racket
Re: racket
But supposing I want to move them - I think there are good reasons for doing so, and it isn't that complicated to do - will slkbuild deal correctly with the relative path to the link target (second version) or does it need to be done the first way?
Re: racket
Leaving that question aside, I still have the same error. This is a fresh install of the new beta iso. All I did to it is add a few things like alpine. I downloaded the SLKBUILD and source from people.salixos.org, just as laprjns must have done - and he didn't have this trouble.
To be specific: the drracket script contains the following -
I did this:
in the directory ~/packages/racket, which contained the SLKBUILD, the source, and the .desktop file.
Here is what I've installed:
EDIT
I think I may have it. I also copied over .bashrc from my normal Salix installation, forgetting it temporarily contained the following line to make the in-place racket there work:
Now that path isn't the one in the drracket script (and there's no such directory here). But I think I'll try building it again with that line commented out
EDIT2
No, still the same. The only thing I can think of is the name of the build directory being "racket" ... but why should that be a problem?
EDIT3
No, it wasn't that - a long shot I suppose. I'm truly stumped - and not a little frustrated
To be specific: the drracket script contains the following -
Code: Select all
bindir="/home/cashew/packages/racket/pkg/usr/bin"
Code: Select all
$fakeroot slkbuild -X
Here is what I've installed:
Code: Select all
root[racket]# twapake -s -c 2012-10-02
vimprobable2-1.0.3-i486-2gv
gsettings-desktop-schemas-3.4.2-i486-2
webkit-1.8.2-i486-1rl
icu4c-49.1.2-i486-1
slk-pkgcheck-1.2.1-noarch-1rl
perl-curses-ui-0.9609-i486-1fg
perl-curses-1.28-i486-3fg
perl-term-readkey-2.30.01-i486-3fg
twapake-0.5-noarch-2fg
gmountman-0.4-noarch-1gv
alpine-2.02-i486-1
I think I may have it. I also copied over .bashrc from my normal Salix installation, forgetting it temporarily contained the following line to make the in-place racket there work:
Code: Select all
export PATH=$PATH:~/downloads/racket-5.3/bin/
EDIT2
No, still the same. The only thing I can think of is the name of the build directory being "racket" ... but why should that be a problem?
EDIT3
No, it wasn't that - a long shot I suppose. I'm truly stumped - and not a little frustrated
Re: racket
mimosa wrote:But supposing I want to move them - I think there are good reasons for doing so, and it isn't that complicated to do - will slkbuild deal correctly with the relative path to the link target (second version) or does it need to be done the first way?
As you said and I just realized myself, the /usr/share/doc directory is already sym'ed link to /usr/doc. So as a result of the build the files will end up in /usr/doc/ and the application links will work. No need to complicate the SLKBUILD with mv's, cp's and ln -s'.
“Don’t you see that the whole aim of Newspeak is to narrow the range of thought?"
Re: racket
I'm more concerned at the moment about my inability to get the thing to build for i486. That isn't caused by moving doc files, because all that happens later on. I've just installed core over the top of current/beta, and when I go to bed, I'm going to set it off building racket with my latest SLKBUILD in one directory, and in another with the version you posted that you said built successfully on your own machine (correct $bindir in drracket script) even if the docs are in the wrong place. The idea of this is just a test.
Regarding the question of moving doc files around, if as you say the symlink from /usr/share/doc to /usr/doc takes care of it at the moment of installation, why have I seen SLKBUILDs that move the stuff around manually? I should explain that at the suggestion of fredg, I have concentrated my packaging efforts on anything packaged by gapan previously. It's a detail, but also, in this case the docs go in /usr/share/doc/$pkgname whereas the normal thing seems to be /usr/doc/$pkgname-$pkgver.
The impression I get is that you (as in you and gapan) are suggesting to leave the docs alone because maybe moving them is somehow causing the more serious problem about the path. I don't think so. Regardless of that - have I got the wrong end of the stick here? If as you say the existing symlink puts everything where it should be, what is the docs line in the SLKBUILD for?
Regarding the question of moving doc files around, if as you say the symlink from /usr/share/doc to /usr/doc takes care of it at the moment of installation, why have I seen SLKBUILDs that move the stuff around manually? I should explain that at the suggestion of fredg, I have concentrated my packaging efforts on anything packaged by gapan previously. It's a detail, but also, in this case the docs go in /usr/share/doc/$pkgname whereas the normal thing seems to be /usr/doc/$pkgname-$pkgver.
The impression I get is that you (as in you and gapan) are suggesting to leave the docs alone because maybe moving them is somehow causing the more serious problem about the path. I don't think so. Regardless of that - have I got the wrong end of the stick here? If as you say the existing symlink puts everything where it should be, what is the docs line in the SLKBUILD for?
Re: racket
I'm building racket now with your SLKBUILD. Will report what I find
“Don’t you see that the whole aim of Newspeak is to narrow the range of thought?"
Re: racket
Which version?
EDIT
Assuming you downloaded it afresh just now, it really is the latest. On my previous attempt I enabled something to do with libraries (as in the PKGBUILD) but decided to revert that for the next test, for greater clarity. I wasn't sure whether I'd uplaoded this latest version, but I see I did. That's the one I'll try tonight.
Assuming you are successful and I'm not, I suppose I'm looking at defective hardware or something ...
EDIT
Assuming you downloaded it afresh just now, it really is the latest. On my previous attempt I enabled something to do with libraries (as in the PKGBUILD) but decided to revert that for the next test, for greater clarity. I wasn't sure whether I'd uplaoded this latest version, but I see I did. That's the one I'll try tonight.
Assuming you are successful and I'm not, I suppose I'm looking at defective hardware or something ...
Last edited by mimosa on 3. Oct 2012, 22:55, edited 1 time in total.
Re: racket
The build just completed (32-bit) using your SLKBUILD that I download from your repo about 45 minutes ago. Looks like the path in /usr/bin/drracket is correct:
Code: Select all
#!/bin/sh
# This script was created by make-mred-launcher
# {{{ bindir
bindir="/usr/bin"
# }}} bindir
“Don’t you see that the whole aim of Newspeak is to narrow the range of thought?"
Re: racket
Yes I understand that moving the doc around is not the cause of the path problem.mimosa wrote:I'm more concerned at the moment about my inability to get the thing to build for i486. That isn't caused by moving doc files, because all that happens later on.
A lot of gapan's SLKBUILDs were generated prior to the many improvements and automation that gapan has added to slkbuid over the last few years.mimosa wrote:Regarding the question of moving doc files around, if as you say the symlink from /usr/share/doc to /usr/doc takes care of it at the moment of installation, why have I seen SLKBUILDs that move the stuff around manually? I should explain that at the suggestion of fredg, I have concentrated my packaging efforts on anything packaged by gapan previously. It's a detail, but also, in this case the docs go in /usr/share/doc/$pkgname whereas the normal thing seems to be /usr/doc/$pkgname-$pkgver.
No, i don't think moving the docs around is causing your problem. I think that moving them around is a unnecessary complication to the SLKBUILD file.mimosa wrote:The impression I get is that you (as in you and gapan) are suggesting to leave the docs alone because maybe moving them is somehow causing the more serious problem about the path.
Actually just found out that I was wrong. I just finishing building racket using your SLKBUILD but commenting out all the cp and mv lines and uncommenting the doc line. This build put the application documents in /usr/share/racket/doc and the Copying, Copying.txt, Changelog, README and TODO files in /usr/doc. So therefore the sym link from /usr/share/doc to /usr/doc is not in play here.If as you say the existing symlink puts everything where it should be...
The doc line, when made active by uncommenting, search the src tree for files called out in the array and copies them to /usr/doc/. So if there is a specific doc that you want copied over then you can just added it to the doc array.mimosa wrote:what is the docs line in the SLKBUILD for?
From http://www.salixos.org/wiki/index.php/B ... h_slkbuild
But you are correct the more important thing is to determine why your getting the wrong path on your i486 build. I built racket twice tonight, the first using your SLKBUILD unchanged(*) and the second time removing the cp and mv line as I already said. In both cases, the path was correct and I was able to start racket. Also in both case, clicking on the help menu items did succeed in calling up the appropriate help pages. So the problem with the path resides in you configuration. I don't think its bad hardware.Salix Wiki article Building packages with slkbuild wrote:docs - put any of the docs that need to be copied over, readme install changelog, etc. are common ones. Don't worry about case or path, it does a recursive case insensitive search for them and moves them over when it finds it.
(*) I did make one change to the SLKBUILD for the first build. There is an error with the ln -s line
Code: Select all
ln -s /usr/doc/racket-5.3 $startdir/pkg/usr/share/racket/doc
Code: Select all
ln -s $startdir/pkg/usr/doc/racket-5.3 $startdir/pkg/usr/share/racket/doc
“Don’t you see that the whole aim of Newspeak is to narrow the range of thought?"
Re: racket
@laprjns
Thanks for all your help. Starting at the end ...
I was unsure about the symlink (asked about it above), and actually built it both ways last night. However, the same error with $bindir occurred. This was on a fresh core install of the new beta, and I did nothing whatsoever to it beforehand.
The only thing I can think of is maybe racket doesn't like being built for i486 with a 64-bit processor. To test that, I will try and find room on my wife's machine for a temporary beta install. I tried asking the people at Racket, but no luck there so far.
Incidentally, I also tried building with the SLKBUILD you posted above. That failed, because it should be:
cd $startdir/src/$pkgname-$pkgver/src
before .configure, but (if memory serves) you omit $pkgver.
Meanwhile, I have reinstalled the beta with a few home comforts, and I'll test the package with the relative path to the symlink target.
Thanks for all your help. Starting at the end ...
I was unsure about the symlink (asked about it above), and actually built it both ways last night. However, the same error with $bindir occurred. This was on a fresh core install of the new beta, and I did nothing whatsoever to it beforehand.
The only thing I can think of is maybe racket doesn't like being built for i486 with a 64-bit processor. To test that, I will try and find room on my wife's machine for a temporary beta install. I tried asking the people at Racket, but no luck there so far.
Incidentally, I also tried building with the SLKBUILD you posted above. That failed, because it should be:
cd $startdir/src/$pkgname-$pkgver/src
before .configure, but (if memory serves) you omit $pkgver.
Meanwhile, I have reinstalled the beta with a few home comforts, and I'll test the package with the relative path to the symlink target.
Re: racket
There's no need to test this as there is absolutely no way it happens.mimosa wrote:The only thing I can think of is maybe racket doesn't like being built for i486 with a 64-bit processor. To test that, I will try and find room on my wife's machine for a temporary beta install.