[arch-dev-public] Add texinfo in base group?
Hi, I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group. Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time. BTW, texinfo provides its own info pages. Currently these files are included in the package but there's no .install script. If I add a script, then technically, it'll need to depend on itself. I'm thinking that it might cause problems for users who are installing it for the first time or with the installer. I could bypass this issue by not running install-info on post-install but only on post-upgrade and removal. Any better idea/comments? Thanks Eric -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
Eric Belanger wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
Seems fine to me.
BTW, texinfo provides its own info pages. Currently these files are included in the package but there's no .install script. If I add a script, then technically, it'll need to depend on itself. I'm thinking that it might cause problems for users who are installing it for the first time or with the installer. I could bypass this issue by not running install-info on post-install but only on post-upgrade and removal. Any better idea/comments?
I do not think that is a problem as the package files have already been installed to their proper place before post-install function from the .install file is called. Just make sure you use the full path to the install-info binary. Allan
Eric Belanger schrieb:
BTW, texinfo provides its own info pages. Currently these files are included in the package but there's no .install script. If I add a script, then technically, it'll need to depend on itself. I'm thinking that it might cause problems for users who are installing it for the first time or with the installer. I could bypass this issue by not running install-info on post-install but only on post-upgrade and removal. Any better idea/comments?
That doesn't make sense to me.
On Sun, Sep 21, 2008 at 12:19 AM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
+1
BTW, texinfo provides its own info pages. Currently these files are included in the package but there's no .install script. If I add a script, then technically, it'll need to depend on itself. I'm thinking that it might cause problems for users who are installing it for the first time or with the installer. I could bypass this issue by not running install-info on post-install but only on post-upgrade and removal. Any better idea/comments?
It is post-install, not pre-install, so there should be no issues at all running it.
On Sun, 21 Sep 2008, Dan McGee wrote:
On Sun, Sep 21, 2008 at 12:19 AM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
+1
BTW, texinfo provides its own info pages. Currently these files are included in the package but there's no .install script. If I add a script, then technically, it'll need to depend on itself. I'm thinking that it might cause problems for users who are installing it for the first time or with the installer. I could bypass this issue by not running install-info on post-install but only on post-upgrade and removal. Any better idea/comments?
It is post-install, not pre-install, so there should be no issues at all running it.
I was probably confusing pre-install with post-install for some reason. I realize now that adding the standard .install file to handle info files won't cause any problems. Eric -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
On Sun, Sep 21, 2008 at 12:19 AM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
That sounds ok to me, but I *do* want to point out something that we've some of us have been irritated about for some time - even IF we assume all of base is installed, we should still probably have proper dependency info. I think we should actually do both.... put it in base, AND make sure we maintain proper deps
On Sun, 21 Sep 2008, Aaron Griffin wrote:
On Sun, Sep 21, 2008 at 12:19 AM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
That sounds ok to me, but I *do* want to point out something that we've some of us have been irritated about for some time - even IF we assume all of base is installed, we should still probably have proper dependency info. I think we should actually do both.... put it in base, AND make sure we maintain proper deps
Yes, doing both would be ideal. The current packages with a missing depends on texinfo: /var/abs/core/binutils /var/abs/core/gcc /var/abs/core/glibc /var/abs/core/libgcrypt /var/abs/core/libtool /var/abs/core/nano /var/abs/extra/enblend-enfuse /var/abs/extra/gdb /var/abs/extra/libidl2 /var/abs/extra/libtasn1 /var/abs/testing/gmp /var/abs/testing/libgcrypt /var/abs/testing/mpfr We should fix the testing packages that are waiting for signoff. As for the rest, they include big packages (i.e. toolchain) that needs to go to testing first. So updating them to just fix the missing depends might be too much work. I could just add the texinfo dependency in svn so it doesn't get forgotten the next time we update these packages. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
On Mon, 22 Sep 2008, Eric Belanger wrote:
On Sun, 21 Sep 2008, Aaron Griffin wrote:
On Sun, Sep 21, 2008 at 12:19 AM, Eric Belanger <belanger@astro.umontreal.ca> wrote:
Hi,
I just want to point out FS#11501. Basically, as we are now adding info pages to packages, we are using a .install file which runs install-info. This tool is part of texinfo. Therefore we should either add a dependency on texinfo to all these packages or alternatively add texinfo to the base group.
Personnally, I prefer adding texinfo to the base group. It is simpler and would make more sense as the tools for the man pages are already in base. Let me know what you think. texinfo is currently out-of-date so I could do that change at the same time.
That sounds ok to me, but I *do* want to point out something that we've some of us have been irritated about for some time - even IF we assume all of base is installed, we should still probably have proper dependency info. I think we should actually do both.... put it in base, AND make sure we maintain proper deps
Yes, doing both would be ideal. The current packages with a missing depends on texinfo:
/var/abs/core/binutils /var/abs/core/gcc /var/abs/core/glibc /var/abs/core/libgcrypt /var/abs/core/libtool /var/abs/core/nano /var/abs/extra/enblend-enfuse /var/abs/extra/gdb /var/abs/extra/libidl2 /var/abs/extra/libtasn1 /var/abs/testing/gmp /var/abs/testing/libgcrypt /var/abs/testing/mpfr
We should fix the testing packages that are waiting for signoff. As for the rest, they include big packages (i.e. toolchain) that needs to go to testing first. So updating them to just fix the missing depends might be too much work. I could just add the texinfo dependency in svn so it doesn't get forgotten the next time we update these packages.
FYI, I've added the texinfo depends for the above packages where it hasn't been fixed yet. This was only done in the svn so it won't be forgotten the next time these are updated. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
participants (5)
-
Aaron Griffin
-
Allan McRae
-
Dan McGee
-
Eric Belanger
-
Thomas Bächler