Re: [E-devel] e16.6 configure script
On Tue, 2003-07-15 at 19:39, Kim Woelders wrote: OK - I have just committed an attempt to make things work in the good old way. Could everybody who cares please try out their pet configuartion :-) I did a ./autogen.sh --prefix=/opt/enlightenment --enable-fsstd here and it did mostly what I wanted. Stuff is now in /opt/enlightenment/share/* that used to be in /opt/enlightenment/*, but I can live with that. The odd thing(s) are: o No code in e/intl, but it still gets used in e/src/Makefile; o Makefile.in in e/po, but no Makefile, so nothing gets made; o Once built (using old po files and Solaris libintl.so) everything seems fine, *except* iconising stuff makes them spin off to just off the bottom left of the screen. My icon box is usually bottom right. It's like it's iconising to a box in a v-screen to the left, which isn't there. Oddly, iconising in the v-screen to the right throws the icon into the first v-screen's icon box. Did someone comment about an off-by-one bug recently? But the fix to building seems good, anyway =O) Ta, -- o o mailto:[EMAIL PROTECTED] /v\ark R. Bowyer http://www.bowyer.screaming.net [EMAIL PROTECTED] `-' --- /\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/19BC72F6 39D0 551D ABE6 1514 0DC1 X Against HTML Mail / \ GALLIPOLI (adj.) Of the behaviour of a bottom lip trying to spit mouthwash after an injection at the dentist. Hence, loose, floppy, useless. 'She went suddenly Gallipoli in his arms' - Noel Coward. signature.asc Description: This is a digitally signed message part
Re: [E-devel] e16.6 configure script
On Wed, 2003-07-16 at 14:02, Kim Woelders wrote: I did a ./autogen.sh --prefix=/opt/enlightenment --enable-fsstd here and it did mostly what I wanted. Stuff is now in /opt/enlightenment/share/* that used to be in /opt/enlightenment/*, but I can live with that. I think you want ./autogen.sh --prefix=/opt/enlightenment I tried that, and saw the path /opt/enlightenment/enlightenment scolling by in the gcc lines... =O( The odd thing(s) are: o No code in e/intl, but it still gets used in e/src/Makefile; Shouldn't gettext put stuff in intl? Possibly. /usr/bin/gettext on Solaris doesn't know what --version means, so I had to hack autogen.sh to get over that, and then saw a test error later. Doesn't matter though, as Sun's libintl.so seems to work fine. o Makefile.in in e/po, but no Makefile, so nothing gets made; More gettext problems, I guess. Somebody else will have to have an opinion about that. OK. If this is the problem, some checks need to be made for which gettext is there and what needs to be done to continue Solaris support. o Once built (using old po files and Solaris libintl.so) everything seems fine, *except* iconising stuff makes them spin off to just off the bottom left of the screen. My icon box is usually bottom right. It's like it's iconising to a box in a v-screen to the left, which isn't there. Oddly, iconising in the v-screen to the right throws the icon into the first v-screen's icon box. Did someone comment about an off-by-one bug recently? I'll take a look at that. Thanks. It could be down to my config or how I restarted Enlightenment, but I don't think so. Since I have the Gnome panel window list (as well as e's), it wasn't a big problem. Except that I managed to iconise gkrellm, which I'd removed from the skip window list thing... =O{ Ta, -- o o mailto:[EMAIL PROTECTED] /v\ark R. Bowyer http://www.bowyer.screaming.net [EMAIL PROTECTED] `-' --- /\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/19BC72F6 39D0 551D ABE6 1514 0DC1 X Against HTML Mail / \ SAVERNAKE (vb.) To sew municipal crests on to a windcheater in the belief that this will make the wearer appear cosmopolitan. signature.asc Description: This is a digitally signed message part
Re: [E-devel] e16.6 configure script
On Wed, 2003-07-16 at 14:18, Kim Woelders wrote: Sorry, I meant ./autogen.sh --prefix=/opt Ahhh! OK, retried, and yes - that works fine =O) Many thanks. I also looked closer at the po/intl problem, as I noticed this at the end of the ./configure output: creating po/Makefile.in Can't open ./po/Makefile.in.in creating intl/Makefile Can't open ./intl/Makefile.in So I looked closer, and po/Makefile.in is zero length... =O{ Something definitely wrong, but I'm not sure what gettext is supposed to do here. As I say, E runs fine best as I can tell, using libintl.so on Solaris. I quickly hacked the Makefile to use that instead (changed the references to ../intl/libintl.a to -lintl and set BUILD_INCLUDED_LIBINTL to no - twice each). Still not building anything in e/po though, so I'm still stuck with old files. But it installs fine and works this way - except for my continuing iconify to iconbox issue =O} Ta, -- o o mailto:[EMAIL PROTECTED] /v\ark R. Bowyer http://www.bowyer.screaming.net [EMAIL PROTECTED] `-' --- /\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/19BC72F6 39D0 551D ABE6 1514 0DC1 X Against HTML Mail / \ KIRBY MISPERTON (n.) One who kindly attempts to wipe an apparent kirby (q.v.) off another's face with a napkin, and then discovers it to be a wart or other permanent fixture, is said to have committed a 'kirby misperton'. signature.asc Description: This is a digitally signed message part
Re: [E-devel] e16.6 configure script
On Wed, 2003-07-16 at 16:01, Michael Jennings wrote: Feel free to modify setup-gettext.sh to resolve the portability issues. Yeah, I've been looking into that. But Solaris's gettext and the GNU one seem so entirely different - like --version just gets repeated out as input, it doesn't elicit a special response of any kind. So much of this shell script becomes redundant. It looks to me that I have a recent enough version of GNU gettext installed in /usr/local/bin to just stick that ahead in my path and let autogen.sh get on with it. It works the standard Solaris way with hacking, but it doesn't install properly, so what the hell. Not exactly hard to fetch and install GNU gettext - it's on the Companion CD I think =O) So I'm building that way now. Which leads to other problems. Like intl/Makefile still has YACC set to : for some reason. Easy to fix. But then I need to fix plural.y so that %expect is 7 instead of 10, for some reason? Bison errors out otherwise. And then Solaris's m4 doesn't accept -I... Now for some reason, /usr/ccs/bin/m4 was hard coded, path and all, into my bison, so I had to move that out the way and soft link to /usr/local/bin/m4 to fix that. Anyway, it works now. In po, I have a Makefile.in and a Makefile.in.in. Both of them have contents now, and it seems that I have to remove a .in each to get that working properly. That's just too weird for words, that is ;O) Fixed that too, and po makes now =O) So, make install, restart E, and everything seems to be great =O) Many thanks once again. Hopefully *some* of my effort today will help someone other than me =O} Anyone else see this weird po/Makefile[.in]* issue, or YACC in intl/Makefile being :? Ta, -- o o mailto:[EMAIL PROTECTED] /v\ark R. Bowyer http://www.bowyer.screaming.net [EMAIL PROTECTED] `-' --- /\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/19BC72F6 39D0 551D ABE6 1514 0DC1 X Against HTML Mail / \ I think when you go on trial they should have a parrot there that says guilty signature.asc Description: This is a digitally signed message part
Re: [E-devel] e16.6 configure script
OK, I understand now. I knew it was broken, I just didn't understand how you intended to fix it ;O) Many thanks again for your continued work here. Mark. On Mon, 2003-07-14 at 18:46, Kim Woelders wrote: If I understand correctly, you want configure --enable-fsstd --prefix=/opt and configure --enable-fsstd (i.e. --prefix=/usr/local) which are both fsstd? What I suggest is to make --enable-fsstd implied so that the configuration wrt. install dirs is the same as in every other package in the universe. To get the old (intended) default behavior one would have to do something like: configure --bindir=/usr/local/enlightenment/bin --datadir=/usr/local /Kim Mark R. Bowyer wrote: On Sun, 2003-07-13 at 22:25, Kim Woelders wrote: Morten Nilsen wrote: if configure is run without parameters, prefix is set to /usr/local enlightenment expects it's binaries to be in prefix/enlightenment/bin, while they get installed in prefix/bin giving configure --enable-fsstd, this works.. If I use --prefix, it doesn't work unless I give --enable-fsstd. without having extensively tested this, I conclude that e requires --with-fsstd to work at all... cheers, I am aware of this problem. Does anybody actually use the non-fsstd stuff, i.e. install to /usr/enlightenment/{bin,themes,...}? I do, and mentioned this when I did my Solaris compile. /usr/local is shared off our team server in the lab here. /opt is local to each machine. Once 0.16.6 is final, I'll be compiling on Solaris 8 and installing in /usr/local for everyone to be able to use it (although those that have moved to Gnome2 already here seem to be making do with metacity =O( ). Before then, I'm compiling on Solaris 10 and installing in my local /opt. If --enable-fsstd lets me do that correctly without having to clean up afterwards, that's nice, but having it just work like everything else would be way nicer... =O} -- o o mailto:[EMAIL PROTECTED] /v\ark R. Bowyer http://www.bowyer.screaming.net [EMAIL PROTECTED] `-' --- /\ ...fingerprint = 7924 9E9E 7B91 225E B065 \ / ASCII Ribbon Campaign 1024D/19BC72F6 39D0 551D ABE6 1514 0DC1 X Against HTML Mail / \ Is tired old cliche one? signature.asc Description: This is a digitally signed message part
Re: [E-devel] e16.6 configure script
OK - I have just committed an attempt to make things work in the good old way. Could everybody who cares please try out their pet configuartion :-) /Kim --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
On Monday, 14 July 2003, at 21:12:39 (+0200), Morten Nilsen wrote: 1) ./configure /usr/local/bin, complains that dox is not in /usr/local/enlightenment/bin I can't reproduce that here. Most likely a problem on your end. Michael -- Michael Jennings (a.k.a. KainX) http://www.kainx.org/ [EMAIL PROTECTED] n + 1, Inc., http://www.nplus1.net/ Author, Eterm (www.eterm.org) --- What if I fell 15 stories? What if my weight wasn't enough to kill me? What if I were sticky enough to walk the ceiling? Maybe I could live through this.-- Mighty Joe Plum, Live Through This --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
Morten Nilsen wrote: I tried using /usr/local/enlightenment, but that wouldn't work without --with-fsstd either... 1) ./configure /usr/local/bin, complains that dox is not in /usr/local/enlightenment/bin 2) ./configure --prefix=/usr/local/enlightenment /usr/local/enlightenment/enlightenment/bin 3) ./configure --with-fsstd --prefix=/usr/local/enlightenment /usr/local/enlightenment/bin 4) ./configure --with-fsstd --prefix=/usr/local /usr/local/bin looks for dox in the right folder. my conclusion; Everyone wants to use --with-fsstd. It's not as simple as that. I accidentally broke the non-fsstd way some time ago. Now it's a question of either fixing it as it was or doing things as they are done in practically all other packages. /Kim --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
On Monday, 14 July 2003, at 21:50:46 (+0200), Morten Nilsen wrote: quite possible, that.. but still, is there really any reason not to allways have --with-fsstd? Yes. As I see it, the people who want enlightenment into a separate folder, still can do that with --prefix... No, they can't. FSSTD uses things like prefix/bin, whereas the default way uses things like prefix/enlightenment/bin. Michael -- Michael Jennings (a.k.a. KainX) http://www.kainx.org/ [EMAIL PROTECTED] n + 1, Inc., http://www.nplus1.net/ Author, Eterm (www.eterm.org) --- You came from heaven to Earth to show the way; from the Earth to the cross, my debt to pay. From the cross to the grave, from the grave to the sky. Lord, I lift your name on high. -- Maranatha --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
On Mon, 14 Jul 2003 16:00:52 EDT, Michael Jennings [EMAIL PROTECTED] said: On Monday, 14 July 2003, at 21:50:46 (+0200), As I see it, the people who want enlightenment into a separate folder, still can do that with --prefix... No, they can't. FSSTD uses things like prefix/bin, whereas the default way uses things like prefix/enlightenment/bin. what's wrong with ./configure --prefix=/usr/whatever/enlightenment and letting it tack the /bin on that? pgp0.pgp Description: PGP signature
Re: [E-devel] e16.6 configure script
At 21:50 2003-07-14 +0200, Morten Nilsen wrote: Michael Jennings wrote: On Monday, 14 July 2003, at 21:12:39 (+0200), Morten Nilsen wrote: 1) ./configure /usr/local/bin, complains that dox is not in /usr/local/enlightenment/bin I can't reproduce that here. Most likely a problem on your end. quite possible, that.. but still, is there really any reason not to allways have --with-fsstd? As I see it, the people who want enlightenment into a separate folder, still can do that with --prefix... I think most ppl would be just as happy (or even totally unaware) if --with-fsstd was removed and the only option in this matter would be --prefix, since these options do _almost_ the same thing. However, I still don't think it SHOULD be removed! Many of us use UNIX/Linux because we are sick and tired of being bullied around in an MS environment, where we are forced to do things the MS way, and where the defaults are set and options do not exist. --with-fsstd and/or --prefix are options, meaning you don't have to use them if you don't want to. But for me and many others, these options are one of the main reasons for using this system, so let's keep it that way: Flexible and optional! Hmm... I didn't mean for this to get so philosophical, but I think you get the point, neither of these options are harmful, they just keep the system flexible! /J PS: BTW, did this issue come up because someone thought it to be hard to maintain two different options, or just because there was nothing else to discuss? If so, read the old article about the color of the shed from one of the FreeBSD lists. It applies anywhere, anytime! http://www.freebsd.org/doc/en_US.ISO8859-1/books/faq/misc.html#BIKESHED-PAINTING Cheers, -- Morten --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
Michael Jennings said: On Monday, 14 July 2003, at 16:10:39 (-0400), [EMAIL PROTECTED] wrote: what's wrong with ./configure --prefix=/usr/whatever/enlightenment behaviortting it tack the /bin on that? if test x$enable_fsstd = xyes; then ENLIGHTENMENT_ROOT=${datadir}/enlightenment ENLIGHTENMENT_BIN=${bindir} LOCALEDIR=${prefix}/${DATADIRNAME}/locale USE_FSSTD=yes else ENLIGHTENMENT_ROOT=${prefix}/enlightenment ENLIGHTENMENT_BIN=${prefix}/enlightenment/bin LOCALEDIR=${prefix}/enlightenment/locale USE_FSSTD=no fi Wouldn't configure --prefix=/opt/enlightenment --datadir=/opt prove the same layout as the non-fsstd code? Perhaps it would be simplest to strip the non-fsstd code out and make a note on how to get the old behaviour back if they really want it. Or just fix the non-fsstd build, although that's just one more thing that needs to be maintained then. Why not hold a vote and go with the majority. --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
On Monday, 14 July 2003, at 18:57:37 (-0400), [EMAIL PROTECTED] wrote: Why not hold a vote and go with the majority. This isn't a democracy. At best, it's a benevolent dictatorship. :) raster has made it very clear in the past how he wants the default to work. So this argument is pointless. Michael -- Michael Jennings (a.k.a. KainX) http://www.kainx.org/ [EMAIL PROTECTED] n + 1, Inc., http://www.nplus1.net/ Author, Eterm (www.eterm.org) --- Sorry, but my karma just ran over your dogma.-- Unknown --- This SF.Net email sponsored by: Parasoft Error proof Web apps, automate testing more. Download eval WebKing and get a free book. www.parasoft.com/bulletproofapps1 ___ enlightenment-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e16.6 configure script
On Mon, 14 Jul 2003 21:45:31 -0400 Michael Jennings [EMAIL PROTECTED] babbled: (B (B On Monday, 14 July 2003, at 18:57:37 (-0400), (B [EMAIL PROTECTED] wrote: (B (B Why not hold a vote and go with the majority. (B (B This isn't a democracy. At best, it's a benevolent dictatorship. :) (B (B raster has made it very clear in the past how he wants the default to (B work. So this argument is pointless. (B (B:-P~~~ :) (B (B-- (B--- Codito, ergo sum - "I code, therefore I am" (BThe Rasterman (Carsten Haitzler)[EMAIL PROTECTED] $B7'<*(B - $Bhttps://lists.sourceforge.net/lists/listinfo/enlightenment-devel