Re: Games::Frozenbubble is on CPAN
But I'm still wondering if it's really needed. Anyone witnessed sound distortion? Or is it happening only on Windows? Would that be a bug specific to the Windows platform in SDL or mikmod? I will work on this with FROGGS. Thanks. Btw we converted the one mikmod file to OGG. The *.xm one. The header for that file was giving us a lot of trouble. Oh. That would mean a rough increase of 100% in size with no real value, it's not too good. What was the trouble? Mikmod doesn't work anymore? Btw the 2 player game music seems to still be xm: frozen-bubble/share/snd/frozen-mainzik-2p.xm - $save = SDL::Surface-new(-width = $image-width, -height = $image-height, -depth = 32, -Amask = 0 but true); #- grrr... this piece of shit of Amask made the surfaces slightly modify along the print/erase of Hurry and Pause took me so much time to debug and find that the problem came from a bug when Amask is set to 0xFF00 (while it's -supposed- to be set to 0xFF00 with 32-bit graphics!!) - $background-blit($drect, $save, $rects{$image}); + $save = SDL::Surface-new( SDL_ANYFORMAT, $image-w, $image-h, 32, 0 , 0, 0, 0); #- grrr... this piece of shit of Amask made the surfaces slightly modify along the print/erase of Hurry and Pause took me so much time to debug and find that the problem came from a bug when Amask is set to 0xFF00 (while it's -supposed- to be set to 0xFF00 with 32-bit graphics!!) + SDL::Video::blit_surface($background, $drect, $save, $rects{$image}); my comment is probably outdated since 0 but true for Amask disappeared? (did you test long-running Hurry/Pause overlay graphics and didn't notice any slight visual modification over time?) Yes. But I think we should have indiviual tests for this stuff. Maybe when we refactor. Yes we should have more tests, but either we have tests then we can rely on them for modifications, or we do modifications first and then we have to manually test. I don't want a new release of FB with regressions. - if ($app-ticks - $$ticks 1000) { - fb_net::gsend('p'); - $$ticks = $app-ticks; + if (SDL::get_ticks() - $ticks 1000) { + Games::FrozenBubble::Net::gsend('p'); + $ticks = SDL::get_ticks(); changing $$ticks by $ticks modifies behaviour, was that really intended? Yes. I don't know why you had the '$$' but it doesn't work like that. The $ticks is now just an integer. Using $$ticks in that subroutine has nothing to do with sdlperl internals. That code: -=-=---=-=---=-=---=-=---=-=-- sub mp_ping_if_needed { my ($ticks) = @_; if ($app-ticks - $$ticks 1000) { fb_net::gsend('p'); $$ticks = $app-ticks; } } ... my $ticks = $app-ticks; while (1) { mp_ping_if_needed(\$ticks); ... } -=-=---=-=---=-=---=-=---=-=-- is broken with your $$ticks = $ticks modification. Do you see what's happening now? + SDL::Video::wm_set_icon(SDL::Video::load_BMP($FPATH/gfx/pinguins/window_icon_penguin.png)); load_BMP for a PNG looks strange Icons can only be set with BMP on windows. We need to change that. I will make a note of it to use a .bmp file for windows. http://sdlperl.ath.cx/projects/SDLPerl/ticket/135 Fine, but does that line of code works on Linux? - my $tmp = SDL::Surface-new(-width = $high_rect-width/4, -height = $high_rect-height/4, - -depth = 32, -Amask = 0 but true)-display_format; - fb_c_stuff::shrink(surf($tmp), surf($background-display_format), 0, 0, rect($high_rect), 4); - $tmp-blit(undef, $app, SDL::Rect-new(-x = $high_posx, '-y' = $high_posy)); + my $tmp = SDL::Video::display_format(SDL::Surface-new($high_rect-w/4, $high_rect-h/4, 32, 0, 0, 0, 0)); + #Games::FrozenBubble::CStuff::shrink($tmp, SDL::Video::display_format($background), 0, 0, $high_rect, 4); + SDL::Video::blit_surface($tmp, SDL::Rect-new(0,0,$tmp-w,$tmp-h), $app, SDL::Rect-new($high_posx, $high_posy, $tmp-w, $tmp-h)); same - with shrinking disabled the highscores are probably broken? No the highscores should work. Wait what branch are you on? Redesign? This was changed later. I use the branch integration as it was in your http URL earlier. Is that incorrect? I doubt the highscores would work without the shrinking stuff activated. It's how screenshots are rendered. Have you tested it with a non empty highscores list? - if ($i) { - print_('menu', $app, $MENUPOS{xpos_panel}, $ypos, $i, $imgbin{void_panel}-width, 'center'); - } - $ypos += $lineheight; + print_('menu', $app, $MENUPOS{xpos_panel}, $ypos, $i, $imgbin{void_panel}-w, 'center') if $i; + $ypos += $lineheight; I think it's better to stick with the old code style (even if it's probably not perfect) than mixing different code styles; in particular, I avoid veeer longg
Re: Games::Frozenbubble is on CPAN
Please join us on #sdl irc.perl.org it makes this much eaiser. I have joined the channel but I'm only occasionally available there.. That is fine. But you can ask questions on there faster. Yes. The only problem is there are no logs for other people (and for us) as there is in an ML. Thank you so much for reviewing. Thanks for your patience. -- Guillaume Cottenceau - http://zarb.org/~gc/
Re: Games::Frozenbubble is on CPAN
Sorry I did not reply sooner, I was very busy. On Tue, Apr 13, 2010 at 3:28 PM, Kartik Thakore thakore.kar...@gmail.com wrote: Hi Guillaume, I did a git rebase between your master and our redesign branch to make the new integration branch. And it has your commits http://github.com/kthakore/frozen-bubble/commits/integration?page=4 I'll try to look at that soon! Also is there a test suite that you use for frozen-bubble? I would help us get frozen-bubble to production faster. No, there's none, sorry. Moreover are you open to allow us to refactor the frozen-bubble file. It is huge and hard to maintain. We would like to make the SDL, game logic, networking code less coupled. If you had plans for this let us know. We are like blind rats in this. :) Why not. I had no plans for this. The main problem would be newly introduced bugs, mainly because you've little experience with that pile of sh^Hcode. So if you could do that in small verifiable patchsets, each of them not introducing regressions, it would in the end be easier to identify if any regression/bug would be found. I was happy to see it's probably already your mindset, e.g. The workflow I am using for this is to keep running frozenbubble and fix each break. in a commit log. Also I would like you to meet Tobias Liech (FROGGS). He is the other lead on this project. _o/ Oh and one off our dev has had some success in bringing our CPAN fb to MacOSX. great! http://skitch.com/edenc/n614c/fullscreen Right now windows and linux are working but we have to port a lot before a good production release. Ok. Thanks for your work! -- Guillaume Cottenceau - http://zarb.org/~gc/
Re: Games::Frozenbubble is on CPAN
On 2010-04-28, at 8:07 AM, Guillaume Cottenceau gcott...@gmail.com wrote: Sorry I did not reply sooner, I was very busy. On Tue, Apr 13, 2010 at 3:28 PM, Kartik Thakore thakore.kar...@gmail.com wrote: Hi Guillaume, I did a git rebase between your master and our redesign branch to make the new integration branch. And it has your commits http://github.com/kthakore/frozen-bubble/commits/integration?page=4 I'll try to look at that soon! Great thanks. Also is there a test suite that you use for frozen-bubble? I would help us get frozen-bubble to production faster. No, there's none, sorry. Froggs has a module SDLx::Test that compares surfaces. Moreover are you open to allow us to refactor the frozen-bubble file. It is huge and hard to maintain. We would like to make the SDL, game logic, networking code less coupled. If you had plans for this let us know. We are like blind rats in this. :) Why not. I had no plans for this. The main problem would be newly introduced bugs, mainly because you've little experience with that pile of sh^Hcode. So if you could do that in small verifiable patchsets, each of them not introducing regressions, it would in the end be easier to identify if any regression/bug would be found. I was happy to see it's probably already your mindset, e.g. The workflow I am using for this is to keep running frozenbubble and fix each break. in a commit log. Yup. Right now we are just trying to keep the code as is an get fb working on Debian. Once that is done we will refactor in TDD fashion. Unit test a component ( using SDLx::Test see github.com/FROGGS ) then refactor. Currently we have single player working on windows and Linux. See the redesign branch in github.com/kthakore/frozen-bubble. Also I would like you to meet Tobias Liech (FROGGS). He is the other lead on this project. _o/ Oh and one off our dev has had some success in bringing our CPAN fb to MacOSX. great! http://skitch.com/edenc/n614c/fullscreen Right now windows and linux are working but we have to port a lot before a good production release. Ok. Thanks for your work! Np. -- Guillaume Cottenceau - http://zarb.org/~gc/
Re: Games::Frozenbubble is on CPAN
I think so. Let me try this today. Kartik Thakore On 2010-04-13, at 4:10 AM, Guillaume Cottenceau gcott...@gmail.com wrote: can you merge changes from master repository from after your cloning, which seems to date a little? for example commit 9253c759f56e63237cf582460d629467491177ec Author: Guillaume Cottenceau gcott...@gmail.com Date: Thu Oct 22 09:37:07 2009 +0200 do not shutdown socket if socket wasn't even created (DNS lookup failure, for example) is not in your repo On Tue, Apr 13, 2010 at 2:33 AM, Kartik Thakore thakore.kar...@gmail.com wrote: Hi, So we released a dev version of Frozenbubble on CPAN. http://search.cpan.org/%7Ekthakore/Games-FrozenBubble-0.001_1/lib/Games/FrozenBubble.pm We still have work to do but I would love your thoughts. Our repo is on http://github.com/kthakore/frozen-bubble . The branch is redesign. Regards, Kartik Thakore -- Guillaume Cottenceau - http://zarb.org/~gc/