Aaron Griffin wrote:

On Feb 4, 2008 8:06 PM, Eric Belanger wrote:

...

On Mon, 4 Feb 2008, Paul Mattal wrote:

...

flex 2.5.34-1 is up for i686. I still haven't gotten myself going for x86_64 builds yet, so would appreciate someone doing that.

Also a minor version update. Drops the patch that was previously needed, as it appears to have been integrated. Cleaned up the install slightly, put the lex shell script in a separate file.

- P

signing off i686

built and added x86_64 package to testing.

I added a fix to have the man page in /usr/share/man/. Perhaps the i686 package should be rebuilt.

Curious, did it explicitly use /usr/man in the PKGBUILD? We have this list here: https://dev.archlinux.org/todo/47/

It did not explicitly. Thanks, Eric, for catching this. Will update i686 in the next build. http://cvs.archlinux.org/cgi-bin/viewcvs.cgi/devel/flex/PKGBUILD.diff?r1=1.22&r2=1.23&cvsroot=Core - P

Jan de Groot wrote:

On Mon, 2008-02-04 at 15:31 -0500, Paul Mattal wrote:

...

flex 2.5.34-1 is up for i686. I still haven't gotten myself going for x86_64 builds yet, so would appreciate someone doing that.

Also a minor version update. Drops the patch that was previously needed, as it appears to have been integrated. Cleaned up the install slightly, put the lex shell script in a separate file.

- P

Sorry, won't signoff:

http://sourceforge.net/tracker/index.php?func=detail&aid=1860773&group_id=97492&atid=618177

OpenOffice doesn't build anymore. It seems flex 2.5.34 can't generate the correct C file for the included scan.l. When touching the file and building flex with flex 2.5.33 installed, the resulting scan.c is totally different than the plain scan.c included in the tarball. The last comment states that flex is broken when touched again, which means that flex can't build itself. How is it supposed to build other things correctly if it can't process its own language file.

Ugh. Thanks for the catch, though someone else had already signed off, so in it went, before you sent this. Let's all see if we can find any solution to this problem in the next day or so-- a patch or some obvious bug. Otherwise, I'll roll back to a 2.5.33 release in the repos. - P

Andreas Radke wrote:

There's no quick solution available.

I reached pretty much the same conclusion. I've just built 2.5.33-3, with force=y, and put it in testing for i686. Can someone sign off? Also if someone could build the x86_64, I'd be grateful. - P

On Thu, 7 Feb 2008, Paul Mattal wrote:

[2.5.33-3 was already used.. but I didn't notice because it wasn't in the CVS log. I've just rebuilt and uploaded as 2.5.33-4. Ignore signoff request for 2.5.33-3. -P]

Andreas Radke wrote:

...

There's no quick solution available.

I reached pretty much the same conclusion.

I've just built 2.5.33-4, with force=y, and put it in testing for i686. Can someone sign off?

Also if someone could build the x86_64, I'd be grateful.

- P

I've built the x86_64 package. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.

2008/2/8, Travis Willard :

On Feb 7, 2008 11:19 PM, Eric Belanger wrote:

...

On Thu, 7 Feb 2008, Paul Mattal wrote:

...

[2.5.33-3 was already used.. but I didn't notice because it wasn't in the CVS log. I've just rebuilt and uploaded as 2.5.33-4. Ignore signoff request for 2.5.33-3. -P]

Andreas Radke wrote:

...

There's no quick solution available.

I reached pretty much the same conclusion.

I've just built 2.5.33-4, with force=y, and put it in testing for i686. Can someone sign off?

Also if someone could build the x86_64, I'd be grateful.

- P

I've built the x86_64 package.

Downloaded fine (i686) - not sure how to test.

Try to build something that requires it to build, e.g. lm_sensors -- Roman Kyrylych (Роман Кирилич)

On Thu, 7 Feb 2008, Eric Belanger wrote:

On Thu, 7 Feb 2008, Paul Mattal wrote:

...

[2.5.33-3 was already used.. but I didn't notice because it wasn't in the CVS log. I've just rebuilt and uploaded as 2.5.33-4. Ignore signoff request for 2.5.33-3. -P]

Andreas Radke wrote:

...

There's no quick solution available.

I reached pretty much the same conclusion.

I've just built 2.5.33-4, with force=y, and put it in testing for i686. Can someone sign off?

Also if someone could build the x86_64, I'd be grateful.

- P

I've built the x86_64 package.

Bump. Still missing a signoff for x86_64. -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.