Re: most commonly used regex lib for awk/frep/mutt/sed?
On Thu, Aug 01, 2002 at 01:47:37PM +0200, Sven Guckes wrote: Is this too ambitious a wish? unfortunately, yes. Ricardo SIGNES is a man of vision, I think. If it is practical to link in a preferred regex suite, as described in his post on this thread, to allow the ERE-proficient amongst us to escape the frustrations and limitations of BREs, then my only questions are: o Do I have to dust off my rusty but revivable 'C' skills to help get this moving? o Should vim be fixed first? 'cos I think Roman is definitely right. (There's still no magic that allows + in lieu of \+) let's assume you could make this change within a day. how many setup files, scripts, shell aliases etc would have to be adjusted? can you give this service? As Ricardo implies, let's not change any. It would require that people change, and I recognise that it is only in accepting the existing tower of babel, that we have any hope of providing a simpler consistent interface for those who prefer it. besides, would everyone gain from this? Oh-oh, is this devil's advocacy? Whether it is: o The time lost by countless users individually clawing their way up the learning curve of a panoply of regex dialects, just to do a simple job. o The countless hours spent by worthy individuals reinventing the wheel for their otherwise great tool. (When the time could be spent on features which help users.) o The resulting bugs, fixes, and re-releases of these tools, impacting both developer and user. o The newsgroup and mailing list traffic due to regex knowledge already acquired not being portable. (Much energy has also been expended on the procmail mailling list, examining how their dialect can be made stranger still, as various deficiencies of the older syntax are addressed.) o The confusion sown by inconsistency. Beginner pain is exacerbated by behavioural variability. Lessons learned on one dialect must be unlearned on others. (After approximately a decade, I still send stuff to grep, rather than use vim's irregular expressions.) there is benefit enough for all. For what it is worth, dissatisfaction with dialects has found its way into the debian regex(7) manpage. It describes POSIX 1003.2 EREs as modern, and is not very flattering of BREs: Obsolete (``basic'') regular expressions differ in several respects. `|', `+', and `?' are ordinary characters and there is no equivalent for their functionality. and Obsolete REs mostly exist for backward compatibility in some old programs; (No, they can't mean vim. It has the backslash work-around, even though it doesn't have modern EREs. ;-) Hmmm. Whither mutt? Regards, Erik -- _,-_|\Erik Christiansen / \ Research Development Division \_,-.__/ Voice Products Department vNEC Business Solutions Pty. Ltd.
Re: most commonly used regex lib for awk/frep/mutt/sed?
Date: Fri, 2 Aug 2002 18:55:42 +1000 From: Erik Christiansen [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: most commonly used regex lib for awk/frep/mutt/sed? On Thu, Aug 01, 2002 at 01:47:37PM +0200, Sven Guckes wrote: Is this too ambitious a wish? unfortunately, yes. Ricardo SIGNES is a man of vision, I think. If it is practical to link in a preferred regex suite, as described in his post on this thread, to allow the ERE-proficient amongst us to escape the frustrations and limitations of BREs, then my only questions are: o Do I have to dust off my rusty but revivable 'C' skills to help get this moving? someone *will* need to do it. i'd definitely love to see this inlined as a ./configure knob, but a (probably quite hard to maintain) patch looks more likely. (oh gosh, how i regret i don't know enough c) o Should vim be fixed first? 'cos I think Roman is definitely right. (There's still no magic that allows + in lieu of \+) vim needs a fix really badly imo, but i don't think any such thing will happen (soon), and even if Bram jumpstarted replacing the strange cousin of re that's used in vim, it would delay the work in mutt too much. let's assume you could make this change within a day. how many setup files, scripts, shell aliases etc would have to be adjusted? can you give this service? As Ricardo implies, let's not change any. It would require that people change, and I recognise that it is only in accepting the existing tower of babel, that we have any hope of providing a simpler consistent interface for those who prefer it. if mutt could be ./configure'd --with-pcre (nondefault, of course), there'd be virtually no problems with confusion between regexps found in various published .muttrc's and the syntax mutt linked with pcre actually expected. besides, would everyone gain from this? Oh-oh, is this devil's advocacy? Whether it is: o The time lost by countless users individually clawing their way up the learning curve of a panoply of regex dialects, just to do a simple job. o The countless hours spent by worthy individuals reinventing the wheel for their otherwise great tool. (When the time could be spent on features which help users.) o The resulting bugs, fixes, and re-releases of these tools, impacting both developer and user. o The newsgroup and mailing list traffic due to regex knowledge already acquired not being portable. (Much energy has also been expended on the procmail mailling list, examining how their dialect can be made stranger still, as various deficiencies of the older syntax are addressed.) o The confusion sown by inconsistency. Beginner pain is exacerbated by behavioural variability. Lessons learned on one dialect must be unlearned on others. (After approximately a decade, I still send stuff to grep, rather than use vim's irregular expressions.) there is benefit enough for all. i'll sign this. -- FreeBSD 4.6-STABLE 1:21PM up 2 days, 20:58, 8 users, load averages: 0.00, 0.00, 0.00
Re: most commonly used regex lib for awk/frep/mutt/sed?
* Erik Christiansen [EMAIL PROTECTED] [2002-08-01 07:48]: is there any enthusiasm for employing a most commonly used regex library, such that any improvements can feed back into all the tools using it? While procmail seems to be out on a long plank in the middle of a swamp, it would be cosmically empowering to be able to move between mutt, vim, awk, and grep without tripping over too many differences in regex dialects. Is this too ambitious a wish? unfortunately, yes. let's assume you could make this change within a day. how many setup files, scripts, shell aliases etc would have to be adjusted? can you give this service? besides, would everyone gain from this? Sven [thinking about how many people use .* who do not know about the greedy convention]
Re: most commonly used regex lib for awk/frep/mutt/sed?
On Thu, Aug 01, 2002 at 01:47:37PM +0200, Sven Guckes wrote: * Erik Christiansen [EMAIL PROTECTED] [2002-08-01 07:48]: While procmail seems to be out on a long plank in the middle of a swamp, it would be cosmically empowering to be able to move between mutt, vim, awk, and grep without tripping over too many differences in regex dialects. Is this too ambitious a wish? unfortunately, yes. let's assume you could make this change within a day. how many setup files, scripts, shell aliases etc would have to be adjusted? can you give this service? ./configure --enable-pcre echo ~/.muttrc set pcre_acknowledge=on # to avoid PCRE warnings on startup -- rjbs msg30052/pgp0.pgp Description: PGP signature
Re: most commonly used regex lib for awk/frep/mutt/sed?
On Thu, Aug 01, 2002 at 08:22:36AM -0400, Ricardo SIGNES wrote: echo ~/.muttrc set pcre_acknowledge=on # to avoid PCRE warnings on startup Ok, so I don't know how to call echo, and s/=on/yes/ Still, you get the picture. -- rjbs msg30054/pgp0.pgp Description: PGP signature
