Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
AZLet me add to this: isn't this what open source is about? The bug AZdid not manifest itself on my system (FreeBSD). You pointed out the AZissue. I fixed it. Cooperation prevailed. What's the big deal? Well, the big deal is that sometimes - probably, not this time, but sometimes - small changes break things that the author didn't even think about it in some entirely different place. And it's always unexpected and unforeseen. That's why there's feature freeze period before release - and, talking about open source, each major open source project you look at - at least, each one I ever looked at - has it. This allows all the system as a whole to be sufficiently tested so that the chance that it is broken is small. If development code breaks, as you said - it's no big deal, you just fix it. But if released code gets broken, non-development people have to live with breakage until the next release. And living with broken version is much tougher than living without a feature. Especially when we talk about such a landmark release as 5.0.0. So I think it justifies some kind of freeze paranoia :) - if we don't have this and that feature 5.0.0 is still good and people would say that's good and probably will be even better, but if it doesn't compile or doesn't work for someone - people would say oh, that's a broken one, let us not use it. -- Stanislav Malyshev, Zend Products Engineer [EMAIL PROTECTED] http://www.zend.com/ +972-3-6139665 ext.115 -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Tuesday 18 May 2004 10:41, Stanislav Malyshev wrote: Well, the big deal is that sometimes - probably, not this time, but sometimes - small changes break things that the author didn't even think about it in some entirely different place. Exactly. Case and point: loading of php extensions from php.ini or via dl() being broken at the moment due to one of those small changes. Edin -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Tue, 18 May 2004, Stanislav Malyshev wrote: Well, the big deal is that sometimes - probably, not this time, but sometimes - small changes break things that the author didn't even think about it in some entirely different place. And it's always unexpected and unforeseen. That's why there's feature freeze period before release - and, talking about open source, each major open source project you look at - at least, each one I ever looked at - has it. This allows all the system as a whole to be sufficiently tested so that the chance that it is broken is small. Once again, that is *not* what my reference to open source was about. I was simply pointing out that we fixed the bug via communication and cooperation. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Sun, 16 May 2004, Andi Gutmans wrote: Yeah I think Zeev is probably right, especially as I don't think we need more than one more RC before we release. Andrei, do you mind if we revisit this after 5.0? How long will 5.1 take after 5.0 is release? A month? A year? - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Derick Rethans wrote: I guess it would be around a year, and yes, that's a long time to wait for a new feature. Whatever happened to release early, release often philosophy? I can't, no, I refuse to believe that we have to wait another year to get a new feature out. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Andrei Zmievski wrote: On Mon, 17 May 2004, Derick Rethans wrote: I guess it would be around a year, and yes, that's a long time to wait for a new feature. Whatever happened to release early, release often philosophy? I can't, no, I refuse to believe that we have to wait another year to get a new feature out. Well, i've no problem with something this small to go into a 0.0.1 release, but IMO only minor new features should go in there and leave the 0.1 releases for larger new features. regards, Derick -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
RE: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
IMO, if we're going to be strict with the feature freeze, Andi/Zeev should have said no in the first place, rather than allowing it to be committed and then asking for it to be reverted. Personally, I'm +1 on this patch getting in PHP 5.0.0, as it is an excellent companion to the additional ini directory scan stuff. --Wez. -Original Message- From: Derick Rethans [mailto:[EMAIL PROTECTED] Sent: 17 May 2004 17:46 To: Andrei Zmievski Cc: PHP Developers Mailing List Subject: Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l On Mon, 17 May 2004, Andrei Zmievski wrote: On Mon, 17 May 2004, Derick Rethans wrote: I guess it would be around a year, and yes, that's a long time to wait for a new feature. Whatever happened to release early, release often philosophy? I can't, no, I refuse to believe that we have to wait another year to get a new feature out. Well, i've no problem with something this small to go into a 0.0.1 release, but IMO only minor new features should go in there and leave the 0.1 releases for larger new features. regards, Derick -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On May 17, 2004, at 1:08 PM, Wez Furlong wrote: IMO, if we're going to be strict with the feature freeze, Andi/Zeev should have said no in the first place, rather than allowing it to be committed and then asking for it to be reverted. Personally, I'm +1 on this patch getting in PHP 5.0.0, as it is an excellent companion to the additional ini directory scan stuff. Feature freeze has always been highly inconsistently applied. There are plenty of functions added/augmented in point-releases of php4. This sudden yearn for a high level of strictness seems to be bad. George -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
Personally, I'm +1 on this patch getting in PHP 5.0.0, as it is an excellent companion to the additional ini directory scan stuff. Feature freeze has always been highly inconsistently applied. There are plenty of functions added/augmented in point-releases of php4. This sudden yearn for a high level of strictness seems to be bad. Sure it's been inconsistently applied, but that doesn't mean that striving for consistency is inherently bad. It's especially not bad when the initial implementation of the exception in question is incompatable with one of the officially recommended set of build tools (bison 1.75 -- see: http://www.php.net/anoncvs.php ) and breaks the Win32 snap generation. Granted that's a minor bug, but it's a perfect illustration of why a feature freeze exists. -Sara -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Andrei Zmievski wrote: On Mon, 17 May 2004, Sara Golemon wrote: Sure it's been inconsistently applied, but that doesn't mean that striving for consistency is inherently bad. It's especially not bad when the initial implementation of the exception in question is incompatable with one of the officially recommended set of build tools (bison 1.75 -- see: http://www.php.net/anoncvs.php ) and breaks the Win32 snap generation. Granted that's a minor bug, but it's a perfect illustration of why a feature freeze exists. There is no such thing as a feature freeze in PHP land, historically. And that minor bug has already been fixed. Let me add to this: isn't this what open source is about? The bug did not manifest itself on my system (FreeBSD). You pointed out the issue. I fixed it. Cooperation prevailed. What's the big deal? - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
This is why I allowed important features in despite the feature freeze. However, it doesn't mean that unimportant stuff should go in as much as people want because otherwise we can't make a good release. I don't think open-source is about adding every single feature at any single point of time. The result would be horrible instability (not that I think your patch fits into this category I'm just saying that even open-source projects need some care). Anyway, due to your patch being pretty much self-contained I thought it wouldn't be a big deal to introduce it. However, thinking it over, I don't think it's a crucial enough patch to force it in. I doubt you see it this way too. What's another 3-4 months going to do? People managed without it up to now. Andi At 11:27 AM 5/17/2004 -0700, Andrei Zmievski wrote: On Mon, 17 May 2004, Andrei Zmievski wrote: On Mon, 17 May 2004, Sara Golemon wrote: Sure it's been inconsistently applied, but that doesn't mean that striving for consistency is inherently bad. It's especially not bad when the initial implementation of the exception in question is incompatable with one of the officially recommended set of build tools (bison 1.75 -- see: http://www.php.net/anoncvs.php ) and breaks the Win32 snap generation. Granted that's a minor bug, but it's a perfect illustration of why a feature freeze exists. There is no such thing as a feature freeze in PHP land, historically. And that minor bug has already been fixed. Let me add to this: isn't this what open source is about? The bug did not manifest itself on my system (FreeBSD). You pointed out the issue. I fixed it. Cooperation prevailed. What's the big deal? - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
But what separates an import feature from a slightly less important one? One hand adding self contained function would not affect any of the code outside of the function, so why not add it. That said, this function may introduce new bugs and further delay the already much overdue release. We maybe talking about a very handy feature, but this is not even self contained code, it affects the ini parsing subsystem which is used in many more places then just reading the PHP config file. People use it for application configs, PHP uses it for browscap stuff that is already flaky as it is, etc... The code could work fine in most cases but not all, we have no way of predicting the weirdness people have in their INI files. Given the number of pending features various people want to add, perhaps now is the time for the 5.1 branch where all the new features can go to and if some prove to be stable and reliable. Then perhaps they can be introduced in 5.0.1, which would probably soon follow the 5.0 release. Without a development branch we are likely to lose many interesting features as the patches would get lost on developers' drivers or stop being compatible with latest code revisions. While it certainly may complicate bug fixing as fixes will need to be ported across 3 trees in some cases, but that's a small price pay for gained flexibility. Ilia -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
I agree with everything Sara says. So we delay it for a few months, what's the big deal there? There are several features that we want to insert to PHP 5 but are not because of the feature freeze (realpath cache, TSRM updates, etc.). There'll be plenty of reasons to go for 5.1 pretty much immediately after the release of 5.0. Zeev At 22:17 17/05/2004, Sara Golemon wrote: There is no such thing as a feature freeze in PHP land, historically. And that minor bug has already been fixed. Oh what wonderful news, that means I can start commit patches again! Let's see, I've got that source binding patch for the network stuff, I've got those compression filters that I thought were going to have to wait till 5.1 and be supported in 5.0 as a PECL package only so that's good. Oh and here's that nice little guy to support HTTP/1.1 chunked encoding, inline gzip deflates, and connection pooling... I guess if there's no feature freeze in PHP land then all this stuff can go in right now. Either there is a feature freeze or there isn't. I appreciate the frustration of having to wait through long and tedious minor version bumps. I also appreciate the fact that historically PHP versioning has been sloppy, but that's not a justification for contuing bad behavior. Let me add to this: isn't this what open source is about? The bug did not manifest itself on my system (FreeBSD). You pointed out the issue. I fixed it. Cooperation prevailed. What's the big deal? Let me respond by saying there wasn't a big deal. I asked if this was too late in the RC cycle to justify a non-critical exception to the feature freeze. The response of Andi/Zeev okayed it was fine. But now you're standing there telling me that there is no such thing as a feature freeze and THAT is a big deal. That promotes chaos and a bad end product. I don't give a rodent's hind quarters about the ini variable substitution patch, I care about putting out a good product. I care about knowing that there's nothing in the final release that hasn't been well tested. I care that I only saw that minor bug because it broke compilation in general, and god knows if it broke something less obvious. Has anyone tried this with the ever-problematic browsecap.ini file? -Sara -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
Using 4.x as a guide, it's around 1 minor release per year. Waiting a month or two for a feature seems completely reasonable. Waiting a year for it seems onerous. George On May 17, 2004, at 3:41 PM, Zeev Suraski wrote: I agree with everything Sara says. So we delay it for a few months, what's the big deal there? There are several features that we want to insert to PHP 5 but are not because of the feature freeze (realpath cache, TSRM updates, etc.). There'll be plenty of reasons to go for 5.1 pretty much immediately after the release of 5.0. -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
At 22:49 17/05/2004, George Schlossnagle wrote: Using 4.x as a guide, it's around 1 minor release per year. Waiting a month or two for a feature seems completely reasonable. Waiting a year for it seems onerous. 4.x is completely irrelevant, because we only switched to the approach that 3rd digit releases contain only bug fixes at a very late stage. In the 4.0/4.1 (and maybe 4.2 as well, I don't remember) branches, each 3rd digit release contained plenty of new functionality. 2nd digit releases symbolized major changes in architecture/compatibility/etc. Zeev -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Andi Gutmans wrote: This is why I allowed important features in despite the feature freeze. However, it doesn't mean that unimportant stuff should go in as much as people want because otherwise we can't make a good release. I don't think open-source is about adding every single feature at any single point of time. The result would be horrible instability (not that I think your patch fits into this category I'm just saying that even open-source projects need some care). I never said that open source is about adding every single feature. My comment was restricted to Sara's reply about my patch being buggy on certain versions of bison. Anyway, due to your patch being pretty much self-contained I thought it wouldn't be a big deal to introduce it. However, thinking it over, I don't think it's a crucial enough patch to force it in. I doubt you see it this way too. What's another 3-4 months going to do? People managed without it up to now. I've gotten some pretty strong responses about this patch being a very useful one. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Ilia Alshanetsky wrote: Given the number of pending features various people want to add, perhaps now is the time for the 5.1 branch where all the new features can go to and if some prove to be stable and reliable. Then perhaps they can be introduced in 5.0.1, which would probably soon follow the 5.0 release. Without a development branch we are likely to lose many interesting features as the patches would get lost on developers' drivers or stop being compatible with latest code revisions. While it certainly may complicate bug fixing as fixes will need to be ported across 3 trees in some cases, but that's a small price pay for gained flexibility. I don't mind if it goes into 5.0.1, but waiting a months and months for 5.1 release (which is how long it might take for all intents and purposes) is not very interesting to me. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Mon, 17 May 2004, Sara Golemon wrote: Let me respond by saying there wasn't a big deal. I asked if this was too late in the RC cycle to justify a non-critical exception to the feature freeze. The response of Andi/Zeev okayed it was fine. But now you're standing there telling me that there is no such thing as a feature freeze and THAT is a big deal. Okay, perhaps I shoulve have said that what we have is an approximation of a feature freeze. We've never had a complete one, because things besides well tested bug fixes do get committed. I appreciate your attention to the commits, but let he who is without a sin cast the first stone.. Speaking generally here, not specifically. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
It's reverted. Hope everyone's happy. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
Thanks Andrei. I'm sorry about this mess (I know it's my fault). I think it's becoming clear things are being held back because of the 5.0 release. As I mentioned I also have some work I'm holding back. I suggest so that we make sure that things aren't being held back for too long we plan on releasing 5.0 ASAP (IMO beginning of July would be a good and reasonable goal. Besides a few bug fixes I don't have anything else on my TODO). Once we release, it's probably best to branch 5.1 where we can start adding the new features which need more testing (like the ones I'm hoping to add) and keep 5.0 for bug fixes. We can then probably release 5.1 within a short period of time. Is that OK? Andi At 01:10 PM 5/17/2004 -0700, Andrei Zmievski wrote: It's reverted. Hope everyone's happy. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
On Tue, 18 May 2004, Andi Gutmans wrote: Thanks Andrei. I'm sorry about this mess (I know it's my fault). I think it's becoming clear things are being held back because of the 5.0 release. As I mentioned I also have some work I'm holding back. I suggest so that we make sure that things aren't being held back for too long we plan on releasing 5.0 ASAP (IMO beginning of July would be a good and reasonable goal. Besides a few bug fixes I don't have anything else on my TODO). Once we release, it's probably best to branch 5.1 where we can start adding the new features which need more testing (like the ones I'm hoping to add) and keep 5.0 for bug fixes. We can then probably release 5.1 within a short period of time. Is that OK? Let's do it. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
It may be worth noting the way MySQL AB has been rolling out MySQL. They released 4.0.0 as a restructured release, did bug fixes and small changes up through 4.0.19, and are approaching the release of 4.1 with significant new features. It sounds like the same might apply here. ~Jason At 5/18/2004 12:03 AM +0300, Andi Gutmans wrote: Thanks Andrei. I'm sorry about this mess (I know it's my fault). I think it's becoming clear things are being held back because of the 5.0 release. As I mentioned I also have some work I'm holding back. I suggest so that we make sure that things aren't being held back for too long we plan on releasing 5.0 ASAP (IMO beginning of July would be a good and reasonable goal. Besides a few bug fixes I don't have anything else on my TODO). Once we release, it's probably best to branch 5.1 where we can start adding the new features which need more testing (like the ones I'm hoping to add) and keep 5.0 for bug fixes. We can then probably release 5.1 within a short period of time. Is that OK? Andi At 01:10 PM 5/17/2004 -0700, Andrei Zmievski wrote: It's reverted. Hope everyone's happy. - Andrei -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
George Schlossnagle wrote: Using 4.x as a guide, it's around 1 minor release per year. Why not follow the good example of for instance the GNOME project? They make 2 minor releases per year (with bugfix release in between) and have a public release schedule / feature roadmap for each release. -- Sebastian Bergmann http://sebastian-bergmann.de/ http://phpOpenTracker.de/ Das Buch zu PHP 5: http://professionelle-softwareentwicklung-mit-php5.de/ -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
At 01:10 PM 5/15/2004 -0700, Sara Golemon wrote: A) Isn't it late in the RC cycle to be adding features? Although personally I don't think this patch is very important/useful, it's quite self-contained so we thought it'd wouldn't be a big deal to add it. It is kind of late and I do prefer not to add new features at this time unless they are tiny or important... I guess I saw this one as tiny. Andi -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
Yeah I think Zeev is probably right, especially as I don't think we need more than one more RC before we release. Andrei, do you mind if we revisit this after 5.0? Andi At 12:00 PM 5/16/2004 +0300, Zeev Suraski wrote: At 09:11 16/05/2004, Andi Gutmans wrote: At 01:10 PM 5/15/2004 -0700, Sara Golemon wrote: A) Isn't it late in the RC cycle to be adding features? Although personally I don't think this patch is very important/useful, it's quite self-contained so we thought it'd wouldn't be a big deal to add it. It is kind of late and I do prefer not to add new features at this time unless they are tiny or important... I guess I saw this one as tiny. Even though I think this patch is fine, I think we're best off waiting with it until 5.1, if only to avoid opening the door to additional tiny patches. Especially when it's not very easy to define the borderline between what's tiny and what isn't. Andrei - do you mind reverting the patch until 5.1? Zeev -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Re: ZendEngine2 / zend_ini_parser.y zend_ini_scanner.l
A) Isn't it late in the RC cycle to be adding features? I guess not. I got OK from Andi Zeev. I realized after sending that it sounded a bit hostile. I apologize for that, but I was under the impression that we're into critical-only territory at this point, and this feature is hardly critical. No strong feelings one way or the other though. B) You're mising a ; (line 259) Where exactly? What file? It compiles fine for me. Zend/zend_ini_parse.y http://cvs.php.net/co.php/ZendEngine2/zend_ini_parser.y And in fact, now that I look again, you're missing a couple of 'em. One to terminate the var_string_list: block and another to terminate the cfg_var_ref: block. Maybe your version of bison is more forgiving than mine, I'm using 1.75. Looks like it's enough to stop the Win32 snap though: http://snaps.php.net/win32/snapshot.log \cygwin\bin\bison.exe --output=Zend/zend_ini_parser.c -v -d -p ini_ Zend/zend_ini_parser.y Zend/zend_ini_parser.y:261.12: parse error, unexpected :, expecting ; or | Zend/zend_ini_parser.y:267.20-54: invalid $ value Zend/zend_ini_parser.y:268.20-54: invalid $ value NMAKE : fatal error U1077: '\cygwin\bin\bison.exe' : return code '0x1' Stop. NMAKE : fatal error U1077: 'c:\Program Files\Microsoft Visual Studio\VC98\bin\NMAKE.EXE' : return code '0x2' Stop. My error message under Linux amounts to the same: bison -y -p ini_ -v -d /home/sarag/cvs/php5/Zend/zend_ini_parser.y -o Zend/zend_ini_parser.c /home/sarag/cvs/php5/Zend/zend_ini_parser.y:261.12: parse error, unexpected :, expecting ; or | /home/sarag/cvs/php5/Zend/zend_ini_parser.y:267.20-54: invalid $ value /home/sarag/cvs/php5/Zend/zend_ini_parser.y:268.20-54: invalid $ value make: *** [Zend/zend_ini_parser.c] Error 1 -Sara -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php