From zeeshanak at gnome.org Thu May 8 16:30:01 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Thu, 8 May 2014 17:30:01 +0100 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' Message-ID: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> Since windows installer only supports 'en-US' language[1], specifying 'en-GB' in installer script breaks unattended installation. [1] http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx --- data/datamaps/windows-lang.xml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/data/datamaps/windows-lang.xml b/data/datamaps/windows-lang.xml index 6b3efe1..9d728a8 100644 --- a/data/datamaps/windows-lang.xml +++ b/data/datamaps/windows-lang.xml @@ -91,7 +91,7 @@ language codes used by Windows. --> - + -- 1.9.0 From fabiano at fidencio.org Thu May 8 22:31:30 2014 From: fabiano at fidencio.org (=?UTF-8?Q?Fabiano_Fid=C3=AAncio?=) Date: Fri, 9 May 2014 00:31:30 +0200 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> Message-ID: Zeeshan, On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) wrote: > Since windows installer only supports 'en-US' language[1], specifying > 'en-GB' in installer script breaks unattended installation. > Did you try mapping based on http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked on IRC? I think a generic solution would be better than this one. Best Regards, -- Fabiano Fid?ncio -------------- next part -------------- An HTML attachment was scrubbed... URL: From zeeshanak at gnome.org Fri May 9 12:05:37 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Fri, 9 May 2014 13:05:37 +0100 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> Message-ID: On Thu, May 8, 2014 at 11:31 PM, Fabiano Fid?ncio wrote: > Zeeshan, > > > On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) > wrote: >> >> Since windows installer only supports 'en-US' language[1], specifying >> 'en-GB' in installer script breaks unattended installation. > > > Did you try mapping based on > http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked on > IRC? > I think a generic solution would be better than this one. Yes, that link is even in the commit log. :) I didn't see any other locale that needs a different mapping. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From fabiano at fidencio.org Fri May 9 21:40:58 2014 From: fabiano at fidencio.org (=?UTF-8?Q?Fabiano_Fid=C3=AAncio?=) Date: Fri, 9 May 2014 23:40:58 +0200 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> Message-ID: On Fri, May 9, 2014 at 2:05 PM, Zeeshan Ali (Khattak) wrote: > On Thu, May 8, 2014 at 11:31 PM, Fabiano Fid?ncio > wrote: > > Zeeshan, > > > > > > On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) < > zeeshanak at gnome.org> > > wrote: > >> > >> Since windows installer only supports 'en-US' language[1], specifying > >> 'en-GB' in installer script breaks unattended installation. > > > > > > Did you try mapping based on > > http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked > on > > IRC? > > I think a generic solution would be better than this one. > > Yes, that link is even in the commit log. :) I didn't see any other > locale that needs a different mapping. > Okay, go for it. Best Regards, -- Fabiano Fid?ncio -------------- next part -------------- An HTML attachment was scrubbed... URL: From cfergeau at redhat.com Mon May 12 07:54:58 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Mon, 12 May 2014 09:54:58 +0200 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140512075458.GI14434@edamame.cdg.redhat.com> On Fri, May 09, 2014 at 01:05:37PM +0100, Zeeshan Ali (Khattak) wrote: > On Thu, May 8, 2014 at 11:31 PM, Fabiano Fid?ncio wrote: > > Zeeshan, > > > > > > On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) > > wrote: > >> > >> Since windows installer only supports 'en-US' language[1], specifying > >> 'en-GB' in installer script breaks unattended installation. > > > > > > Did you try mapping based on > > http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked on > > IRC? > > I think a generic solution would be better than this one. > > Yes, that link is even in the commit log. :) I didn't see any other > locale that needs a different mapping. The commit log does not tell the full story and could mention that this is not a complete solution as language cannot be en-GB while locale can be en-GB, and InputLocale needs to be something different as well: en-US 0809:00000809 en-GB en-US en-GB Christophe From zeeshanak at gnome.org Mon May 12 15:09:33 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 16:09:33 +0100 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: <20140512075458.GI14434@edamame.cdg.redhat.com> References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> <20140512075458.GI14434@edamame.cdg.redhat.com> Message-ID: On Mon, May 12, 2014 at 8:54 AM, Christophe Fergeau wrote: > On Fri, May 09, 2014 at 01:05:37PM +0100, Zeeshan Ali (Khattak) wrote: >> On Thu, May 8, 2014 at 11:31 PM, Fabiano Fid?ncio wrote: >> > Zeeshan, >> > >> > >> > On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) >> > wrote: >> >> >> >> Since windows installer only supports 'en-US' language[1], specifying >> >> 'en-GB' in installer script breaks unattended installation. >> > >> > >> > Did you try mapping based on >> > http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked on >> > IRC? >> > I think a generic solution would be better than this one. >> >> Yes, that link is even in the commit log. :) I didn't see any other >> locale that needs a different mapping. > > The commit log does not tell the full story and could mention that this is not > a complete solution as language cannot be en-GB while locale can be en-GB, and Sure, I'll update the log. > InputLocale needs to be something different as well: We don't set that so not relevant to this patch. > > > en-US > > 0809:00000809 > en-GB > en-US > en-GB > > > Christophe -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Mon May 12 15:14:55 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 16:14:55 +0100 Subject: [Libosinfo] [PATCH] datamap,windows: Map 'en_GB' to 'en-US' In-Reply-To: References: <1399566601-15958-1-git-send-email-zeeshanak@gnome.org> <20140512075458.GI14434@edamame.cdg.redhat.com> Message-ID: On Mon, May 12, 2014 at 4:09 PM, Zeeshan Ali (Khattak) wrote: > On Mon, May 12, 2014 at 8:54 AM, Christophe Fergeau wrote: >> On Fri, May 09, 2014 at 01:05:37PM +0100, Zeeshan Ali (Khattak) wrote: >>> On Thu, May 8, 2014 at 11:31 PM, Fabiano Fid?ncio wrote: >>> > Zeeshan, >>> > >>> > >>> > On Thu, May 8, 2014 at 6:30 PM, Zeeshan Ali (Khattak) >>> > wrote: >>> >> >>> >> Since windows installer only supports 'en-US' language[1], specifying >>> >> 'en-GB' in installer script breaks unattended installation. >>> > >>> > >>> > Did you try mapping based on >>> > http://technet.microsoft.com/library/dd744369(WS.10).aspx, as we talked on >>> > IRC? >>> > I think a generic solution would be better than this one. >>> >>> Yes, that link is even in the commit log. :) I didn't see any other >>> locale that needs a different mapping. >> >> The commit log does not tell the full story and could mention that this is not >> a complete solution as language cannot be en-GB while locale can be en-GB, and > > Sure, I'll update the log. Thinking more about this, since its only the *UILanguage that needs this mapping and en-GB should be set for others, I wonder if my patch is a good solution at all. Maybe we can just do a simple template for *UILanguage that does the mapping? >> InputLocale needs to be something different as well: > > We don't set that so not relevant to this patch. > >> >> >> en-US >> >> 0809:00000809 >> en-GB >> en-US >> en-GB >> >> >> Christophe > > > > -- > Regards, > > Zeeshan Ali (Khattak) > ________________________________________ > Befriend GNOME: http://www.gnome.org/friends/ -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Mon May 12 15:38:34 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 16:38:34 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters Message-ID: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> Since 'ComputerName' is supposed to be max 15 characters, unattended installation breaks if this is set to a string longer than that. Lets crop the string to ensure that doesn't happen. http://technet.microsoft.com/en-us/library/ff715676.aspx --- data/install-scripts/windows-sif.xml | 9 ++++++++- data/install-scripts/windows-unattend.xml | 9 ++++++++- 2 files changed, 16 insertions(+), 2 deletions(-) diff --git a/data/install-scripts/windows-sif.xml b/data/install-scripts/windows-sif.xml index dc28533..71aa731 100644 --- a/data/install-scripts/windows-sif.xml +++ b/data/install-scripts/windows-sif.xml @@ -122,6 +122,13 @@ + + + + + + + ;SetupMgrTag [Data] @@ -177,7 +184,7 @@ FullName="" - ComputerName="" + ComputerName="" [Display] diff --git a/data/install-scripts/windows-unattend.xml b/data/install-scripts/windows-unattend.xml index 046bf0f..3159281 100644 --- a/data/install-scripts/windows-unattend.xml +++ b/data/install-scripts/windows-unattend.xml @@ -305,6 +305,13 @@ + + + + + + + @@ -388,7 +395,7 @@ - + -- 1.9.0 From zeeshanak at gnome.org Mon May 12 15:38:35 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 16:38:35 +0100 Subject: [Libosinfo] [PATCH 2/2] install-config: Document implicit hostname changes In-Reply-To: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> Message-ID: <1399909115-32661-2-git-send-email-zeeshanak@gnome.org> Since we need to chop the hostname for windows, lets warn the app developers about this in the API documentation. --- osinfo/osinfo_install_config.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c index de4ee42..c7d41d0 100644 --- a/osinfo/osinfo_install_config.c +++ b/osinfo/osinfo_install_config.c @@ -343,6 +343,20 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); } +/** + * osinfo_install_config_set_hostname: + * @config: the install config + * @hostname: the desired hostname + * + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. + * + * Note that some operating systems have restrictions on maximum number of + * characters in hostname (or its equivalent) so the actual hostname set might + * be different from the value you specified here. For example if you specify + * 'my-not-so-long-hostname' as hostname for Microsoft Windows, the actual + * hostname will be 'my-not-so-long-' since Microsoft Window does not allow more + * than 15 characters in hostnames. + */ void osinfo_install_config_set_hostname(OsinfoInstallConfig *config, const gchar *hostname) { -- 1.9.0 From zeeshanak at gnome.org Mon May 12 19:29:36 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 20:29:36 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, installer: 'EN-GB' not supported as UI language Message-ID: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> Apparently 'EN-GB' is not supported as UI language but only as locale by win7 installer. Lets override 'EN-GB' to 'EN-US' for UI language so unattended installation doesn't break on EN-GB locale. http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx --- data/install-scripts/windows-unattend.xml | 30 ++++++++++++++++++++++++++---- 1 file changed, 26 insertions(+), 4 deletions(-) diff --git a/data/install-scripts/windows-unattend.xml b/data/install-scripts/windows-unattend.xml index f3cb7f8..f3030d2 100644 --- a/data/install-scripts/windows-unattend.xml +++ b/data/install-scripts/windows-unattend.xml @@ -69,6 +69,17 @@ + + + + EN-US + + + + + + + @@ -132,10 +143,10 @@ - + - + @@ -305,6 +316,17 @@ + + + + EN-US + + + + + + + @@ -368,10 +390,10 @@ - + - + -- 1.9.0 From zeeshanak at gnome.org Mon May 12 19:29:37 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Mon, 12 May 2014 20:29:37 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> Message-ID: <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx --- data/datamaps/windows-lang.xml | 1 + 1 file changed, 1 insertion(+) diff --git a/data/datamaps/windows-lang.xml b/data/datamaps/windows-lang.xml index 6b3efe1..96a7baf 100644 --- a/data/datamaps/windows-lang.xml +++ b/data/datamaps/windows-lang.xml @@ -104,6 +104,7 @@ + -- 1.9.0 From cfergeau at redhat.com Tue May 13 08:05:08 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 13 May 2014 10:05:08 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> Message-ID: <20140513080508.GB19281@edamame.cdg.redhat.com> On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: > Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language > or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of > 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. Are you 100% sure about that? Some web searches gave http://www.msfn.org/board/topic/149673-unattendxml-fr-ca-0c0c00001009/ for example, but it's unclear whether this can work for unmodified ISOs. Christophe From cfergeau at redhat.com Tue May 13 08:27:36 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 13 May 2014 10:27:36 +0200 Subject: [Libosinfo] [PATCH 1/2] win7, installer: 'EN-GB' not supported as UI language In-Reply-To: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140513082736.GC19281@edamame.cdg.redhat.com> On Mon, May 12, 2014 at 08:29:36PM +0100, Zeeshan Ali (Khattak) wrote: > Apparently 'EN-GB' is not supported as UI language but only as locale by > win7 installer. Lets override 'EN-GB' to 'EN-US' for UI language so > unattended installation doesn't break on EN-GB locale. > > http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ > http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx ACK, though if more special casing is needed, I'd tend to add another datamap for that. Christophe From cfergeau at redhat.com Tue May 13 08:30:01 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 13 May 2014 10:30:01 +0200 Subject: [Libosinfo] [PATCH 2/2] install-config: Document implicit hostname changes In-Reply-To: <1399909115-32661-2-git-send-email-zeeshanak@gnome.org> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <1399909115-32661-2-git-send-email-zeeshanak@gnome.org> Message-ID: <20140513083001.GD19281@edamame.cdg.redhat.com> On Mon, May 12, 2014 at 04:38:35PM +0100, Zeeshan Ali (Khattak) wrote: > Since we need to chop the hostname for windows, lets warn the app > developers about this in the API documentation. Do we want to make the limitation programmatically queriable? Silently truncating the hostname in some cases feels a bit unnice. ACK to the doc change. Christophe > --- > osinfo/osinfo_install_config.c | 14 ++++++++++++++ > 1 file changed, 14 insertions(+) > > diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c > index de4ee42..c7d41d0 100644 > --- a/osinfo/osinfo_install_config.c > +++ b/osinfo/osinfo_install_config.c > @@ -343,6 +343,20 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf > OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); > } > > +/** > + * osinfo_install_config_set_hostname: > + * @config: the install config > + * @hostname: the desired hostname > + * > + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. > + * > + * Note that some operating systems have restrictions on maximum number of > + * characters in hostname (or its equivalent) so the actual hostname set might > + * be different from the value you specified here. For example if you specify > + * 'my-not-so-long-hostname' as hostname for Microsoft Windows, the actual > + * hostname will be 'my-not-so-long-' since Microsoft Window does not allow more > + * than 15 characters in hostnames. > + */ > void osinfo_install_config_set_hostname(OsinfoInstallConfig *config, > const gchar *hostname) > { > -- > 1.9.0 > > _______________________________________________ > Libosinfo mailing list > Libosinfo at redhat.com > https://www.redhat.com/mailman/listinfo/libosinfo From cfergeau at redhat.com Tue May 13 09:20:14 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 13 May 2014 11:20:14 +0200 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140513092014.GE19281@edamame.cdg.redhat.com> On Mon, May 12, 2014 at 04:38:34PM +0100, Zeeshan Ali (Khattak) wrote: > Since 'ComputerName' is supposed to be max 15 characters, unattended > installation breaks if this is set to a string longer than that. Lets crop the > string to ensure that doesn't happen. > > http://technet.microsoft.com/en-us/library/ff715676.aspx This link also lists forbidden characters, I don't think this is enforced? I'm also not sure substring will do the right thing with non-ASCII characters, which are potentially encoded as multiple bytes (the limit is 15 bytes, not 15 chars). Christophe From zeeshanak at gnome.org Tue May 13 11:29:32 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Tue, 13 May 2014 12:29:32 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, installer: 'EN-GB' not supported as UI language In-Reply-To: <20140513082736.GC19281@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <20140513082736.GC19281@edamame.cdg.redhat.com> Message-ID: On Tue, May 13, 2014 at 9:27 AM, Christophe Fergeau wrote: > On Mon, May 12, 2014 at 08:29:36PM +0100, Zeeshan Ali (Khattak) wrote: >> Apparently 'EN-GB' is not supported as UI language but only as locale by >> win7 installer. Lets override 'EN-GB' to 'EN-US' for UI language so >> unattended installation doesn't break on EN-GB locale. >> >> http://www.wincert.net/forum/topic/9550-add-en-gb-language-and-kb-to-autounattend/ >> http://technet.microsoft.com/library/dd744369%28WS.10%29.aspx > > ACK, though if more special casing is needed, I'd tend to add another > datamap for that. Agreed but I would have gone for that approach already for this but each config param has one datamap attached to it so we'll need to introduce a different parameter for locale with that approach. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Tue May 13 11:31:10 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Tue, 13 May 2014 12:31:10 +0100 Subject: [Libosinfo] [PATCH 2/2] install-config: Document implicit hostname changes In-Reply-To: <20140513083001.GD19281@edamame.cdg.redhat.com> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <1399909115-32661-2-git-send-email-zeeshanak@gnome.org> <20140513083001.GD19281@edamame.cdg.redhat.com> Message-ID: On Tue, May 13, 2014 at 9:30 AM, Christophe Fergeau wrote: > On Mon, May 12, 2014 at 04:38:35PM +0100, Zeeshan Ali (Khattak) wrote: >> Since we need to chop the hostname for windows, lets warn the app >> developers about this in the API documentation. > > Do we want to make the limitation programmatically queriable? Silently > truncating the hostname in some cases feels a bit unnice. If we can come up with a not-so-complicated API, why not? -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Tue May 13 11:50:28 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Tue, 13 May 2014 12:50:28 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: <20140513092014.GE19281@edamame.cdg.redhat.com> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> Message-ID: On Tue, May 13, 2014 at 10:20 AM, Christophe Fergeau wrote: > On Mon, May 12, 2014 at 04:38:34PM +0100, Zeeshan Ali (Khattak) wrote: >> Since 'ComputerName' is supposed to be max 15 characters, unattended >> installation breaks if this is set to a string longer than that. Lets crop the >> string to ensure that doesn't happen. >> >> http://technet.microsoft.com/en-us/library/ff715676.aspx > > This link also lists forbidden characters, I don't think this is > enforced? No its not. I didn't bother to enforce that (read below). > I'm also not sure substring will do the right thing with > non-ASCII characters, which are potentially encoded as multiple bytes > (the limit is 15 bytes, not 15 chars). That is true. Either I go work on my XSLT-foo and enforce this or we can just document this in set_hostname. Reading this, I'd go for the latter as non-ASCII won't be acceptable on other OSs either: https://en.wikipedia.org/wiki/Hostname#Restrictions_on_valid_host_names -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Tue May 13 11:57:44 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Tue, 13 May 2014 12:57:44 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140513080508.GB19281@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> Message-ID: On Tue, May 13, 2014 at 9:05 AM, Christophe Fergeau wrote: > On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: >> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language >> or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of >> 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. > > Are you 100% sure about that? No but I first tried to the do the same as with EN-GB: modify FR-CA to FR-FR for *UILanguage but express installation broke. It also breaks with FR-CA as both language and locale. > Some web searches gave > http://www.msfn.org/board/topic/149673-unattendxml-fr-ca-0c0c00001009/ > for example, but it's unclear whether this can work for unmodified ISOs. Maybe it works for some ISOs but unless we have some means of differentiating, I'd rather we go for an approach that works for all ISOs. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Wed May 14 14:06:14 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:06:14 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> Message-ID: <20140514140614.GK29308@edamame.cdg.redhat.com> On Tue, May 13, 2014 at 12:57:44PM +0100, Zeeshan Ali (Khattak) wrote: > On Tue, May 13, 2014 at 9:05 AM, Christophe Fergeau wrote: > > On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: > >> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language > >> or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of > >> 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. > > > > Are you 100% sure about that? > > No but I first tried to the do the same as with EN-GB: modify FR-CA to > FR-FR for *UILanguage but express installation broke. It also breaks > with FR-CA as both language and locale. I'm actually confused as to where you got your FR-CA from. Only the values listed in data/datamaps/windows-lang.xml are valid, we are not going to be able to list all variants in that file. Just for 'fr', I get fr_LU, fr_BE and fr_CH in addition to fr_CA and fr_FR. I expect these are an issue as well? Or am I missing something? Christophe From cfergeau at redhat.com Wed May 14 14:09:52 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:09:52 +0200 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> Message-ID: <20140514140952.GL29308@edamame.cdg.redhat.com> On Tue, May 13, 2014 at 12:50:28PM +0100, Zeeshan Ali (Khattak) wrote: > > I'm also not sure substring will do the right thing with > > non-ASCII characters, which are potentially encoded as multiple bytes > > (the limit is 15 bytes, not 15 chars). > > That is true. Either I go work on my XSLT-foo and enforce this or we > can just document this in set_hostname. Reading this, I'd go for the > latter as non-ASCII won't be acceptable on other OSs either: > > https://en.wikipedia.org/wiki/Hostname#Restrictions_on_valid_host_names Having some function which would be able to validate install script parameters, and report an error if the parameter does not fit the constraints for the current install script would probably be very useful. I'm not exactly sure how this could be done though :( Christophe From zeeshanak at gnome.org Wed May 14 14:14:04 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:14:04 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140514140614.GK29308@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 3:06 PM, Christophe Fergeau wrote: > On Tue, May 13, 2014 at 12:57:44PM +0100, Zeeshan Ali (Khattak) wrote: >> On Tue, May 13, 2014 at 9:05 AM, Christophe Fergeau wrote: >> > On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: >> >> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language >> >> or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of >> >> 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. >> > >> > Are you 100% sure about that? >> >> No but I first tried to the do the same as with EN-GB: modify FR-CA to >> FR-FR for *UILanguage but express installation broke. It also breaks >> with FR-CA as both language and locale. > > I'm actually confused as to where you got your FR-CA from. Only the > values listed in data/datamaps/windows-lang.xml are valid, Without this patch, you never get 'FR-CA' but rather fr_CA as its not mapped for lack of a mapping. > we are not > going to be able to list all variants in that file. I'm afraid we'll have to (at least try) if we want the express installation not to break. > Just for 'fr', I get > fr_LU, fr_BE and fr_CH in addition to fr_CA and fr_FR. I expect these > are an issue as well? Or am I missing something? I haven't tested those but I'm sure they suffer the same fate. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 14 14:17:55 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:17:55 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: <20140514140952.GL29308@edamame.cdg.redhat.com> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> <20140514140952.GL29308@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 3:09 PM, Christophe Fergeau wrote: > On Tue, May 13, 2014 at 12:50:28PM +0100, Zeeshan Ali (Khattak) wrote: >> > I'm also not sure substring will do the right thing with >> > non-ASCII characters, which are potentially encoded as multiple bytes >> > (the limit is 15 bytes, not 15 chars). >> >> That is true. Either I go work on my XSLT-foo and enforce this or we >> can just document this in set_hostname. Reading this, I'd go for the >> latter as non-ASCII won't be acceptable on other OSs either: >> >> https://en.wikipedia.org/wiki/Hostname#Restrictions_on_valid_host_names > > Having some function which would be able to validate install script > parameters, and report an error if the parameter does not fit the > constraints for the current install script would probably be very > useful. I'm not exactly sure how this could be done though :( Sure but me neither. In the meantime, can we go with this patch and documenting limitation set_hostname? -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 14 14:19:56 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:19:56 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> <20140514140952.GL29308@edamame.cdg.redhat.com> Message-ID: Totally forgot to mention that this patch fixes win7 express installation in Boxes as Boxes now a days adds name of variant to hostname. On Wed, May 14, 2014 at 3:17 PM, Zeeshan Ali (Khattak) wrote: > On Wed, May 14, 2014 at 3:09 PM, Christophe Fergeau wrote: >> On Tue, May 13, 2014 at 12:50:28PM +0100, Zeeshan Ali (Khattak) wrote: >>> > I'm also not sure substring will do the right thing with >>> > non-ASCII characters, which are potentially encoded as multiple bytes >>> > (the limit is 15 bytes, not 15 chars). >>> >>> That is true. Either I go work on my XSLT-foo and enforce this or we >>> can just document this in set_hostname. Reading this, I'd go for the >>> latter as non-ASCII won't be acceptable on other OSs either: >>> >>> https://en.wikipedia.org/wiki/Hostname#Restrictions_on_valid_host_names >> >> Having some function which would be able to validate install script >> parameters, and report an error if the parameter does not fit the >> constraints for the current install script would probably be very >> useful. I'm not exactly sure how this could be done though :( > > Sure but me neither. In the meantime, can we go with this patch and > documenting limitation set_hostname? > > -- > Regards, > > Zeeshan Ali (Khattak) > ________________________________________ > Befriend GNOME: http://www.gnome.org/friends/ -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Wed May 14 14:24:52 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:24:52 +0200 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> <20140514140952.GL29308@edamame.cdg.redhat.com> Message-ID: <20140514142452.GM29308@edamame.cdg.redhat.com> On Wed, May 14, 2014 at 03:19:56PM +0100, Zeeshan Ali (Khattak) wrote: > Totally forgot to mention that this patch fixes win7 express > installation in Boxes as Boxes now a days adds name of variant to > hostname. If you document set_hostname as expecting a <= 15 chars parameter, Boxes is broken, not libosinfo... Christophe From cfergeau at redhat.com Wed May 14 14:27:03 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:27:03 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> Message-ID: <20140514142703.GN29308@edamame.cdg.redhat.com> On Tue, May 13, 2014 at 12:57:44PM +0100, Zeeshan Ali (Khattak) wrote: > On Tue, May 13, 2014 at 9:05 AM, Christophe Fergeau wrote: > > On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: > >> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language > >> or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of > >> 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. > > > > Are you 100% sure about that? > > No but I first tried to the do the same as with EN-GB: modify FR-CA to > FR-FR for *UILanguage but express installation broke. It also breaks > with FR-CA as both language and locale. With a non-unattended French Windows install, "Language to install" only lists "France", but "time/currency format" and "keyboard/input method" both have a "French (Canada)" option, so it must be possible to specify this with an unattended install as well. Christophe From zeeshanak at gnome.org Wed May 14 14:42:02 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:42:02 +0100 Subject: [Libosinfo] [PATCH 1/2] win7, winxp, installer: 'ComputerName' must be <= 15 characters In-Reply-To: <20140514142452.GM29308@edamame.cdg.redhat.com> References: <1399909115-32661-1-git-send-email-zeeshanak@gnome.org> <20140513092014.GE19281@edamame.cdg.redhat.com> <20140514140952.GL29308@edamame.cdg.redhat.com> <20140514142452.GM29308@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 3:24 PM, Christophe Fergeau wrote: > On Wed, May 14, 2014 at 03:19:56PM +0100, Zeeshan Ali (Khattak) wrote: >> Totally forgot to mention that this patch fixes win7 express >> installation in Boxes as Boxes now a days adds name of variant to >> hostname. > > If you document set_hostname as expecting a <= 15 chars parameter, Boxes > is broken, not libosinfo... Hmm.. you probably have a point. I imagine there could be usecases for apps to rely on the set hostname. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Wed May 14 14:46:57 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:46:57 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> Message-ID: <20140514144657.GA3061@edamame.cdg.redhat.com> On Wed, May 14, 2014 at 03:14:04PM +0100, Zeeshan Ali (Khattak) wrote: > On Wed, May 14, 2014 at 3:06 PM, Christophe Fergeau wrote: > > I'm actually confused as to where you got your FR-CA from. Only the > > values listed in data/datamaps/windows-lang.xml are valid, > > Without this patch, you never get 'FR-CA' but rather fr_CA as its not > mapped for lack of a mapping. Actually, see osinfo_media_get_languages() documentation: * If media is an installer, this property indicates the languages that * can be used during automatic installations. I don't think fr_CA is ever returned by that method, so that would make it an invalid value to use as l10n-language. Boxes seems to be trying hard to return a good value from UnattendedInstaller::get_preferred_language(), so I'm not sure how we get to the point where fr_CA gets passed to an unattended install script? One thing I noticed in get_preferred_language() is that it should probably strip ".UTF-8" when trying to match system locale with media languages, but there is still a fallback in place to return one of the media languages if there is no match. Christophe From zeeshanak at gnome.org Wed May 14 14:48:52 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:48:52 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140514142703.GN29308@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514142703.GN29308@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 3:27 PM, Christophe Fergeau wrote: > On Tue, May 13, 2014 at 12:57:44PM +0100, Zeeshan Ali (Khattak) wrote: >> On Tue, May 13, 2014 at 9:05 AM, Christophe Fergeau wrote: >> > On Mon, May 12, 2014 at 08:29:37PM +0100, Zeeshan Ali (Khattak) wrote: >> >> Unlike 'EN-GB', 'FR-CA' is not supported at all, neither as UI language >> >> or locale by win7 installer so lets map 'fr_CA' to 'FR-FR' instead of >> >> 'FR-CA' so unattended installation doesn't break on 'fr_CA' locale. >> > >> > Are you 100% sure about that? >> >> No but I first tried to the do the same as with EN-GB: modify FR-CA to >> FR-FR for *UILanguage but express installation broke. It also breaks >> with FR-CA as both language and locale. > > With a non-unattended French Windows install, "Language to install" only lists "France", but "time/currency format" and "keyboard/input method" both have a "French (Canada)" option, so it must be possible to specify this with an unattended install as well. Come to think of it, I tested this against EN ISO so you might be right. I feel too lazy to download ISOs of all locales and testing so I'll leave this issue (if there is one) to French-speaking people (Do you know anyone? ;) I think Boxes should choose the first supported lang on the media (if media has any lang associated) rather than system lang when it fails to match the lang from system and media. Wouldn't you agree? -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 14 14:59:32 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 15:59:32 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140514144657.GA3061@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> <20140514144657.GA3061@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 3:46 PM, Christophe Fergeau wrote: > On Wed, May 14, 2014 at 03:14:04PM +0100, Zeeshan Ali (Khattak) wrote: >> On Wed, May 14, 2014 at 3:06 PM, Christophe Fergeau wrote: >> > I'm actually confused as to where you got your FR-CA from. Only the >> > values listed in data/datamaps/windows-lang.xml are valid, >> >> Without this patch, you never get 'FR-CA' but rather fr_CA as its not >> mapped for lack of a mapping. > > Actually, see osinfo_media_get_languages() documentation: > > * If media is an installer, this property indicates the languages that > * can be used during automatic installations. > > I don't think fr_CA is ever returned by that method, so that would make it > an invalid value to use as l10n-language. Ah that answers my question to you in my previous mail. :) > Boxes seems to be trying hard > to return a good value from UnattendedInstaller::get_preferred_language(), > so I'm not sure how we get to the point where fr_CA gets passed to an > unattended install script? > One thing I noticed in get_preferred_language() is that it should probably strip ".UTF-8" when trying to match system locale with media languages, but there is still a fallback in place to return one of the media languages if there is no match. Actually, i hardcoded 'fr_CA' to be passed as I didn't think hard enough that its not going to actually happen. Sorry about the false alarm. :( Thinking about this, I wonder why en_GB gets passed. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Wed May 14 14:59:36 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 16:59:36 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514142703.GN29308@edamame.cdg.redhat.com> Message-ID: <20140514145936.GA5965@edamame.cdg.redhat.com> On Wed, May 14, 2014 at 03:48:52PM +0100, Zeeshan Ali (Khattak) wrote: > I think Boxes should choose the first supported lang on the media (if > media has any lang associated) rather than system lang when it fails > to match the lang from system and media. Wouldn't you agree? My understanding of UnattendedInstaller::get_preferred_language() is that it's already doing that: private string get_preferred_language () { var system_langs = Intl.get_language_names (); var media_langs = new HashTable (str_hash, str_equal); var media_langs_list = os_media.languages; foreach (var lang in media_langs_list) media_langs.add (lang); foreach (var lang in system_langs) { ==> here we should probably strip .UTF-8 from lang as is done below if (lang in media_langs) { debug ("matched %s", lang); return lang; } } ==> media language fallback below if (media_langs_list != null) { debug ("Failed to match system locales with media languages, falling back to %s media language", media_langs_list.nth_data (0)); return media_langs_list.nth_data (0); } ==> system locale fallback below var lang = system_langs[0].replace (".utf8", ""); lang = lang.replace (".UTF-8", ""); debug ("No media language, using %s locale", lang); return lang; } From cfergeau at redhat.com Wed May 14 17:07:32 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 14 May 2014 19:07:32 +0200 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> <20140514144657.GA3061@edamame.cdg.redhat.com> Message-ID: <20140514170732.GB6772@edamame.cdg.redhat.com> On Wed, May 14, 2014 at 03:59:32PM +0100, Zeeshan Ali (Khattak) wrote: > Actually, i hardcoded 'fr_CA' to be passed as I didn't think hard > enough that its not going to actually happen. Sorry about the false > alarm. :( Thinking about this, I wonder why en_GB gets passed. There are some en-gb win7 ISOs, this probably comes from these. And your fr_CA experiments are another indication that noone reads API doc about valid values, and that it would be much easier if the install scripts parameters could be validated by libosinfo before being used ;) Imo it's a good experiment with what users of libosinfo could go through :( Christophe From zeeshanak at gnome.org Wed May 14 17:51:16 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 18:51:16 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140514170732.GB6772@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> <20140514144657.GA3061@edamame.cdg.redhat.com> <20140514170732.GB6772@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 6:07 PM, Christophe Fergeau wrote: > On Wed, May 14, 2014 at 03:59:32PM +0100, Zeeshan Ali (Khattak) wrote: >> Actually, i hardcoded 'fr_CA' to be passed as I didn't think hard >> enough that its not going to actually happen. Sorry about the false >> alarm. :( Thinking about this, I wonder why en_GB gets passed. > > There are some en-gb win7 ISOs, this probably comes from these. And your > fr_CA experiments are another indication that noone reads API doc about > valid values, and that it would be much easier if the install scripts > parameters could be validated by libosinfo before being used ;) Imo it's > a good experiment with what users of libosinfo could go through :( There is a GError parameter to osinfo_install_script_generate* so we could check the params in there but that easy for generic stuff (e.g hostname) but in this case, we'd need some new xml at least (if not, API). -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 14 19:28:39 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 20:28:39 +0100 Subject: [Libosinfo] [PATCH 2/2] docs: Add InstallScript/Config to docs In-Reply-To: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> References: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> Message-ID: <1400095719-31883-2-git-send-email-zeeshanak@gnome.org> --- docs/reference/Libosinfo-docs.xml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml index 74b3f5f..6b6886a 100644 --- a/docs/reference/Libosinfo-docs.xml +++ b/docs/reference/Libosinfo-docs.xml @@ -25,6 +25,8 @@ + + -- 1.9.0 From zeeshanak at gnome.org Wed May 14 19:28:38 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 20:28:38 +0100 Subject: [Libosinfo] [PATCH 1/2] install-config: Document hostname limitations Message-ID: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> Since Microsoft Windows has several limitations on allowed characters and number of characters in ComputerName (which is what hostname is translated to), lets inform the app developers about these limitations. http://technet.microsoft.com/en-us/library/ff715676.aspx --- osinfo/osinfo_install_config.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c index de4ee42..4fc8b35 100644 --- a/osinfo/osinfo_install_config.c +++ b/osinfo/osinfo_install_config.c @@ -343,6 +343,22 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); } +/** + * osinfo_install_config_set_hostname: + * @config: the install config + * @hostname: the desired hostname + * + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. + * + * Note that some operating systems have restrictions on maximum number of, and + * allowed characters in hostname (or its equivalent) so it is highly + * recommended that you keep the actual hostname less than or equal to 15 + * characters long and does not contain non-ASCII, spaces or any of the + * following characters to ensure the generated install script using @config + * works as expected: + * + * { | } ~ [ \ ] ^ ' : ; < = > ? @ ! " # $ % ` ( ) + / . , * & + */ void osinfo_install_config_set_hostname(OsinfoInstallConfig *config, const gchar *hostname) { -- 1.9.0 From zeeshanak at gnome.org Wed May 14 20:05:10 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 14 May 2014 21:05:10 +0100 Subject: [Libosinfo] [PATCH 2/2] win7,installer: 'FR-CA' not supported In-Reply-To: <20140514170732.GB6772@edamame.cdg.redhat.com> References: <1399922977-29553-1-git-send-email-zeeshanak@gnome.org> <1399922977-29553-2-git-send-email-zeeshanak@gnome.org> <20140513080508.GB19281@edamame.cdg.redhat.com> <20140514140614.GK29308@edamame.cdg.redhat.com> <20140514144657.GA3061@edamame.cdg.redhat.com> <20140514170732.GB6772@edamame.cdg.redhat.com> Message-ID: On Wed, May 14, 2014 at 6:07 PM, Christophe Fergeau wrote: > On Wed, May 14, 2014 at 03:59:32PM +0100, Zeeshan Ali (Khattak) wrote: >> Actually, i hardcoded 'fr_CA' to be passed as I didn't think hard >> enough that its not going to actually happen. Sorry about the false >> alarm. :( Thinking about this, I wonder why en_GB gets passed. > > There are some en-gb win7 ISOs, this probably comes from these. Actually I don't have one of those ISOs but now I recall why I get en_GB: It was against a media we get from Tracker and until yesterday, Tracker didn't extract language info and Boxes still doesn't fetch it from Tracker[1] so what happened was that Boxes assumed media has no languages associated and fell back to system lang. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ [1] I'm using this part now but its a bit tricky since just adding the parameter to our sparql query will break Boxes against older Tracker. From gscrivan at redhat.com Fri May 16 07:18:19 2014 From: gscrivan at redhat.com (Giuseppe Scrivano) Date: Fri, 16 May 2014 09:18:19 +0200 Subject: [Libosinfo] [PATCH] openbsd: remove reference to missing version Message-ID: <1400224699-8148-1-git-send-email-gscrivan@redhat.com> OpenBSD 5.3 is not present in the DB, so remove a reference to it. Signed-off-by: Giuseppe Scrivano --- data/oses/openbsd.xml.in | 2 -- 1 file changed, 2 deletions(-) diff --git a/data/oses/openbsd.xml.in b/data/oses/openbsd.xml.in index 062eb47..bf97a44 100644 --- a/data/oses/openbsd.xml.in +++ b/data/oses/openbsd.xml.in @@ -214,8 +214,6 @@ openbsd openbsd <_vendor>OpenBSD Project - - 2013-11-01 -- 1.9.0 From cfergeau at redhat.com Fri May 16 08:06:51 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Fri, 16 May 2014 10:06:51 +0200 Subject: [Libosinfo] [PATCH] openbsd: remove reference to missing version In-Reply-To: <1400224699-8148-1-git-send-email-gscrivan@redhat.com> References: <1400224699-8148-1-git-send-email-gscrivan@redhat.com> Message-ID: <20140516080651.GA5244@edamame.cdg.redhat.com> On Fri, May 16, 2014 at 09:18:19AM +0200, Giuseppe Scrivano wrote: > OpenBSD 5.3 is not present in the DB, so remove a reference to it. > > Signed-off-by: Giuseppe Scrivano > --- > data/oses/openbsd.xml.in | 2 -- > 1 file changed, 2 deletions(-) > > diff --git a/data/oses/openbsd.xml.in b/data/oses/openbsd.xml.in > index 062eb47..bf97a44 100644 > --- a/data/oses/openbsd.xml.in > +++ b/data/oses/openbsd.xml.in > @@ -214,8 +214,6 @@ > openbsd > openbsd > <_vendor>OpenBSD Project > - > - It would be better to add it ;) At least a stub entry without volume ids c&p'ed from one of the other openbsd entries. Christophe From gscrivan at redhat.com Fri May 16 08:50:23 2014 From: gscrivan at redhat.com (Giuseppe Scrivano) Date: Fri, 16 May 2014 10:50:23 +0200 Subject: [Libosinfo] [PATCH] openbsd: add missing 5.* releases In-Reply-To: <20140516080651.GA5244@edamame.cdg.redhat.com> References: <20140516080651.GA5244@edamame.cdg.redhat.com> Message-ID: <1400230223-12069-1-git-send-email-gscrivan@redhat.com> Signed-off-by: Giuseppe Scrivano --- I've added the missing versions, I took the release date from: http://www.openbsd.org/5[1-3].html data/oses/openbsd.xml.in | 93 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 93 insertions(+) diff --git a/data/oses/openbsd.xml.in b/data/oses/openbsd.xml.in index 062eb47..d9f88ea 100644 --- a/data/oses/openbsd.xml.in +++ b/data/oses/openbsd.xml.in @@ -207,6 +207,99 @@ + + openbsd5.1 + <_name>OpenBSD 5.1 + 5.1 + openbsd + openbsd + <_vendor>OpenBSD Project + + + + 2012-05-01 + + + + The OpenBSD project + OpenBSD + OpenBSD/i386 5.1 Install CD + OpenBSD 5.1 i386 Install CD + + + + + + The OpenBSD project + OpenBSD + OpenBSD/amd64 5.1 Install CD + OpenBSD 5.1 amd64 Install CD + + + + + + openbsd5.2 + <_name>OpenBSD 5.2 + 5.2 + openbsd + openbsd + <_vendor>OpenBSD Project + + + + 2012-11-01 + + + + The OpenBSD project + OpenBSD + OpenBSD/i386 5.2 Install CD + OpenBSD 5.2 i386 Install CD + + + + + + The OpenBSD project + OpenBSD + OpenBSD/amd64 5.2 Install CD + OpenBSD 5.2 amd64 Install CD + + + + + + openbsd5.3 + <_name>OpenBSD 5.3 + 5.3 + openbsd + openbsd + <_vendor>OpenBSD Project + + + + 2013-05-01 + + + + The OpenBSD project + OpenBSD + OpenBSD/i386 5.3 Install CD + OpenBSD 5.3 i386 Install CD + + + + + + The OpenBSD project + OpenBSD + OpenBSD/amd64 5.3 Install CD + OpenBSD 5.3 amd64 Install CD + + + + openbsd5.4 <_name>OpenBSD 5.4 -- 1.9.0 From cfergeau at redhat.com Fri May 16 09:31:43 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Fri, 16 May 2014 11:31:43 +0200 Subject: [Libosinfo] [PATCH] openbsd: add missing 5.* releases In-Reply-To: <1400230223-12069-1-git-send-email-gscrivan@redhat.com> References: <20140516080651.GA5244@edamame.cdg.redhat.com> <1400230223-12069-1-git-send-email-gscrivan@redhat.com> Message-ID: <20140516093143.GB5244@edamame.cdg.redhat.com> On Fri, May 16, 2014 at 10:50:23AM +0200, Giuseppe Scrivano wrote: > Signed-off-by: Giuseppe Scrivano > --- > I've added the missing versions, I took the release date from: > > http://www.openbsd.org/5[1-3].html Great, thanks for the patch! Thanks for going the extra mile and adding the release dates. ACK. Christophe From gscrivan at redhat.com Fri May 16 09:59:40 2014 From: gscrivan at redhat.com (Giuseppe Scrivano) Date: Fri, 16 May 2014 11:59:40 +0200 Subject: [Libosinfo] [PATCH] openbsd: add missing 5.* releases In-Reply-To: <20140516093143.GB5244@edamame.cdg.redhat.com> (Christophe Fergeau's message of "Fri, 16 May 2014 11:31:43 +0200") References: <20140516080651.GA5244@edamame.cdg.redhat.com> <1400230223-12069-1-git-send-email-gscrivan@redhat.com> <20140516093143.GB5244@edamame.cdg.redhat.com> Message-ID: <87oayyavsj.fsf@redhat.com> Christophe Fergeau writes: > On Fri, May 16, 2014 at 10:50:23AM +0200, Giuseppe Scrivano wrote: >> Signed-off-by: Giuseppe Scrivano >> --- >> I've added the missing versions, I took the release date from: >> >> http://www.openbsd.org/5[1-3].html > > Great, thanks for the patch! Thanks for going the extra mile and adding the release dates. ACK. thanks for the review, pushed now. Giuseppe From cfergeau at redhat.com Tue May 20 13:14:12 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 20 May 2014 15:14:12 +0200 Subject: [Libosinfo] [PATCH 1/2] install-config: Document hostname limitations In-Reply-To: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> References: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140520131412.GA23115@edamame.cdg.redhat.com> On Wed, May 14, 2014 at 08:28:38PM +0100, Zeeshan Ali (Khattak) wrote: > Since Microsoft Windows has several limitations on allowed characters and > number of characters in ComputerName (which is what hostname is > translated to), lets inform the app developers about these limitations. > > http://technet.microsoft.com/en-us/library/ff715676.aspx > --- > osinfo/osinfo_install_config.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c > index de4ee42..4fc8b35 100644 > --- a/osinfo/osinfo_install_config.c > +++ b/osinfo/osinfo_install_config.c > @@ -343,6 +343,22 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf > OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); > } > > +/** > + * osinfo_install_config_set_hostname: > + * @config: the install config > + * @hostname: the desired hostname > + * > + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. > + * > + * Note that some operating systems have restrictions on maximum number of, and > + * allowed characters in hostname (or its equivalent) so it is highly > + * recommended that you keep the actual hostname less than or equal to 15 > + * characters long and does not contain non-ASCII, spaces or any of the "and that it does not contain" > + * following characters to ensure the generated install script using @config > + * works as expected: > + * > + * { | } ~ [ \ ] ^ ' : ; < = > ? @ ! " # $ % ` ( ) + / . , * & Maybe we should just recommend to use alphanumeric characters and '-' instead of a blacklist? ACK with the first comment addressed. Christophe From zeeshanak at gnome.org Tue May 20 13:26:37 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Tue, 20 May 2014 14:26:37 +0100 Subject: [Libosinfo] [PATCH 1/2] install-config: Document hostname limitations In-Reply-To: <20140520131412.GA23115@edamame.cdg.redhat.com> References: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> <20140520131412.GA23115@edamame.cdg.redhat.com> Message-ID: On Tue, May 20, 2014 at 2:14 PM, Christophe Fergeau wrote: > On Wed, May 14, 2014 at 08:28:38PM +0100, Zeeshan Ali (Khattak) wrote: >> Since Microsoft Windows has several limitations on allowed characters and >> number of characters in ComputerName (which is what hostname is >> translated to), lets inform the app developers about these limitations. >> >> http://technet.microsoft.com/en-us/library/ff715676.aspx >> --- >> osinfo/osinfo_install_config.c | 16 ++++++++++++++++ >> 1 file changed, 16 insertions(+) >> >> diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c >> index de4ee42..4fc8b35 100644 >> --- a/osinfo/osinfo_install_config.c >> +++ b/osinfo/osinfo_install_config.c >> @@ -343,6 +343,22 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf >> OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); >> } >> >> +/** >> + * osinfo_install_config_set_hostname: >> + * @config: the install config >> + * @hostname: the desired hostname >> + * >> + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. >> + * >> + * Note that some operating systems have restrictions on maximum number of, and >> + * allowed characters in hostname (or its equivalent) so it is highly >> + * recommended that you keep the actual hostname less than or equal to 15 >> + * characters long and does not contain non-ASCII, spaces or any of the > > "and that it does not contain" > >> + * following characters to ensure the generated install script using @config >> + * works as expected: >> + * >> + * { | } ~ [ \ ] ^ ' : ; < = > ? @ ! " # $ % ` ( ) + / . , * & > > Maybe we should just recommend to use alphanumeric characters and '-' > instead of a blacklist? Agreed. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Tue May 20 13:30:44 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 20 May 2014 15:30:44 +0200 Subject: [Libosinfo] [PATCH 2/2] docs: Add InstallScript/Config to docs In-Reply-To: <1400095719-31883-2-git-send-email-zeeshanak@gnome.org> References: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> <1400095719-31883-2-git-send-email-zeeshanak@gnome.org> Message-ID: <20140520133044.GB23115@edamame.cdg.redhat.com> Is there some command which can be ran that lists C files which have doc comments but are not listed in -docs.xml? If we are not missing more files, ACK. Christophe On Wed, May 14, 2014 at 08:28:39PM +0100, Zeeshan Ali (Khattak) wrote: > --- > docs/reference/Libosinfo-docs.xml | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml > index 74b3f5f..6b6886a 100644 > --- a/docs/reference/Libosinfo-docs.xml > +++ b/docs/reference/Libosinfo-docs.xml > @@ -25,6 +25,8 @@ > > > > + > + > > > > -- > 1.9.0 > > _______________________________________________ > Libosinfo mailing list > Libosinfo at redhat.com > https://www.redhat.com/mailman/listinfo/libosinfo From zeeshanak at gnome.org Wed May 21 00:58:54 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 01:58:54 +0100 Subject: [Libosinfo] [PATCH 2/2] docs: Add InstallScript/Config to docs In-Reply-To: <20140520133044.GB23115@edamame.cdg.redhat.com> References: <1400095719-31883-1-git-send-email-zeeshanak@gnome.org> <1400095719-31883-2-git-send-email-zeeshanak@gnome.org> <20140520133044.GB23115@edamame.cdg.redhat.com> Message-ID: On Tue, May 20, 2014 at 2:30 PM, Christophe Fergeau wrote: > Is there some command which can be ran that lists C files which have > doc comments but are not listed in -docs.xml? Don't think so. I'll manually check if each file in docs/reference/xml/ is listed in the morning. :) -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 21 00:59:49 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 01:59:49 +0100 Subject: [Libosinfo] [PATCHv2] docs: Add InstallScript/Config to docs Message-ID: <1400633989-17752-1-git-send-email-zeeshanak@gnome.org> --- docs/reference/Libosinfo-docs.xml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml index 74b3f5f..6b6886a 100644 --- a/docs/reference/Libosinfo-docs.xml +++ b/docs/reference/Libosinfo-docs.xml @@ -25,6 +25,8 @@ + + -- 1.9.0 From cfergeau at redhat.com Wed May 21 12:03:02 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 21 May 2014 14:03:02 +0200 Subject: [Libosinfo] [PATCHv2] docs: Add InstallScript/Config to docs In-Reply-To: <1400633989-17752-1-git-send-email-zeeshanak@gnome.org> References: <1400633989-17752-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140521120302.GJ23115@edamame.cdg.redhat.com> Isn't this the same as v1 ? If it is, my ACK still stands. Christophe On Wed, May 21, 2014 at 01:59:49AM +0100, Zeeshan Ali (Khattak) wrote: > --- > docs/reference/Libosinfo-docs.xml | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml > index 74b3f5f..6b6886a 100644 > --- a/docs/reference/Libosinfo-docs.xml > +++ b/docs/reference/Libosinfo-docs.xml > @@ -25,6 +25,8 @@ > > > > + > + > > > > -- > 1.9.0 > > _______________________________________________ > Libosinfo mailing list > Libosinfo at redhat.com > https://www.redhat.com/mailman/listinfo/libosinfo From zeeshanak at gnome.org Wed May 21 12:05:22 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 13:05:22 +0100 Subject: [Libosinfo] [PATCHv2] docs: Add InstallScript/Config to docs In-Reply-To: <20140521120302.GJ23115@edamame.cdg.redhat.com> References: <1400633989-17752-1-git-send-email-zeeshanak@gnome.org> <20140521120302.GJ23115@edamame.cdg.redhat.com> Message-ID: On Wed, May 21, 2014 at 1:03 PM, Christophe Fergeau wrote: > Isn't this the same as v1 ? If it is, my ACK still stands. Oh, I sent the wrong patch. :) -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From zeeshanak at gnome.org Wed May 21 12:06:02 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 13:06:02 +0100 Subject: [Libosinfo] [PATCHv2] install-config: Document hostname limitations Message-ID: <1400673962-19905-1-git-send-email-zeeshanak@gnome.org> Since Microsoft Windows has several limitations on allowed characters and number of characters in ComputerName (which is what hostname is translated to), lets inform the app developers about these limitations. http://technet.microsoft.com/en-us/library/ff715676.aspx --- osinfo/osinfo_install_config.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c index de4ee42..af0f007 100644 --- a/osinfo/osinfo_install_config.c +++ b/osinfo/osinfo_install_config.c @@ -343,6 +343,19 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); } +/** + * osinfo_install_config_set_hostname: + * @config: the install config + * @hostname: the desired hostname + * + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. + * + * Note that some operating systems have restrictions on maximum number of, and + * allowed characters in hostname (or its equivalent) so it is highly + * recommended that you keep the actual hostname less than or equal to 15 + * characters long and make sure that it does not contain any characters other + * than ASCII alphanumeric and '-'. + */ void osinfo_install_config_set_hostname(OsinfoInstallConfig *config, const gchar *hostname) { -- 1.9.0 From cfergeau at redhat.com Wed May 21 12:19:57 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 21 May 2014 14:19:57 +0200 Subject: [Libosinfo] [PATCHv2] install-config: Document hostname limitations In-Reply-To: <1400673962-19905-1-git-send-email-zeeshanak@gnome.org> References: <1400673962-19905-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140521121957.GK23115@edamame.cdg.redhat.com> On Wed, May 21, 2014 at 01:06:02PM +0100, Zeeshan Ali (Khattak) wrote: > Since Microsoft Windows has several limitations on allowed characters and > number of characters in ComputerName (which is what hostname is > translated to), lets inform the app developers about these limitations. > > http://technet.microsoft.com/en-us/library/ff715676.aspx > --- > osinfo/osinfo_install_config.c | 13 +++++++++++++ > 1 file changed, 13 insertions(+) > > diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c > index de4ee42..af0f007 100644 > --- a/osinfo/osinfo_install_config.c > +++ b/osinfo/osinfo_install_config.c > @@ -343,6 +343,19 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf > OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); > } > > +/** > + * osinfo_install_config_set_hostname: > + * @config: the install config > + * @hostname: the desired hostname > + * > + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. > + * > + * Note that some operating systems have restrictions on maximum number of, and > + * allowed characters in hostname (or its equivalent) so it is highly > + * recommended that you keep the actual hostname less than or equal to 15 > + * characters long and make sure that it does not contain any characters other > + * than ASCII alphanumeric and '-'. I'd add "Automatic installation may fail if the OS has restrictions on this value and it contains invalid characters." Christophe From zeeshanak at gnome.org Wed May 21 12:57:38 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 13:57:38 +0100 Subject: [Libosinfo] [PATCHv2] install-config: Document hostname limitations In-Reply-To: <20140521121957.GK23115@edamame.cdg.redhat.com> References: <1400673962-19905-1-git-send-email-zeeshanak@gnome.org> <20140521121957.GK23115@edamame.cdg.redhat.com> Message-ID: On Wed, May 21, 2014 at 1:19 PM, Christophe Fergeau wrote: > On Wed, May 21, 2014 at 01:06:02PM +0100, Zeeshan Ali (Khattak) wrote: >> Since Microsoft Windows has several limitations on allowed characters and >> number of characters in ComputerName (which is what hostname is >> translated to), lets inform the app developers about these limitations. >> >> http://technet.microsoft.com/en-us/library/ff715676.aspx >> --- >> osinfo/osinfo_install_config.c | 13 +++++++++++++ >> 1 file changed, 13 insertions(+) >> >> diff --git a/osinfo/osinfo_install_config.c b/osinfo/osinfo_install_config.c >> index de4ee42..af0f007 100644 >> --- a/osinfo/osinfo_install_config.c >> +++ b/osinfo/osinfo_install_config.c >> @@ -343,6 +343,19 @@ const gchar *osinfo_install_config_get_reg_product_key(OsinfoInstallConfig *conf >> OSINFO_INSTALL_CONFIG_PROP_REG_PRODUCTKEY); >> } >> >> +/** >> + * osinfo_install_config_set_hostname: >> + * @config: the install config >> + * @hostname: the desired hostname >> + * >> + * Sets the #OSINFO_INSTALL_CONFIG_PROP_HOSTNAME parameter. >> + * >> + * Note that some operating systems have restrictions on maximum number of, and >> + * allowed characters in hostname (or its equivalent) so it is highly >> + * recommended that you keep the actual hostname less than or equal to 15 >> + * characters long and make sure that it does not contain any characters other >> + * than ASCII alphanumeric and '-'. > > I'd add > "Automatic installation may fail if the OS has restrictions on this value > and it contains invalid characters." How about I just append " Otherwise unattended installation might fail." to above? -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ From cfergeau at redhat.com Wed May 21 12:59:47 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Wed, 21 May 2014 14:59:47 +0200 Subject: [Libosinfo] [PATCHv2] install-config: Document hostname limitations In-Reply-To: References: <1400673962-19905-1-git-send-email-zeeshanak@gnome.org> <20140521121957.GK23115@edamame.cdg.redhat.com> Message-ID: <20140521125947.GM23115@edamame.cdg.redhat.com> On Wed, May 21, 2014 at 01:57:38PM +0100, Zeeshan Ali (Khattak) wrote: > On Wed, May 21, 2014 at 1:19 PM, Christophe Fergeau wrote: > > > > I'd add > > "Automatic installation may fail if the OS has restrictions on this value > > and it contains invalid characters." > > How about I just append " Otherwise unattended installation might > fail." to above? Sure, ACK. Christophe From zeeshanak at gnome.org Wed May 21 16:54:34 2014 From: zeeshanak at gnome.org (Zeeshan Ali (Khattak)) Date: Wed, 21 May 2014 17:54:34 +0100 Subject: [Libosinfo] [PATCH] docs: Add docs links from main docs page Message-ID: <1400691274-21396-1-git-send-email-zeeshanak@gnome.org> --- docs/reference/Libosinfo-docs.xml | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml index 74b3f5f..1a239a9 100644 --- a/docs/reference/Libosinfo-docs.xml +++ b/docs/reference/Libosinfo-docs.xml @@ -15,6 +15,9 @@ Libosinfo + + + @@ -23,12 +26,22 @@ + + + + + + + + + + @@ -38,6 +51,9 @@ + + + Object Hierarchy @@ -47,6 +63,10 @@ API Index + + Deprecated API Index + + -- 1.9.0 From cfergeau at redhat.com Thu May 22 07:14:02 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Thu, 22 May 2014 09:14:02 +0200 Subject: [Libosinfo] [PATCH] docs: Add docs links from main docs page In-Reply-To: <1400691274-21396-1-git-send-email-zeeshanak@gnome.org> References: <1400691274-21396-1-git-send-email-zeeshanak@gnome.org> Message-ID: <20140522071402.GN23115@edamame.cdg.redhat.com> ACK On Wed, May 21, 2014 at 05:54:34PM +0100, Zeeshan Ali (Khattak) wrote: > --- > docs/reference/Libosinfo-docs.xml | 20 ++++++++++++++++++++ > 1 file changed, 20 insertions(+) > > diff --git a/docs/reference/Libosinfo-docs.xml b/docs/reference/Libosinfo-docs.xml > index 74b3f5f..1a239a9 100644 > --- a/docs/reference/Libosinfo-docs.xml > +++ b/docs/reference/Libosinfo-docs.xml > @@ -15,6 +15,9 @@ > > > Libosinfo > + > + > + > > > > @@ -23,12 +26,22 @@ > > > > + > + > > + > > + > + > + > + > + > > > > > + > + > > > > @@ -38,6 +51,9 @@ > > > > + > + > + > > > Object Hierarchy > @@ -47,6 +63,10 @@ > API Index > > > + > + Deprecated API Index > + > + > > > > -- > 1.9.0 > > _______________________________________________ > Libosinfo mailing list > Libosinfo at redhat.com > https://www.redhat.com/mailman/listinfo/libosinfo From fidencio at redhat.com Sat May 24 00:49:39 2014 From: fidencio at redhat.com (=?UTF-8?q?Fabiano=20Fid=C3=AAncio?=) Date: Sat, 24 May 2014 02:49:39 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound Message-ID: <1400892579-7208-1-git-send-email-fidencio@redhat.com> osinfo_loader_nodeset() can return -1 without set an error Caught by: https://scan.coverity.com --- osinfo/osinfo_loader.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/osinfo/osinfo_loader.c b/osinfo/osinfo_loader.c index aa5e48f..859328b 100644 --- a/osinfo/osinfo_loader.c +++ b/osinfo/osinfo_loader.c @@ -673,7 +673,7 @@ static void osinfo_loader_datamap(OsinfoLoader *loader, OsinfoDatamap *map = osinfo_loader_get_datamap(loader, id); nnodes = osinfo_loader_nodeset("./entry", ctxt, &nodes, err); - if (error_is_set(err)) + if (nnodes < 0 || error_is_set(err)) goto cleanup; for (i = 0 ; i < nnodes ; i++) { @@ -911,7 +911,7 @@ static OsinfoMedia *osinfo_loader_media (OsinfoLoader *loader, } gint nnodes = osinfo_loader_nodeset("./variant", ctxt, &nodes, err); - if (error_is_set(err)) { + if (nnodes < 0 || error_is_set(err)) { g_object_unref(media); return NULL; } @@ -1002,7 +1002,7 @@ static OsinfoTree *osinfo_loader_tree (OsinfoLoader *loader, osinfo_loader_entity(loader, OSINFO_ENTITY(tree), keys, ctxt, root, err); gint nnodes = osinfo_loader_nodeset("./treeinfo/*", ctxt, &nodes, err); - if (error_is_set(err)) { + if (nnodes < 0 || error_is_set(err)) { g_object_unref(G_OBJECT(tree)); return NULL; } @@ -1176,7 +1176,7 @@ static OsinfoDeviceDriver *osinfo_loader_driver(OsinfoLoader *loader, } gint nnodes = osinfo_loader_nodeset("./*", ctxt, &nodes, err); - if (error_is_set(err)) { + if (nnodes < 0 || error_is_set(err)) { g_object_unref(G_OBJECT(driver)); return NULL; } @@ -1284,7 +1284,7 @@ static void osinfo_loader_os(OsinfoLoader *loader, g_free(nodes); nnodes = osinfo_loader_nodeset("./variant", ctxt, &nodes, err); - if (error_is_set(err)) + if (nnodes < 0 || error_is_set(err)) goto cleanup; for (i = 0 ; i < nnodes ; i++) { -- 1.9.0 From cfergeau at redhat.com Mon May 26 08:52:02 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Mon, 26 May 2014 10:52:02 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: <1400892579-7208-1-git-send-email-fidencio@redhat.com> References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> Message-ID: <20140526085202.GA21125@edamame.cdg.redhat.com> On Sat, May 24, 2014 at 02:49:39AM +0200, Fabiano Fid?ncio wrote: > osinfo_loader_nodeset() can return -1 without set an error Do you have more details about the scenario in which it occurs? It seems this can only happen happening through one of the g_return_val_if_fail()? g_return_if_fail() are used to flag programmers' errors, after such an error is reported, behaviour is undefined, so we generally don't try to handle them. Christophe -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From fabiano at fidencio.org Mon May 26 09:24:13 2014 From: fabiano at fidencio.org (=?UTF-8?Q?Fabiano_Fid=C3=AAncio?=) Date: Mon, 26 May 2014 11:24:13 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: <20140526085202.GA21125@edamame.cdg.redhat.com> References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> <20140526085202.GA21125@edamame.cdg.redhat.com> Message-ID: On Mon, May 26, 2014 at 10:52 AM, Christophe Fergeau wrote: > On Sat, May 24, 2014 at 02:49:39AM +0200, Fabiano Fid?ncio wrote: > > osinfo_loader_nodeset() can return -1 without set an error > > Do you have more details about the scenario in which it occurs? It seems > this can only happen happening through one of the > g_return_val_if_fail()? > You're right, this is the only possible scenario. > g_return_if_fail() are used to flag programmers' errors, after such an > error is reported, behaviour is undefined, so we generally don't try to > handle them. > I tried to handle the g_return_var_if_fail(). If we generally don't try to handle then, just ignore the patch. Best Regards, -- Fabiano Fid?ncio -------------- next part -------------- An HTML attachment was scrubbed... URL: From cfergeau at redhat.com Mon May 26 16:11:57 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Mon, 26 May 2014 18:11:57 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> <20140526085202.GA21125@edamame.cdg.redhat.com> Message-ID: <20140526161157.GM28598@edamame.cdg.redhat.com> On Mon, May 26, 2014 at 11:24:13AM +0200, Fabiano Fid?ncio wrote: > On Mon, May 26, 2014 at 10:52 AM, Christophe Fergeau > wrote: > > g_return_if_fail() are used to flag programmers' errors, after such an > > error is reported, behaviour is undefined, so we generally don't try to > > handle them. > > > > I tried to handle the g_return_var_if_fail(). > If we generally don't try to handle then, just ignore the patch. If this is causing warnings from coverity, it would be better if we did something about it. Christophe -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From fabiano at fidencio.org Mon May 26 22:18:29 2014 From: fabiano at fidencio.org (=?UTF-8?Q?Fabiano_Fid=C3=AAncio?=) Date: Tue, 27 May 2014 00:18:29 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: <20140526161157.GM28598@edamame.cdg.redhat.com> References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> <20140526085202.GA21125@edamame.cdg.redhat.com> <20140526161157.GM28598@edamame.cdg.redhat.com> Message-ID: On Mon, May 26, 2014 at 6:11 PM, Christophe Fergeau wrote: > On Mon, May 26, 2014 at 11:24:13AM +0200, Fabiano Fid?ncio wrote: > > On Mon, May 26, 2014 at 10:52 AM, Christophe Fergeau < > cfergeau at redhat.com> > > wrote: > > > g_return_if_fail() are used to flag programmers' errors, after such an > > > error is reported, behaviour is undefined, so we generally don't try to > > > handle them. > > > > > > > I tried to handle the g_return_var_if_fail(). > > If we generally don't try to handle then, just ignore the patch. > > If this is causing warnings from coverity, it would be better if we did > something about it. > What is your suggestion? Best Regards, -- Fabiano Fid?ncio -------------- next part -------------- An HTML attachment was scrubbed... URL: From cfergeau at redhat.com Tue May 27 08:19:29 2014 From: cfergeau at redhat.com (Christophe Fergeau) Date: Tue, 27 May 2014 10:19:29 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> <20140526085202.GA21125@edamame.cdg.redhat.com> <20140526161157.GM28598@edamame.cdg.redhat.com> Message-ID: <20140527081929.GA25829@edamame.cdg.redhat.com> On Tue, May 27, 2014 at 12:18:29AM +0200, Fabiano Fid?ncio wrote: > On Mon, May 26, 2014 at 6:11 PM, Christophe Fergeau > wrote: > > > If this is causing warnings from coverity, it would be better if we did > > something about it. > > > > What is your suggestion? What is coverity saying exactly? Is it only complaining about this particular use of g_return_val_if_fail? Christophe -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 819 bytes Desc: not available URL: From fabiano at fidencio.org Tue May 27 09:04:06 2014 From: fabiano at fidencio.org (=?UTF-8?Q?Fabiano_Fid=C3=AAncio?=) Date: Tue, 27 May 2014 11:04:06 +0200 Subject: [Libosinfo] [PATCH] loader: Negative loop bound In-Reply-To: <20140527081929.GA25829@edamame.cdg.redhat.com> References: <1400892579-7208-1-git-send-email-fidencio@redhat.com> <20140526085202.GA21125@edamame.cdg.redhat.com> <20140526161157.GM28598@edamame.cdg.redhat.com> <20140527081929.GA25829@edamame.cdg.redhat.com> Message-ID: On Tue, May 27, 2014 at 10:19 AM, Christophe Fergeau wrote: > On Tue, May 27, 2014 at 12:18:29AM +0200, Fabiano Fid?ncio wrote: > > On Mon, May 26, 2014 at 6:11 PM, Christophe Fergeau > > > wrote: > > > > > If this is causing warnings from coverity, it would be better if we did > > > something about it. > > > > > > > What is your suggestion? > > What is coverity saying exactly? Is it only complaining about this > particular use of g_return_val_if_fail? > This is exactly what coverity says: "62584 Negative loop bound Loop may run for a long time or not at all. In osinfo_loader_os: Negative value used as a loop upper bound (CWE-606)" Best Regards, -- Fabiano Fid?ncio -------------- next part -------------- An HTML attachment was scrubbed... URL: From lasse.schuirmann at gmail.com Wed May 28 12:42:02 2014 From: lasse.schuirmann at gmail.com (Lasse Schuirmann) Date: Wed, 28 May 2014 14:42:02 +0200 Subject: [Libosinfo] [PATCH] debian: Add desktop install-script Message-ID: Hello, the patch below introduces an installation script for debian. I tested it with gnome-boxes after applying the patches from https://bugzilla.gnome.org/show_bug.cgi?id=730640 It currently installs debian without asking any questions. However there are two issues I couldn't solve yet: 1. The script does install GNOME as the only desktop environment. As I understand it the desktop package I use in this command: tasksel tasksel/first multiselect standard desktop should use the appropriate desktop package for the given ISO. This doesn't work. Furthermore the several debian ISOs (e.g. KDE and lxde) have the same volume ID. 2. The script does not work without questions when no internet connection is present. I appreciate any comments. Greetings, Lasse Schuirmann >From 154d8a0c946938f389416f81b0301c6e663c3e84 Mon Sep 17 00:00:00 2001 From: Lasse Schuirmann Date: Tue, 27 May 2014 13:42:47 +0200 Subject: [PATCH] debian: Add desktop installation script --- data/install-scripts/debian.xml | 183 ++++++++++++++++++++++++++++++++++++++++ data/oses/debian.xml.in | 1 + 2 files changed, 184 insertions(+) diff --git a/data/install-scripts/debian.xml b/data/install-scripts/debian.xml index c5c8592..c1dbfa6 100644 --- a/data/install-scripts/debian.xml +++ b/data/install-scripts/debian.xml @@ -1,4 +1,5 @@ + jeos preseed.cfg @@ -70,4 +71,186 @@ d-i finish-install/reboot_in_progress note + + + + desktop + preseed.cfg + true + + + + + + + + + + + + + + + + + + + initrd + + + diff --git a/data/oses/debian.xml.in b/data/oses/debian.xml.in index 1ecb2dd..dc97d81 100644 --- a/data/oses/debian.xml.in +++ b/data/oses/debian.xml.in @@ -256,6 +256,7 @@