From: David B. <db...@ta...> - 2006-06-12 20:41:54
|
Dear Tuxmath colleagues, I have just committed revision 9 to the svn repository. Highlights include: 1. Better support for playing through a defined list of questions - the player wins if the list is finished with at least one surviving city, with a simple "victory" screen showing Tux dancing. If all cities get blown up, the old "Game Over" graphic appears. In this mode, there is now a counter at the top of the screen showing the number of remaining questions, and the LED numbers display in a "flat panel monitor" on top of Tux's console. This mode is selectable with the "--norepeats" command line argument. 2. Support for questions in the forms num1 + ? = num3 and ? + num2 = num3. These are off by default but can be selected with the command line arguments "--answersfirst" and "--answersmiddle", respectively. Everything seems to work correctly by my own testing. I would like to get this revision into Debian unstable, if possible, so perhaps some more kids and parents can test it. Best wishes, -- David Bruce p.s. Holger, is there anything else I need to do to "release" any of these revisions to the public? Am I supposed to be putting binaries in the svn repository as well as my updated source? On a related note, I have had several local parents ask about a version for (gulp) Windows, so I am starting to look into learning to do a Windows build. |
From: Bill K. <nb...@so...> - 2006-06-12 20:47:36
|
On Mon, Jun 12, 2006 at 04:41:16PM -0400, David Bruce wrote: > Dear Tuxmath colleagues, > > I have just committed revision 9 to the svn repository. Highlights include: David, you rock! Thanks so much for taking the reigns and adding some much-needed improvement to my little game! :^) -- -bill! bi...@ne... http://www.newbreedsoftware.com/ |
From: Holger L. <de...@la...> - 2006-06-13 10:44:23
|
Hi David, On Monday 12 June 2006 22:41, you wrote: > I have just committed revision 9 to the svn repository. Highlights includ= e: Cool! > 1. Better support for playing through a defined list of questions - the > player wins if the list is finished with at least one surviving city, with > a simple "victory" screen showing Tux dancing. If all cities get blown u= p, > the old "Game Over" graphic appears. In this mode, there is now a counter > at the top of the screen showing the number of remaining questions, and t= he > LED numbers display in a "flat panel monitor" on top of Tux's console.=20 > This mode is selectable with the "--norepeats" command line argument. I saw the README was last updated in may, is that new option reflected ther= e? (BTW, you might also want to take a look at the TODO file.) > 2. Support for questions in the forms num1 + ? =3D num3 and ? + num2 =3D= num3. > These are off by default but can be selected with the command line > arguments "--answersfirst" and "--answersmiddle", respectively. ditto.. > Everything seems to work correctly by my own testing. I would like to get > this revision into Debian unstable, if possible, so perhaps some more kids > and parents can test it. Ack, and I will do so soonish, hopefully this weekend, (I'm very busy atm,= =20 sorry). I also definitly want a better version in etch (the next debian=20 version which is supposed to be released in December 2oo6), so we need to=20 have it in unstable long before that, so it gets proper testing. > p.s. Holger, is there anything else I need to do to "release" any of thes= e=20 > revisions to the public?=20 No strictly a need, but a very nice to have: give tuxmath a version number!= :) So far, it has only been released with a date-revision like "2001.09.07" or= =20 "0.0.20050316". I'm not sure if "1.0" is yet appropriate (but you decide),= =20 but something like "1.0" or "0.9" or $whatever would surely be very nice. Having the version number somewhere in the source (setup.h?) is sufficient,= =20 having the svn version in the README is fine, though having it there as wel= l=20 is also nice. Your choice. Also, you could have a look at http://tuxmath.sourceforge.net/ and try to=20 change that or point to a better page. > Am I supposed to be putting binaries in the svn =20 > repository as well as my updated source?=20 No, thats not neccessary. (And not useful.) regards & kudos & thanks for your work, Holger |
From: Andrew Z. <an...@sp...> - 2006-06-14 16:26:31
|
David Bruce wrote: > Dear Tuxmath colleagues, > > I have just committed revision 9 to the svn repository. Highlights include: > > 1. Better support for playing through a defined list of questions - the player > wins if the list is finished with at least one surviving city, with a > simple "victory" screen showing Tux dancing. If all cities get blown up, the > old "Game Over" graphic appears. In this mode, there is now a counter at the > top of the screen showing the number of remaining questions, and the LED > numbers display in a "flat panel monitor" on top of Tux's console. This mode > is selectable with the "--norepeats" command line argument. > > 2. Support for questions in the forms num1 + ? = num3 and ? + num2 = num3. > These are off by default but can be selected with the command line arguments > "--answersfirst" and "--answersmiddle", respectively. > > Everything seems to work correctly by my own testing. I would like to get > this revision into Debian unstable, if possible, so perhaps some more kids > and parents can test it. > Some things I noticed while packaging an .rpm: * The SVN URL was hard to find because Tuxmath has so many web sites on Google. * RPM (at least Fedora's) doesn't like it when you strip the binaries because RPM itself separates the debuginfo into its own .rpm package. * Using "chown root" causes permissions errors while packaging. By default, RPM assumes all files belong to root. * You probably don't want .xvpics in SVN. * rpmlint found non standard permissions /usr/share/tuxmath 0751. It's executable but not readable by others. Also, it would be nice (for testing, for example) to be able to do "make; ./tuxmath" without doing an install, but tuxmath uses absolute pathnames. For other people to test, here's RPMs for Fedora Core 5, and the source RPM should rebuild on a variety of RPM systems. http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.i386.rpm http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.src.rpm http://www.filelodge.com/files/room18/463092/tuxmath-debuginfo-0.9.svn.9-1fc5.i386.rpm Andrew |
From: David B. <db...@ta...> - 2006-06-17 16:50:16
|
Dear Andrew, On Wednesday 14 June 2006 12:25, Andrew Ziem wrote: > Some things I noticed while packaging an .rpm: > * The SVN URL was hard to find because Tuxmath has so many web sites on > Google. > * RPM (at least Fedora's) doesn't like it when you strip the binaries > because RPM itself separates the debuginfo into its own .rpm package. > * Using "chown root" causes permissions errors while packaging. By > default, RPM assumes all files belong to root. I take it you are referring to the "make install" part of the makefile. I haven't touched this since I started maintaining tuxmath, and I honestly don't (yet) know enough about this to feel comfortable changing it, being a self-taught hacker rather than a professional. > * You probably don't want .xvpics in SVN. > * rpmlint found non standard permissions /usr/share/tuxmath 0751. It's > executable but not readable by others. At least on my Debian system, the tuxmath binary is a 0755, both the one installed by the official Debian packaging system and the one under /usr/local installed by "make install". > Also, it would be nice (for testing, for example) to be able to do > "make; ./tuxmath" without doing an install, but tuxmath uses absolute > pathnames. I'll look into this. > For other people to test, here's RPMs for Fedora Core 5, and the source > RPM should rebuild on a variety of RPM systems. > > http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.i386.rp >m > http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.src.rpm > http://www.filelodge.com/files/room18/463092/tuxmath-debuginfo-0.9.svn.9-1f >c5.i386.rpm I have been thinking of using some of my empty disc space for small installs of various distributions and maybe *BSDs for testing purposes. I've never had an rpm-based system. Sorry that I don't yet know enough to address these issues very well, but I'll do my best. Thanks, -- David Bruce |
From: Andrew Z. <an...@sp...> - 2006-06-19 16:13:22
|
David Bruce wrote: >> Some things I noticed while packaging an .rpm: >> * The SVN URL was hard to find because Tuxmath has so many web sites on >> Google. >> * RPM (at least Fedora's) doesn't like it when you strip the binaries >> because RPM itself separates the debuginfo into its own .rpm package. >> * Using "chown root" causes permissions errors while packaging. By >> default, RPM assumes all files belong to root. >> > > I take it you are referring to the "make install" part of the makefile. I > haven't touched this since I started maintaining tuxmath, and I honestly > don't (yet) know enough about this to feel comfortable changing it, being a > self-taught hacker rather than a professional. > Yes, "make install" is what I am referring to. Typically, open source Unix-ish programs use autoconf, but a few use "make install" too. Tuxmath's "make install" just needs some clean up. Maybe these will help: http://www.gnu.org/software/make/manual/html_node/Directory-Variables.html#Directory-Variables http://www.gnu.org/software/make/manual/html_node/Install-Command-Categories.html#Install-Command-Categories http://tinyxml.cvs.sourceforge.net/tinyxml/tinyxml/Makefile?revision=1.14&view=markup (does not include install, but does show some "make" features) http://cvs.sourceforge.net/viewcvs.py/cday/cdayc/makefile?rev=1.12&view=auto (temporarily down right now) ... or find some other BSD/GPL "make install" program to copy. > >> * You probably don't want .xvpics in SVN. >> * rpmlint found non standard permissions /usr/share/tuxmath 0751. It's >> executable but not readable by others. >> > > At least on my Debian system, the tuxmath binary is a 0755, both the one > installed by the official Debian packaging system and the one > under /usr/local installed by "make install". > The tuxmath binary is OK here too, but what about the directory /usr/share/tuxmath (might also be /usr/local/share/tuxmath)? >> For other people to test, here's RPMs for Fedora Core 5, and the source >> RPM should rebuild on a variety of RPM systems. >> >> http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.i386.rp >> m >> http://www.filelodge.com/files/room18/463092/tuxmath-0.9.svn.9-1fc5.src.rpm >> http://www.filelodge.com/files/room18/463092/tuxmath-debuginfo-0.9.svn.9-1f >> c5.i386.rpm >> > I have been thinking of using some of my empty disc space for small installs > of various distributions and maybe *BSDs for testing purposes. I've never > had an rpm-based system. Sorry that I don't yet know enough to address these > issues very well, but I'll do my best. > Maybe you can get the community to do the testing for you or try VMPlayer. (You can try VMPlayer with EasyVMX or http://www.vmware.com/vmtn/appliances/community.html .) Andrew |
From: David B. <db...@ta...> - 2006-06-17 16:36:01
|
Hi Holger, On Tuesday 13 June 2006 06:42, Holger Levsen wrote: > I saw the README was last updated in may, is that new option reflected > there? (BTW, you might also want to take a look at the TODO file.) > > 2. Support for questions in the forms num1 + ? = num3 and ? + num2 = > > num3. These are off by default but can be selected with the command line > > arguments "--answersfirst" and "--answersmiddle", respectively. These were in there but I forgot to change the date. I have just commited svn revision 10 which includes this minor update. > > p.s. Holger, is there anything else I need to do to "release" any of > > these revisions to the public? > > No strictly a need, but a very nice to have: give tuxmath a version number! > :) So far, it has only been released with a date-revision like "2001.09.07" > or "0.0.20050316". I'm not sure if "1.0" is yet appropriate (but you > decide), but something like "1.0" or "0.9" or $whatever would surely be > very nice. > > Having the version number somewhere in the source (setup.h?) is sufficient, > having the svn version in the README is fine, though having it there as > well is also nice. Your choice. I have made the executive decision that we are at version 0.8, and mentioned it in README.txt, tuxmath.h, and CHANGES.txt. I figure we can work on it some more so that version 1.0 may coincide with Etch. > > Also, you could have a look at http://tuxmath.sourceforge.net/ and try to > change that or point to a better page. Yes, that ought to be updated. > > regards & kudos & thanks for your work, > Holger Thanks -- David Bruce |
From: Holger L. <de...@la...> - 2006-06-18 12:28:40
|
Hi, On Saturday 17 June 2006 18:35, you wrote: > > > 2. Support for questions in the forms num1 + ? =3D num3 and ? + num2= =3D > > > num3. These are off by default but can be selected with the command > > > line arguments "--answersfirst" and "--answersmiddle", respectively. > These were in there but I forgot to change the date. I have just commited > svn revision 10 which includes this minor update. Thanks. > I have made the executive decision that we are at version 0.8, and > mentioned it in README.txt, tuxmath.h, and CHANGES.txt. I figure we can > work on it some more so that version 1.0 may coincide with Etch. Yay! I was thinking about removing the quote-signs in the subject when=20 replying :) I will package and upload it soonish. > > Also, you could have a look at http://tuxmath.sourceforge.net/ and try = to > > change that or point to a better page. > Yes, that ought to be updated. Do you have access to it?=20 btw, I just subscribed again to the tuxmath malinglist, somehow I didnt get= =20 any posts there since march 31st :( (Please still cc: me until this is=20 resolved..) regards, Holger |