[gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 405 bytes --]

looks like bash-4.0 has broken semicolon escaping in subshells.  this comes up 
when using find's -exec like we do in a few places in eclasses:
                ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift
you can work around the issue in a couple of ways:
 - quote the semicolon:
	.... ';')
 - use backticks
	`find .... \;`

i'll tweak the eclasses to use quoting for now
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 835 bytes --]

[gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Ryan Hill]

[-- Attachment #1: Type: text/plain, Size: 761 bytes --]

On Sat, 21 Feb 2009 18:27:10 -0500
Mike Frysinger <vapier@gentoo.org> wrote:

> looks like bash-4.0 has broken semicolon escaping in subshells.  this
> comes up when using find's -exec like we do in a few places in
> eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift
> you can work around the issue in a couple of ways:
>  - quote the semicolon:
> 	.... ';')
>  - use backticks
> 	`find .... \;`
> 
> i'll tweak the eclasses to use quoting for now
> -mike
> 

is this a bug or broken on purpose?

-- 
gcc-porting,                                      by design, by neglect
treecleaner,                              for a fact or just for effect
wxwidgets @ gentoo     EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 729 bytes --]

On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > comes up when using find's -exec like we do in a few places in
> > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift
> > you can work around the issue in a couple of ways:
> >  - quote the semicolon:
> > 	.... ';')
> >  - use backticks
> > 	`find .... \;`
> >
> > i'll tweak the eclasses to use quoting for now
>
> is this a bug or broken on purpose?

i say it's a bug, but i'm not the bash maintainer

i imagine it's fall out from attempts to fix support for case statements in 
subshells
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 835 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mart Raudsepp]

[-- Attachment #1: Type: text/plain, Size: 1272 bytes --]

On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > comes up when using find's -exec like we do in a few places in
> > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift
> > > you can work around the issue in a couple of ways:
> > >  - quote the semicolon:
> > > 	.... ';')
> > >  - use backticks
> > > 	`find .... \;`
> > >
> > > i'll tweak the eclasses to use quoting for now
> >
> > is this a bug or broken on purpose?
> 
> i say it's a bug, but i'm not the bash maintainer
> 
> i imagine it's fall out from attempts to fix support for case statements in 
> subshells

Then the bug should be fixed, instead of changing usage to something
apparently less common, as the conversion could miss some.  And more
importantly users still want to use \; for find -exec ending on their
command line and their very own scripts.
And who knows how many shell scripts shipped by packages use the
escaping method.


-- 
Mart Raudsepp
Gentoo Developer
Mail: leio@gentoo.org
Weblog: http://planet.gentoo.org/developers/leio

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1385 bytes --]

On Saturday 21 February 2009 19:00:19 Mart Raudsepp wrote:
> On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> > On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > > comes up when using find's -exec like we do in a few places in
> > > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;);
> > > > shift you can work around the issue in a couple of ways:
> > > >  - quote the semicolon:
> > > > 	.... ';')
> > > >  - use backticks
> > > > 	`find .... \;`
> > > >
> > > > i'll tweak the eclasses to use quoting for now
> > >
> > > is this a bug or broken on purpose?
> >
> > i say it's a bug, but i'm not the bash maintainer
> >
> > i imagine it's fall out from attempts to fix support for case statements
> > in subshells
>
> Then the bug should be fixed, instead of changing usage to something
> apparently less common, as the conversion could miss some.  And more
> importantly users still want to use \; for find -exec ending on their
> command line and their very own scripts.
> And who knows how many shell scripts shipped by packages use the
> escaping method.

i think you missed the entire point of this thread: there's a bug in bash-4.0 
that code is likely to hit.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 835 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mart Raudsepp]

[-- Attachment #1: Type: text/plain, Size: 1790 bytes --]

On Sat, 2009-02-21 at 19:29 -0500, Mike Frysinger wrote:
> On Saturday 21 February 2009 19:00:19 Mart Raudsepp wrote:
> > On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> > > On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > > > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > > > comes up when using find's -exec like we do in a few places in
> > > > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;);
> > > > > shift you can work around the issue in a couple of ways:
> > > > >  - quote the semicolon:
> > > > > 	.... ';')
> > > > >  - use backticks
> > > > > 	`find .... \;`
> > > > >
> > > > > i'll tweak the eclasses to use quoting for now
> > > >
> > > > is this a bug or broken on purpose?
> > >
> > > i say it's a bug, but i'm not the bash maintainer
> > >
> > > i imagine it's fall out from attempts to fix support for case statements
> > > in subshells
> >
> > Then the bug should be fixed, instead of changing usage to something
> > apparently less common, as the conversion could miss some.  And more
> > importantly users still want to use \; for find -exec ending on their
> > command line and their very own scripts.
> > And who knows how many shell scripts shipped by packages use the
> > escaping method.
> 
> i think you missed the entire point of this thread: there's a bug in bash-4.0 
> that code is likely to hit.

I think you missed the entire point of my reply.
That bug should be fixed, not workarounds applied all over the tree, as
users still want to be able to escape semi-colons.


-- 
Mart Raudsepp
Gentoo Developer
Mail: leio@gentoo.org
Weblog: http://planet.gentoo.org/developers/leio

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 2292 bytes --]

On Saturday 21 February 2009 19:38:33 Mart Raudsepp wrote:
> On Sat, 2009-02-21 at 19:29 -0500, Mike Frysinger wrote:
> > On Saturday 21 February 2009 19:00:19 Mart Raudsepp wrote:
> > > On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> > > > On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > > > > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > > > > looks like bash-4.0 has broken semicolon escaping in subshells. 
> > > > > > this comes up when using find's -exec like we do in a few places
> > > > > > in eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po
> > > > > > \;); shift you can work around the issue in a couple of ways: -
> > > > > > quote the semicolon:
> > > > > > 	.... ';')
> > > > > >  - use backticks
> > > > > > 	`find .... \;`
> > > > > >
> > > > > > i'll tweak the eclasses to use quoting for now
> > > > >
> > > > > is this a bug or broken on purpose?
> > > >
> > > > i say it's a bug, but i'm not the bash maintainer
> > > >
> > > > i imagine it's fall out from attempts to fix support for case
> > > > statements in subshells
> > >
> > > Then the bug should be fixed, instead of changing usage to something
> > > apparently less common, as the conversion could miss some.  And more
> > > importantly users still want to use \; for find -exec ending on their
> > > command line and their very own scripts.
> > > And who knows how many shell scripts shipped by packages use the
> > > escaping method.
> >
> > i think you missed the entire point of this thread: there's a bug in
> > bash-4.0 that code is likely to hit.
>
> I think you missed the entire point of my reply.
> That bug should be fixed, not workarounds applied all over the tree, as
> users still want to be able to escape semi-colons.

no one suggested doing any of this crap you're talking about.  if you want to 
get all retarded, dont install the masked ebuild.  i gave a heads up to people 
who might want to experiment so they wouldnt have to figure out weird errors.  
in the mean time, i tweaked a few common files so people wouldnt hit errors 
and could investigate even further.

i guess in the future i simply wont post heads up so i dont have to listen to 
people whine about non-existent issues.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 835 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mart Raudsepp]

[-- Attachment #1: Type: text/plain, Size: 2601 bytes --]

On Sat, 2009-02-21 at 19:44 -0500, Mike Frysinger wrote:
> On Saturday 21 February 2009 19:38:33 Mart Raudsepp wrote:
> > On Sat, 2009-02-21 at 19:29 -0500, Mike Frysinger wrote:
> > > On Saturday 21 February 2009 19:00:19 Mart Raudsepp wrote:
> > > > On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> > > > > On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > > > > > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > > > > > looks like bash-4.0 has broken semicolon escaping in subshells. 
> > > > > > > this comes up when using find's -exec like we do in a few places
> > > > > > > in eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po
> > > > > > > \;); shift you can work around the issue in a couple of ways: -
> > > > > > > quote the semicolon:
> > > > > > > 	.... ';')
> > > > > > >  - use backticks
> > > > > > > 	`find .... \;`
> > > > > > >
> > > > > > > i'll tweak the eclasses to use quoting for now
> > > > > >
> > > > > > is this a bug or broken on purpose?
> > > > >
> > > > > i say it's a bug, but i'm not the bash maintainer
> > > > >
> > > > > i imagine it's fall out from attempts to fix support for case
> > > > > statements in subshells
> > > >
> > > > Then the bug should be fixed, instead of changing usage to something
> > > > apparently less common, as the conversion could miss some.  And more
> > > > importantly users still want to use \; for find -exec ending on their
> > > > command line and their very own scripts.
> > > > And who knows how many shell scripts shipped by packages use the
> > > > escaping method.
> > >
> > > i think you missed the entire point of this thread: there's a bug in
> > > bash-4.0 that code is likely to hit.
> >
> > I think you missed the entire point of my reply.
> > That bug should be fixed, not workarounds applied all over the tree, as
> > users still want to be able to escape semi-colons.
> 
> no one suggested doing any of this crap you're talking about.  if you want to 
> get all retarded, dont install the masked ebuild.  i gave a heads up to people 
> who might want to experiment so they wouldnt have to figure out weird errors.  
> in the mean time, i tweaked a few common files so people wouldnt hit errors 
> and could investigate even further.

Perhaps you should actually state those intentions at the start instead
of starting to rant out on people replying.
Sounds good now that we actually know what the plan is.


-- 
Mart Raudsepp
Gentoo Developer
Mail: leio@gentoo.org
Weblog: http://planet.gentoo.org/developers/leio

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

[gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Ryan Hill]

[-- Attachment #1: Type: text/plain, Size: 1094 bytes --]

On Sat, 21 Feb 2009 18:55:37 -0500
Mike Frysinger <vapier@gentoo.org> wrote:

> On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > looks like bash-4.0 has broken semicolon escaping in subshells.
> > > this comes up when using find's -exec like we do in a few places
> > > in eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po
> > > \;); shift you can work around the issue in a couple of ways:
> > >  - quote the semicolon:
> > > 	.... ';')
> > >  - use backticks
> > > 	`find .... \;`
> > >
> > > i'll tweak the eclasses to use quoting for now
> >
> > is this a bug or broken on purpose?
> 
> i say it's a bug, but i'm not the bash maintainer
> 
> i imagine it's fall out from attempts to fix support for case
> statements in subshells

thanks for the heads up.

-- 
gcc-porting,                                      by design, by neglect
treecleaner,                              for a fact or just for effect
wxwidgets @ gentoo     EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 2852 bytes --]

On Saturday 21 February 2009 19:54:51 Mart Raudsepp wrote:
> On Sat, 2009-02-21 at 19:44 -0500, Mike Frysinger wrote:
> > On Saturday 21 February 2009 19:38:33 Mart Raudsepp wrote:
> > > On Sat, 2009-02-21 at 19:29 -0500, Mike Frysinger wrote:
> > > > On Saturday 21 February 2009 19:00:19 Mart Raudsepp wrote:
> > > > > On Sat, 2009-02-21 at 18:55 -0500, Mike Frysinger wrote:
> > > > > > On Saturday 21 February 2009 18:38:55 Ryan Hill wrote:
> > > > > > > On Sat, 21 Feb 2009 18:27:10 -0500 Mike Frysinger wrote:
> > > > > > > > looks like bash-4.0 has broken semicolon escaping in
> > > > > > > > subshells. this comes up when using find's -exec like we do
> > > > > > > > in a few places in eclasses: ls=$(find "$1" -name '*.po'
> > > > > > > > -exec basename {} .po \;); shift you can work around the
> > > > > > > > issue in a couple of ways: - quote the semicolon:
> > > > > > > > 	.... ';')
> > > > > > > >  - use backticks
> > > > > > > > 	`find .... \;`
> > > > > > > >
> > > > > > > > i'll tweak the eclasses to use quoting for now
> > > > > > >
> > > > > > > is this a bug or broken on purpose?
> > > > > >
> > > > > > i say it's a bug, but i'm not the bash maintainer
> > > > > >
> > > > > > i imagine it's fall out from attempts to fix support for case
> > > > > > statements in subshells
> > > > >
> > > > > Then the bug should be fixed, instead of changing usage to
> > > > > something apparently less common, as the conversion could miss
> > > > > some.  And more importantly users still want to use \; for find
> > > > > -exec ending on their command line and their very own scripts.
> > > > > And who knows how many shell scripts shipped by packages use the
> > > > > escaping method.
> > > >
> > > > i think you missed the entire point of this thread: there's a bug in
> > > > bash-4.0 that code is likely to hit.
> > >
> > > I think you missed the entire point of my reply.
> > > That bug should be fixed, not workarounds applied all over the tree, as
> > > users still want to be able to escape semi-colons.
> >
> > no one suggested doing any of this crap you're talking about.  if you
> > want to get all retarded, dont install the masked ebuild.  i gave a heads
> > up to people who might want to experiment so they wouldnt have to figure
> > out weird errors. in the mean time, i tweaked a few common files so
> > people wouldnt hit errors and could investigate even further.
>
> Perhaps you should actually state those intentions at the start instead
> of starting to rant out on people replying.
> Sounds good now that we actually know what the plan is.

i guess i used too many code words like "bash-4.0 is broken" and "workaround".  
i'll address this in the future by just mailing base-system@gentoo.org as they 
should be familiar with these insider terms.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 835 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Nirbheek Chauhan]

On Sun, Feb 22, 2009 at 6:58 AM, Mike Frysinger <vapier@gentoo.org> wrote:
> i guess i used too many code words like "bash-4.0 is broken" and "workaround".
> i'll address this in the future by just mailing base-system@gentoo.org as they
> should be familiar with these insider terms.

Cool it with the war of words folks :)

Mike, thanks for the heads-up regarding bash-4.0, it's really helpful
to have exchange of information to prevent a communication gap (which
only results in conflict).

Mart, if something is unclear, explicitly asking it would be better
than replying with (possibly false) assumptions. Although I agree that
context *is* important for us mere mortals (mike, *hint hint* ;)

This was your daily dose of chuckles, huggles, and fun,

Cheers,

~Nirbheek Chauhan

Re: [gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Dawid Węgliński]

On Sunday 22 of February 2009 00:27:10 Mike Frysinger wrote:
> looks like bash-4.0 has broken semicolon escaping in subshells.  this comes
> up when using find's -exec like we do in a few places in eclasses:
>                 ls=$(find "$1" -name '*.po' -exec basename {} .po \;);
> shift you can work around the issue in a couple of ways:
>  - quote the semicolon:
> 	.... ';')
>  - use backticks
> 	`find .... \;`
>
> i'll tweak the eclasses to use quoting for now
> -mike

FYI. Not only find's semicolons are affected. It also happens in case ;; 
construction.

-- 
Cheers,
Dawid Węgliński

Re: [gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 842 bytes --]

On Sunday 22 February 2009 17:30:09 Dawid Węgliński wrote:
> On Sunday 22 of February 2009 00:27:10 Mike Frysinger wrote:
> > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > comes up when using find's -exec like we do in a few places in eclasses:
> > ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift you can work
> > around the issue in a couple of ways:
> >  - quote the semicolon:
> > 	.... ';')
> >  - use backticks
> > 	`find .... \;`
> >
> > i'll tweak the eclasses to use quoting for now
>
> FYI. Not only find's semicolons are affected. It also happens in case ;;
> construction.

embedded case statements in $(...) subshells have always been broken.  
bash-4.0 is supposed to fix that.  if you have some code that is broken, 
please post it so i can push it upstream.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Dawid Węgliński]

On Sunday 22 of February 2009 23:39:11 Mike Frysinger wrote:
> On Sunday 22 February 2009 17:30:09 Dawid Węgliński wrote:
> > On Sunday 22 of February 2009 00:27:10 Mike Frysinger wrote:
> > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > comes up when using find's -exec like we do in a few places in
> > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;); shift
> > > you can work around the issue in a couple of ways:
> > >  - quote the semicolon:
> > > 	.... ';')
> > >  - use backticks
> > > 	`find .... \;`
> > >
> > > i'll tweak the eclasses to use quoting for now
> >
> > FYI. Not only find's semicolons are affected. It also happens in case ;;
> > construction.
>
> embedded case statements in $(...) subshells have always been broken.
> bash-4.0 is supposed to fix that.  if you have some code that is broken,
> please post it so i can push it upstream.
> -mike

It wasn't me who experienced that, but a user:

13:50 < diabel-> <<<          dir /usr/share/doc/wxGTK-2.8.9.1-r3
13:50 < diabel-> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: 
line 2989: błąd składni przy nieoczekiwanym znaczniku `;;'
13:50 < diabel-> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: 
line 2989: `        ;;' * * ERROR: x11-libs/wxGTK-2.8.9.1-r3 failed.

All it states is syntax error near double semicolons.

Re: [gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1741 bytes --]

On Sunday 22 February 2009 18:03:23 Dawid Węgliński wrote:
> On Sunday 22 of February 2009 23:39:11 Mike Frysinger wrote:
> > On Sunday 22 February 2009 17:30:09 Dawid Węgliński wrote:
> > > On Sunday 22 of February 2009 00:27:10 Mike Frysinger wrote:
> > > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > > comes up when using find's -exec like we do in a few places in
> > > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;);
> > > > shift you can work around the issue in a couple of ways:
> > > >  - quote the semicolon:
> > > > 	.... ';')
> > > >  - use backticks
> > > > 	`find .... \;`
> > > >
> > > > i'll tweak the eclasses to use quoting for now
> > >
> > > FYI. Not only find's semicolons are affected. It also happens in case
> > > ;; construction.
> >
> > embedded case statements in $(...) subshells have always been broken.
> > bash-4.0 is supposed to fix that.  if you have some code that is broken,
> > please post it so i can push it upstream.
>
> It wasn't me who experienced that, but a user:
>
> 13:50 < diabel-> <<<          dir /usr/share/doc/wxGTK-2.8.9.1-r3
> 13:50 < diabel->
> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: line 2989:
> błąd składni przy nieoczekiwanym znaczniku `;;'
> 13:50 < diabel->
> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: line 2989: `  
>      ;;' * * ERROR: x11-libs/wxGTK-2.8.9.1-r3 failed.
>
> All it states is syntax error near double semicolons.

unfortunately, that info isnt useful in any way.  it looks like it's a binpkg 
though, so getting the files again should be possible.  please have the user 
find that file from the binpkg and report a bug.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-dev] bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 1629 bytes --]

On Sunday 22 February 2009 18:03:23 Dawid Węgliński wrote:
> On Sunday 22 of February 2009 23:39:11 Mike Frysinger wrote:
> > On Sunday 22 February 2009 17:30:09 Dawid Węgliński wrote:
> > > On Sunday 22 of February 2009 00:27:10 Mike Frysinger wrote:
> > > > looks like bash-4.0 has broken semicolon escaping in subshells.  this
> > > > comes up when using find's -exec like we do in a few places in
> > > > eclasses: ls=$(find "$1" -name '*.po' -exec basename {} .po \;);
> > > > shift you can work around the issue in a couple of ways:
> > > >  - quote the semicolon:
> > > > 	.... ';')
> > > >  - use backticks
> > > > 	`find .... \;`
> > > >
> > > > i'll tweak the eclasses to use quoting for now
> > >
> > > FYI. Not only find's semicolons are affected. It also happens in case
> > > ;; construction.
> >
> > embedded case statements in $(...) subshells have always been broken.
> > bash-4.0 is supposed to fix that.  if you have some code that is broken,
> > please post it so i can push it upstream.
>
> It wasn't me who experienced that, but a user:
>
> 13:50 < diabel-> <<<          dir /usr/share/doc/wxGTK-2.8.9.1-r3
> 13:50 < diabel->
> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: line 2989:
> błąd składni przy nieoczekiwanym znaczniku `;;'
> 13:50 < diabel->
> /var/tmp/binpkgs/x11-libs/wxGTK-2.8.9.1-r3/temp/environment: line 2989: `  
>      ;;' * * ERROR: x11-libs/wxGTK-2.8.9.1-r3 failed.
>
> All it states is syntax error near double semicolons.

this is probably the same issue i pointed out originally.  re-emerge bash and 
it should work.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Daniel Gryniewicz]

On Sat, 2009-02-21 at 19:44 -0500, Mike Frysinger wrote:

<snip>

> > > > > > >
> > > > > > > i'll tweak the eclasses to use quoting for now

> 
> no one suggested doing any of this crap you're talking about.  if you want to 
> get all retarded, dont install the masked ebuild.  i gave a heads up to people 
> who might want to experiment so they wouldnt have to figure out weird errors.  
> in the mean time, i tweaked a few common files so people wouldnt hit errors 
> and could investigate even further.
> 
> i guess in the future i simply wont post heads up so i dont have to listen to 
> people whine about non-existent issues.


I personally took the quoted line above to indicate that changes would
be made to the tree.  I can understand how confusion would occur.

Dan

Re: [gentoo-dev] Re: bash-4.0 regression heads up (escaped semicolons in subshells)

[Mike Frysinger]

On Tuesday 24 February 2009 10:52:55 Daniel Gryniewicz wrote:
> On Sat, 2009-02-21 at 19:44 -0500, Mike Frysinger wrote:
> > > > > > > > i'll tweak the eclasses to use quoting for now
> >
> > no one suggested doing any of this crap you're talking about.  if you
> > want to get all retarded, dont install the masked ebuild.  i gave a heads
> > up to people who might want to experiment so they wouldnt have to figure
> > out weird errors. in the mean time, i tweaked a few common files so
> > people wouldnt hit errors and could investigate even further.
> >
> > i guess in the future i simply wont post heads up so i dont have to
> > listen to people whine about non-existent issues.
>
> I personally took the quoted line above to indicate that changes would
> be made to the tree.

that's because i did make changes to the tree.  "for now".  temporal clauses 
change the meaning significantly which is why the whole sentence needs to be 
digested rather than scraping a few bits.
-mike