch/36266/ You're a few days late.
|
# ? Nov 5, 2014 20:27 |
|
hobbesmaster posted:Qt is great compared to a lot of frameworks out there *cough*gtk*cough* I'm starting development on BlackBerry 10 later this week. I'm holding you responsible for the optimism you're making me feel
|
# ? Nov 5, 2014 22:38 |
|
Suspicious Dish posted:http://patchwork.freedesktop.org/patch/36266/ Curious how does this thing work - avp is an array of pointers to strings? *++*a = *(++(*a)) - *a pointer moves to the next string?
|
# ? Nov 5, 2014 22:46 |
|
hackbunny posted:I'm starting development on BlackBerry 10 later this week. I'm holding you responsible for the optimism you're making me feel ... out of curiosity, why? I assumed that BB was pretty much dead at this point.
|
# ? Nov 5, 2014 23:34 |
|
Skuto posted:If Perforce is the worst VCS you've worked with, you're not allowed to complain about it. Not the worst for sure. Hard to "beat" CVS, or sccs. (I'm at FB/Oculus, not Google.)
|
# ? Nov 5, 2014 23:47 |
|
Volmarias posted:Please do OK, OK, I give up, here is it. I hope it will not turn out to be a mistake omeg posted:Edit: actual semi serious question to hackbunny or anyone else maybe more knowledgeable than me: how hard would be to make gui processes on non-active desktops still process messages and user/gdi apis to actually work on them? Since pretty much all of win32k state is global most user/gdi apis just return ACCESS_DENIED when targeted to non-active desktops, and threads on those inactive desktops seem to not process any (or most) window messages. Well I guess I answered myself. I think you can't send messages cross-desktop. Have you tried setting the target desktop as the desktop of the sending thread? Or maybe the desktop has the wrong permissions? but I admit it's a stab in the dark Volmarias posted:... out of curiosity, why? I assumed that BB was pretty much dead at this point. A rather important customer demands it. They are in too deep with BES and can't get rid of BlackBerry devices easily
|
# ? Nov 6, 2014 01:02 |
|
hackbunny posted:I think you can't send messages cross-desktop. Have you tried setting the target desktop as the desktop Changing the desktop should work but I specifically want to avoid that. It's a long story -- basically a dream of implementing meaningful GUI isolation within one active session. Like a seamless mode for VMs but without actual VMs. I've done a lot of research and have come to the conclusion it's just not possible with the way win32k is built. On topic of windowing systems, is there any OS with a non-awful architecture of GUI stuff?
|
# ? Nov 6, 2014 01:24 |
|
OS X's is pretty good, considering. I mean, it's better than X11 or Windows. That's not exactly a high bar.
|
# ? Nov 6, 2014 01:25 |
|
omeg posted:Changing the desktop should work but I specifically want to avoid that. It's a long story -- basically a dream of implementing meaningful GUI isolation within one active session. Like a seamless mode for VMs but without actual VMs. I've done a lot of research and have come to the conclusion it's just not possible with the way win32k is built. Wayland is quite good for a modern architecture.
|
# ? Nov 6, 2014 01:48 |
|
Sinestro posted:OS X's is pretty good, considering. I mean, it's better than X11 or Windows. That's not exactly a high bar. We do some weird stuff, but modern Windows (esp Win10) provides a much better system for doing that stuff (scheduling/preemption, multi-process/API sharing, devices over HDMI that don't show up as regular monitors, hooking apps' stuff) than X11 or OS X, at least so far. E: I don't know if Wayland sucks or rules for our purposes, yet.
|
# ? Nov 6, 2014 02:08 |
|
I loving hate winerror.hcode:
|
# ? Nov 6, 2014 02:25 |
|
I think I want to say that I've been at my co-op job for about 1.5 months, and I haven't done a single commit yet. This may be good for me, as this place uses Clearcase.
|
# ? Nov 6, 2014 02:35 |
|
hackbunny posted:I'm starting development on BlackBerry 10 later this week. I'm holding you responsible for the optimism you're making me feel Uh I've only used Qt on "desktop" platforms. I suspect blackberries may be a bit different.
|
# ? Nov 6, 2014 02:42 |
|
QuarkJets posted:How does this thread generally feel about Qt? I sometimes work on a large Qt-based application, and it seems to be pretty coherent overall, but are there some good examples of horror there that I just haven't stumbled upon yet? I've had way more experience with PyQt than Qt, and on the whole, it's relatively straight forward, if not laborious. You have to manage an acylic memory graph by hand of parent and child QObjects, rather than managing ref-counts or gc. Any time you delete an object, all of the objects children are cleaned out. Although there is threads, you live in an event loop, so you also have to mash your code until it's co-operatively scheduled. The real Qt horror might be that it is a framework, not a library for c++. Everything is different, down to the data structures http://doc.qt.digia.com/qq/qq19-containers.html (no stl here mate) and there is the c++ preprocessor moc http://qt-project.org/doc/qt-4.8/moc.html too
|
# ? Nov 6, 2014 03:58 |
|
You have a double called tmpRec.Value. If its negative put it in moneyOut otherwise pop it into MoneyIn. What do you do??????????code:
How does that even pop into your head?
|
# ? Nov 6, 2014 05:44 |
|
bpower posted:You have a double called tmpRec.Value. If its negative put it in moneyOut otherwise pop it into MoneyIn. What do you do?????????? No check of tmpRec.HasValue may be considered worse, if it's not immediately above your first line.
|
# ? Nov 6, 2014 06:28 |
omeg posted:I won't post any links in case it's considered but I highly recommend to browse through the leaked win2k sources. They're old but clearly show the difference in quality between the kernel and the windowing/graphics subsystem. Granted it's almost all in the name of compatibility with this or that. Hard to find a function without "HACK HACK HACK" or "THIS IS SO LOTUS123 CAN loving RUN". My personal favorite is, "TERRIBLE HORRIBLE NO GOOD VERY BAD HACK." Edit: tef posted:I have a nifty avatar How long have you had that avatar, and how did you end up with it? I've been obsessed with L-Systems this past week. I've been posting shots from my current project in the screenshots thread. Centripetal Horse fucked around with this message at 07:28 on Nov 6, 2014 |
|
# ? Nov 6, 2014 07:25 |
|
Subjunctive posted:Hard to "beat" CVS, or sccs. zergstain posted:...I haven't done a single commit yet. This may be good for me, as this place uses Clearcase. f;b
|
# ? Nov 6, 2014 08:14 |
|
Suspicious Dish posted:You're a few days late. Note that the relevant code is still there. quote:Curious how does this thing work - avp is an array of pointers to strings? *++*a = *(++(*a)) - *a pointer moves to the next string? ..and we check it isn't empty by seeing if the first char is 0, yes.
|
# ? Nov 6, 2014 08:19 |
|
Subjunctive posted:Not the worst for sure. Hard to "beat" CVS, or sccs. (I'm at FB/Oculus, not Google.) You obviously havent tried Source(un)Safe (the number of times we had to start again because the DB corrupted) or Clearcase. I don't want to even go there....
|
# ? Nov 6, 2014 10:08 |
|
If you hate clearcase, wait till you've tried clearcase where branching is disabled in all vobs and the use of commit messages is strictly forbidden in favor of a separate central manually-updated commit log that in practice has barely been touched since 1997!
|
# ? Nov 6, 2014 10:25 |
|
Soricidus posted:If you hate clearcase, wait till you've tried clearcase where branching is disabled in all vobs and the use of commit messages is strictly forbidden in favor of a separate central manually-updated commit log that in practice has barely been touched since 1997! sounds like when i used to work at ATOS
|
# ? Nov 6, 2014 10:27 |
|
Yeah, I haven't had the pleasure of using them directly, though I've felt both those bullets graze me.
|
# ? Nov 6, 2014 10:30 |
|
Someone tell me about CMVC because after the current project is done we're moving to a product that is managed by it. All I know is they tried and failed to convert the project from CMVC to our current source control package very recently.
|
# ? Nov 6, 2014 17:12 |
|
zergstain in Dec 2012 posted:Place I'm starting at in about a week uses CVS. I'm not sure how I should feel about that. Progressive JPEG in Dec 2012 posted:Just be glad you aren't using ClearCase. Fate is a cruel mistress.
|
# ? Nov 6, 2014 19:02 |
|
zergstain posted:Fate is a cruel mistress.
|
# ? Nov 6, 2014 19:19 |
|
Same dude from before. He knows one "technique". Every area is designed like the following.code:
Edit: \/\/ hahahah. "Conceptually", you're not willing to grant it the status of theory. bpower fucked around with this message at 21:35 on Nov 6, 2014 |
# ? Nov 6, 2014 19:37 |
|
Well, I'm not going to post any actual code, instead I'll just say:
|
# ? Nov 6, 2014 20:54 |
|
I used Clearcase on Unix for years and I like it way better than loving Perforce
|
# ? Nov 6, 2014 21:49 |
|
Xenoveritas posted:Well, I'm not going to post any actual code, instead I'll just say: quote:
Good luck with refactoring that.
|
# ? Nov 6, 2014 23:07 |
|
Xenoveritas posted:[*]Rather than throwing exceptions, all error handling is done using System.exit(1)
|
# ? Nov 6, 2014 23:53 |
|
Xenoveritas posted:Well, I'm not going to post any actual code, instead I'll just say: Until you got to the last line there, I thought I knew which program you were talking about. Now I'm just going to sob gently at the horrifying revelation that there's more than one of them.
|
# ? Nov 7, 2014 00:39 |
|
|
# ? May 29, 2024 06:10 |
|
Sounds like a (tiny) throwaway demonstration/test app/spike, what's the horror?
|
# ? Nov 7, 2014 02:20 |