Re: [Geeqie-devel] Release 1.3
> On Tue, May 31, 2016 at 07:39:15PM +0100, Colin Clark wrote: >>> There may be people trying to use Geeqie who struggle just to download a >>> tar and do a make. For them, compiler warnings are an unnecessary worry, >>> because they do not understand that the warnings are in fact irrelevant >>> to them. >> I don't think that people who struggle over that output will ever >> compile geeqie themself. They use the version, the distribution is >> giving him. >> >I understand your point of view, but I am not in complete agreement. > >I can do a few things with software, but I have no patience when I am >trying to get other people's software to run. If I download something >from the interweb, do a configure/make and get a screen-full of >messages, I will not spend any time trying to understand it. I will find >another app and try that. > >I would rather leave endless messages that may or may not be meaningful >to the profis. > >I much prefer programs that, if they create a run-able binary, show no >message other than "It works, try it". I would prefer Geeqie to be that >way, but I am also happy for it to stay as it is. Those are what those warnings are meant for, for the developer to migrate from functions that are going to be deprecated within the future. Sometimes those warnings may be displayed only on a person's unique platform or the person's build environment which is different from the developer's build environment. If you think you're only seeing the warnings (due to building on sparc, bsd, ...), then maybe forwarding the warnings to the mailing list or developers would help bring them to light. (Correct me if I'm wrong here, but just realized this perspective just now.) I think the warnings I see while building, are clearly indicated by "Warning: " prefix. When compiling breaks, the warnings are usually a strong indicator just prior to a failed compiled statement. Usually you'll see 20-40 warnings when scanning the compiler stdout or compiler text logs, immediately prior to a statement compiling failure. Without these sometimes numerous warnings prior to breaking, the person compiling is left for looking for one compiler error amongst compiler stdout. (This usually occurs within when using the option "-j" with large values, etc. Most times, users can just ignore the warnings unless the compiler fails. Sort of like buying a used car with it's documented history showing. Hiding the warnings, would be likely buying a used car without knowing it's history. Also, warnings are sometimes indicative of unmaintained code, likely to break within the future. Most users never see the compiler messages, as most just use the software. If I'm not mistaken, those compiling can further customize the compiler messages by modifying their make.conf file. Most just leave them at defaults, while developers tend to significantly increase verbosity... especially when things break. -- Roger http://rogerx.freeshell.org/ -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
Colin> My view, which is but one of many, is that all compiler Colin> messages other than errors should be inhibited for a standard Colin> build. I think we should inhibit the deprecated warnings, they're useless unless we're doing a new release and/or moving to a new version of GTK. Otherwise they just provide no useful information at this time. The (whitespace) dammaged patch I sent goes a long way to making things quieter, so we can work on the real errors. Colin> There may be people trying to use Geeqie who struggle just to Colin> download a tar and do a make. For them, compiler warnings are Colin> an unnecessary worry, because they do not understand that the Colin> warnings are in fact irrelevant to them. Exactly. These days it's so easy to do: git clone ... && ./configure && make that we should support that model if at all possible. Colin> They are only relevant to the developers - and unfortunately Colin> there seems to be very few of those around at the moment. I'm slowly trying to help out where I can. I don't know GTK at all. Colin> Deprecated warnings will never go away, because as fast as they Colin> are fixed, the GTK developers are at work creating a few more. Heh. Colin> My view is that all warnings should only be enabled with the Colin> --enable-developer option. And just update the readme to reflect that change. -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 Hi, Am So den 29. Mai 2016 um 18:02 schrieb Colin Clark: > My view, which is but one of many, is that all compiler messages other > than errors should be inhibited for a standard build. You mean like the kernel messages? > There may be people trying to use Geeqie who struggle just to download a > tar and do a make. For them, compiler warnings are an unnecessary worry, > because they do not understand that the warnings are in fact irrelevant > to them. I don't think that people who struggle over that output will ever compile geeqie themself. They use the version, the distribution is giving him. > Deprecated warnings will never go away, because as fast as they are > fixed, the GTK developers are at work creating a few more. That is a sad fact, yes. Regards Klaus - -- Klaus Ethgen http://www.ethgen.ch/ pub 4096R/4E20AF1C 2011-05-16 Klaus EthgenFingerprint: 85D4 CA42 952C 949B 1753 62B3 79D0 B06F 4E20 AF1C -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Charset: ISO-8859-1 iQGcBAEBCgAGBQJXTGWiAAoJEKZ8CrGAGfasnOcMAKmL+6UfkRghMBl6syhXplmv q1dR2ZRw2aa8grbg5L5IfUThbo/uGKBo3WzRlyoq6MHGDw6HzlvPxJc8qkv/NwWP IFdwpBTO1reVIO37NByMh2X+bfvpFJZmoyY+i0UtXUwZ5zzYqClF7m9qd02W1fkZ IGyH6dFaU404mZ2pOb11GvlEz9wmk3d/oAWE2GvzMvizJSLwY3vJOd7XMpkZIWt0 EJKZTiA7m1MGqhdK0Pb1/zjUxmGT91ii7/jZ0/ts8ywbmDVAR+kWjkbnotbI2bRh LDSSZiCk3bpdeDv/LhgXggXuwRwJLCkJXTX8KF0R+h9EJboseq6Ekk5g8cmi+azX /WrbxxeUxX3xXgeN6XZuqHGrw2PiUa+/9Gmu/2E93I+CRkrgJeJGKAWLSwo265Hj pGnx3GXftU3LIi35fEYugDKP2En4+j1P43t4Wy0/NMVfkWCiM0m676VXMjWC02qT 3ho0NNda5wEBsI4Im6NJP4zNFYZJisxMhbAxMaDDvQ== =+ta/ -END PGP SIGNATURE- -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
> On Sun, May 29, 2016 at 06:02:00PM +0100, Colin Clark wrote: >On 29/05/16 14:36, Klaus Ethgen wrote: >> -BEGIN PGP SIGNED MESSAGE- >> Hash: SHA512 >> >> Am So den 29. Mai 2016 um 4:01 schrieb John Stoffel: >>> Just gettings the warnings fixed would be a good goal in my mind. >> True. >> >> As you might have noticed, I already did some cleanup in merges 2943df6 >> and 6164a63. >> >> Regards >> Klaus >> - -- >My view, which is but one of many, is that all compiler messages other >than errors should be inhibited for a standard build. > >There may be people trying to use Geeqie who struggle just to download a >tar and do a make. For them, compiler warnings are an unnecessary worry, >because they do not understand that the warnings are in fact irrelevant >to them. > >They are only relevant to the developers - and unfortunately there seems >to be very few of those around at the moment. > >Deprecated warnings will never go away, because as fast as they are >fixed, the GTK developers are at work creating a few more. > >My view is that all warnings should only be enabled with the >--enable-developer option. The only problem with this configure option, albeit this is the first time I've ever heard of such a "configure --enable-developer" option, nobody will be prepared for when something then breaks. With the omission of printing warnings during compiling, developers will effectively loose hindsight, unless they spend more time to further state further configure time options to print the warnings. Most who compile, need to keep apprised of deprecated warnings, for future problems or newer version releases of depends. Furthermore, you're asking to change something that has been default for more than 10-20 years. It would probably be better to inhibit a "--no-warnings" configure option. In which I think there already is if you manually augment the gcc/gcc++ parameters. (ie. cc --help=warnings; -W/Wextra) But when I'm compiling, I usually always add the -W/Wextra, and further when debugging -ggdb, etc... The more the merrier.) I think if somebody wants to make the compiling statements prettier or more readable, I would suggest using Color GCC. Color GCC compiler statements are more well readable too! The other option used for source based distributions, is providing the package manager responsible for handling "extracting, configuring, building and installing", a quiet parameter. (ie. --quiet) Most people compiling, would likely use the package manager for quieting output. When compiling every other day packages, I do use --quiet. -- Roger http://rogerx.freeshell.org/ -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
On 29/05/16 14:36, Klaus Ethgen wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA512 > > Am So den 29. Mai 2016 um 4:01 schrieb John Stoffel: >> Just gettings the warnings fixed would be a good goal in my mind. > True. > > As you might have noticed, I already did some cleanup in merges 2943df6 > and 6164a63. > > Regards > Klaus > - -- My view, which is but one of many, is that all compiler messages other than errors should be inhibited for a standard build. There may be people trying to use Geeqie who struggle just to download a tar and do a make. For them, compiler warnings are an unnecessary worry, because they do not understand that the warnings are in fact irrelevant to them. They are only relevant to the developers - and unfortunately there seems to be very few of those around at the moment. Deprecated warnings will never go away, because as fast as they are fixed, the GTK developers are at work creating a few more. My view is that all warnings should only be enabled with the --enable-developer option. -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 Am So den 29. Mai 2016 um 4:01 schrieb John Stoffel: > Just gettings the warnings fixed would be a good goal in my mind. True. As you might have noticed, I already did some cleanup in merges 2943df6 and 6164a63. Regards Klaus - -- Klaus Ethgen http://www.ethgen.ch/ pub 4096R/4E20AF1C 2011-05-16 Klaus EthgenFingerprint: 85D4 CA42 952C 949B 1753 62B3 79D0 B06F 4E20 AF1C -BEGIN PGP SIGNATURE- Version: GnuPG v1 Comment: Charset: ISO-8859-1 iQGcBAEBCgAGBQJXSvA+AAoJEKZ8CrGAGfaszFoMAKtDm4hYx+rs22DY7DC485DK aBwAO+0PTYJsrUOBk0gZXpllrPXVM2U5hwamyiRjc7mwu/loqVAVxdqS7j428L80 +BM+8oOLXJCDtJaAShM5+U4PeGH0Ny7bZfPr7lv2tPA7EP7iyez8l0ein2bJ0Eqe ISQKxNz2o+ueAGMylPD7moM9cOrQjCUFcsrJNOgJ7xVQAiZO03HQcqgeaEqodN83 cKU+awPjf6JCmHHREn199ppzdSduNK4auq27gXNMu+aR+AasR8rYL5jdmOUrb78u GKUQ/Iq94suu5BN6tDF70KMgiFs75FhrRo8i/ooyFUqsF7aYcqLBwFNtzD9akgoL 4xH0+0MtrD/GdRc7YtX7empZ1vO8ibqYNF9rR3pSQcGH/bG1V9jFckay8vSYX3r9 SNosdcA9xipkFmua8XjDXfl/N4mgt65DRD5gcKGrHgrfARVzqXEvKA1Z8tph21en TBTyxtO/5NMCcpOLQmmu6e2wPjAqB1OJZjdGNzGsLg== =YVMi -END PGP SIGNATURE- -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
> "Klaus" == Klaus Ethgenwrites: Klaus> -BEGIN PGP SIGNED MESSAGE- Klaus> Hash: SHA512 Klaus> Hi, Klaus> Am Fr den 27. Mai 2016 um 19:33 schrieb John Stoffel: >> Just to let you know, I'm seeing a bunch of warnings when I did a test >> compile on Debian Jesse x86_64 system. I probably don't have all the >> right libraries installed or something, but I figured cleaning this up >> would be a good start. Klaus> Yes, we use some deprecated stuff. Some of them I even don't Klaus> want to replace. (The inline image stuff) It would be nice if we could turn those warnings off somehow, because they will just hide other more important errors. Klaus> There are even some warnings in rcfile.c.. Nothing critical but Klaus> might need some small work. Well now I think I can try to work on those cleanups if I can find some time. Just gettings the warnings fixed would be a good goal in my mind. John -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel
Re: [Geeqie-devel] Release 1.3
Just to let you know, I'm seeing a bunch of warnings when I did a test compile on Debian Jesse x86_64 system. I probably don't have all the right libraries installed or something, but I figured cleaning this up would be a good start. Maybe I can do something this weekend... Sorry for the crappy cut'n'paste, my terminals are doing something wonky... and this is only a small subset of the warnings I'm getting. It would be nice to get rid of these deprecated warnings for regular builds image.c: In function ‘image_select’: image.c:1679:3: warning: ‘gtk_widget_set_state’ is deprecated (declared at /usr/include/gtk-3.0/gtk] gtk_widget_set_state(imd->widget, GTK_STATE_SELECTED); ^ image.c:1680:3: warning: ‘gtk_widget_set_state’ is deprecated (declared at /usr/include/gtk-3.0/gtk] gtk_widget_set_state(imd->pr, GTK_STATE_NORMAL); /* do not propagate */ ^ image.c:1683:3: warning: ‘gtk_widget_set_state’ is deprecated (declared at /usr/include/gtk-3.0/gtk] gtk_widget_set_state(imd->widget, GTK_STATE_NORMAL); ^ image.c: In function ‘selectable_frame_draw_cb’: image.c:1788:2: warning: ‘gtk_paint_flat_box’ is deprecated (declared at /usr/include/gtk-3.0/gtk/d] gtk_paint_flat_box(gtk_widget_get_style(widget), ^ image.c:1788:2: warning: ‘gtk_widget_get_style’ is deprecated (declared at /usr/include/gtk-3.0/gtk] image.c:1790:7: warning: ‘gtk_widget_get_state’ is deprecated (declared at /usr/include/gtk-3.0/gtk] gtk_widget_get_state(widget), ^ image.c:1799:3: warning: ‘gtk_paint_focus’ is deprecated (declared at /usr/include/gtk-3.0/gtk/depr] gtk_paint_focus(gtk_widget_get_style(widget), cr, GTK_STATE_ACTIVE, ^ image.c:1799:3: warning: ‘gtk_widget_get_style’ is deprecated (declared at /usr/include/gtk-3.0/gtk] image.c:1806:3: warning: ‘gtk_paint_shadow’ is deprecated (declared at /usr/include/gtk-3.0/gtk/dep] gtk_paint_shadow(gtk_widget_get_style(widget), cr, GTK_STATE_NORMAL, GTK_SHADOW_IN, ^ image.c:1806:3: warning: ‘gtk_widget_get_style’ is deprecated (declared at /usr/include/gtk-3.0/gtk] image.c: In function ‘image_new’: image.c:1916:2: warning: ‘gtk_vbox_new’ is deprecated (declared at /usr/include/gtk-3.0/gtk/depreca] imd->widget = gtk_vbox_new(0, 0); ^ mv -f .deps/image.Tpo .deps/image.Po -- What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports. https://ad.doubleclick.net/ddm/clk/305295220;132659582;e ___ Geeqie-devel mailing list Geeqie-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geeqie-devel