From: Michael R. <re...@eu...> - 2007-02-24 13:18:10
|
Hi there, Thanks to Martin Hejl (who kicked me in the ass some days ago), I released the first release candidate of LCD4Linux-0.10.1 Have a look at http://lcd4linux.bulix.org for details, download, and a (probably incomplete) list of changes. The 'trunk' was bumped to version 0.10.2, but there will be no development until 0.10.1 is released. I'd appreciate any tests and feedback on RC1, so we can release 0.10.1 shortly, and continue work on 0.10.2 (there are plans for scrolling, finally!) Note that I got two cool displays recently, but I withstood to develop a driver for them, but released RC1. So that's the main reason for releasing that seldom: every time I'm thinking of releasing something, someone sends me a new and unsupportd display :-) Have fun, Michael -- Michael Reinelt <re...@eu...> http://home.pages.at/reinelt GPG-Key 0xDF13BA50 ICQ #288386781 |
From: Robert B. <rb...@ge...> - 2007-02-24 23:22:07
|
Michael Reinelt wrote: > Thanks to Martin Hejl (who kicked me in the ass some days ago), I > released the first release candidate of LCD4Linux-0.10.1 > > Have a look at http://lcd4linux.bulix.org for details, download, and a > (probably incomplete) list of changes. Finally! I just bumped the version in Gentoo[1] and ran into few minor problems: 1. ChangeLog As you sure noticed, the ChangeLog shipped with the RC was not current. 2. Filename and URL There's a problem with you hosting the download files in the wiki: Besides the download being pretty slow (about 10kb/s), with a wget query the downloaded file's name will be "lcd4linux-0.10.1-RC1.tar.gz?format=raw". As renaming does not work with the Gentoo package management and mirror scripts, I cannot give your link as "source link" in the package. I worked around that problem by manually introducing the file into the Gentoo mirroring, but if it is possible in the future, please host the file outside the wiki. 3. gcc-Warning Maybe you saw them yourself, but in case you missed them, here they are: drv_BeckmannEgle.c: In function 'drv_BuE_CT_start': drv_BeckmannEgle.c:355: warning: overflow in implicit constant conversion drv_WincorNixdorf.c: In function 'drv_WN_start': drv_WincorNixdorf.c:143: warning: overflow in implicit constant conversion drv_Image.c: In function 'drv_IMG_flush_PPM': drv_Image.c:117: warning: format '%d' expects type 'int', but argument 4 has type 'long unsigned int' drv_Image.c:124: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int' plugin_mysql.c: In function 'configure_mysql': plugin_mysql.c:93: warning: pointer targets in passing argument 6 of 'cfg_number' differ in signedness > The 'trunk' was bumped to version 0.10.2, but there will be no > development until 0.10.1 is released. > > I'd appreciate any tests and feedback on RC1, so we can release 0.10.1 > shortly, and continue work on 0.10.2 (there are plans for scrolling, > finally!) Is there a reason you didn't svn-tag the rc? It's always nice to have the old versions tagged to get the changes at the next release :-) > So that's the main reason for releasing that seldom: every time I'm > thinking of releasing something, someone sends me a new and unsupportd > display :-) Damn. I just get to see the pictures of the displays in the wiki. Must be nice to get some examples ;-P Anyway, nice work. Robert [1] http://sources.gentoo.org/viewcvs.py/gentoo-x86/app-misc/lcd4linux/ |
From: Michael R. <re...@eu...> - 2007-02-25 07:39:13
|
Hi Robert, > 1. ChangeLog > As you sure noticed, the ChangeLog shipped with the RC was not current. Yes, I've noticed that, fill be fixed in RC2 > 2. Filename and URL > There's a problem with you hosting the download files in the wiki: > Besides the download being pretty slow (about 10kb/s), with a wget query > the downloaded file's name will be > "lcd4linux-0.10.1-RC1.tar.gz?format=raw". As renaming does not work with > the Gentoo package management and mirror scripts, I cannot give your > link as "source link" in the package. I worked around that problem by > manually introducing the file into the Gentoo mirroring, but if it is > possible in the future, please host the file outside the wiki. Hmm, thanks for pointing that out. Well, I could use SourceForge's Download Service, although I don't like to :-) Sam, any ideas? > 3. gcc-Warning > Maybe you saw them yourself, but in case you missed them, here they are: Strange, I don't see this warnings. What version of gcc do you use? here its gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) As long as I don't get the warnings, it will be hard to fix'em. > Is there a reason you didn't svn-tag the rc? It's always nice to have > the old versions tagged to get the changes at the next release :-) Yes, I don't want to tag RC's, but I will tag 0.10.1. I hope that there will be no more RC's in the future, but a much more frequent release cycle. > Damn. I just get to see the pictures of the displays in the wiki. Must > be nice to get some examples ;-P Let me tell you that this photo is quite old, and my collection is about 5 times greater by now :-) Thanks for your feedback! bye, Michael -- Michael Reinelt <re...@eu...> http://home.pages.at/reinelt GPG-Key 0xDF13BA50 ICQ #288386781 |
From: Robert B. <rb...@ge...> - 2007-02-25 11:48:56
Attachments:
signature.asc
|
(taking this off -users) Michael Reinelt wrote: >> 3. gcc-Warning >> Maybe you saw them yourself, but in case you missed them, here they ar= e: > Strange, I don't see this warnings. What version of gcc do you use? her= e > its gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) I'm using 'gcc (GCC) 4.1.1 (Gentoo 4.1.1-r3)' for x86_64-pc-linux-gnu. > As long as I don't get the warnings, it will be hard to fix'em. I can compile-test if you have any patches, but I don't have the time to look into the code right now. Got an exam on tuesday. >> Is there a reason you didn't svn-tag the rc? It's always nice to have >> the old versions tagged to get the changes at the next release :-) > Yes, I don't want to tag RC's, but I will tag 0.10.1. I hope that there= > will be no more RC's in the future, but a much more frequent release cy= cle. Sounds good. Robert |
From: Michael R. <re...@eu...> - 2007-02-25 12:49:41
|
Hi Robert, >> Strange, I don't see this warnings. What version of gcc do you use? here >> its gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) > > I'm using 'gcc (GCC) 4.1.1 (Gentoo 4.1.1-r3)' for x86_64-pc-linux-gnu. Oh, you're using x86_64... that may explain things... In the meantime, I upgraded to gcc-4.2; changed the warning options to very very pedantic, and fixed lots of other warnings (eg the struc initialization from obsolete 'field:value' to '.field=value') I didn't get your warnings, but I tried to fix them "blindly". fixes are in trunk and backported to branches/0.10.1 > I can compile-test if you have any patches Please do so! TIA, Michael btw, do you know of a way to reduce the "chattyness" of the compilation process? Since I upgraded autoconf/automake/autowhatever some time ago, the compilation of one single file emits 3-5 lines, and I regularly overlook these warnings... -- Michael Reinelt <re...@eu...> http://home.pages.at/reinelt GPG-Key 0xDF13BA50 ICQ #288386781 |
From: Ernst B. <e.b...@xe...> - 2007-02-26 10:03:10
|
On Sonntag, 25. Februar 2007, Michael Reinelt wrote: > I didn't get your warnings, but I tried to fix them "blindly". I think you can pass GCC an option to pretend to be on 64Bit to see that warning, IIRC it was "-m64". Never tested that however, on 64Bit myself ;) > > btw, do you know of a way to reduce the "chattyness" of the compilation > process? Since I upgraded autoconf/automake/autowhatever some time ago, > the compilation of one single file emits 3-5 lines, and I regularly > overlook these warnings... Easy: # make > /dev/null Warnings go to STDERR, and are still visible. Alternatively, you can tell GCC to bail out on warnings, "-pedantic-errors" and "-Wfatal-errors" seem to do that. /Ernst |
From: Michael R. <re...@eu...> - 2007-02-26 10:38:15
|
Hi Ernst, >> I didn't get your warnings, but I tried to fix them "blindly". > I think you can pass GCC an option to pretend to be on 64Bit to see that > warning, IIRC it was "-m64". Great thanks! Now the warnings should be fixed! (and commited to SVN, both to trunk and 0.10.1 branch) > Never tested that however, on 64Bit myself ;) with -m64 I get two strange assembler errors, but I think the result from the fact that I'm on 32Bit (at least I hope so :-) >> btw, do you know of a way to reduce the "chattyness" of the compilation >> process? > Easy: > # make > /dev/null Uh... shame, shame, shame on me :-) -- Michael Reinelt <re...@eu...> http://home.pages.at/reinelt GPG-Key 0xDF13BA50 ICQ #288386781 |
From: Robert B. <rb...@ge...> - 2007-03-11 11:46:36
Attachments:
signature.asc
|
Hi Michael and Ernst, Michael Reinelt wrote: >>> I didn't get your warnings, but I tried to fix them "blindly". >> I think you can pass GCC an option to pretend to be on 64Bit to see th= at=20 >> warning, IIRC it was "-m64". >=20 > Great thanks! Now the warnings should be fixed! (and commited to SVN, > both to trunk and 0.10.1 branch) That fixed all of the warnings I posted before, except one: drv_Image.c: In function 'drv_IMG_flush_PPM': drv_Image.c:117: warning: format '%d' expects type 'int', but argument 4 has type 'long unsigned int' drv_Image.c:124: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int' >> Never tested that however, on 64Bit myself ;) > with -m64 I get two strange assembler errors, but I think the result > from the fact that I'm on 32Bit (at least I hope so :-) I guess that does not work because you need a complete 64bit toolchain. On Gentoo, you can build it using the "crossdev" package, but I don't know how to do that on other platforms :-) >>> btw, do you know of a way to reduce the "chattyness" of the compilati= on >>> process? >=20 >> Easy: >> # make > /dev/null > Uh... shame, shame, shame on me :-) You can also call make -s to silence the make calls except for warnings and errors (which is the same as > /dev/null). There are also patches to automake to let the user choose whether (s)he wants a pretty output (as in the Linux kernel). This, however, should break again when autoreconf'ing the created files on other systems. See the following project for details: http://kim.tensta.gannert.se/projects/pretty-am/ Regards, Robert |